Humanness Voice Annotator turns audio into a whole-recording description and a timed performance script: spoken words, local delivery and emotion cues, free-form vocal events and pauses.
Original audio is embedded as Base64 MP3. No model or API runs in your browser. The downloadable listening HTML includes its own styles, diagram and players.
Humaneness Ears is a Whisper encoder with task heads, not an ASR decoder. MOSS event descriptions receive the full audio and both endpoints; MOSS does not discover those endpoints. This release connects existing models and preserves their raw predictions.
The original full-recording MOSS caption: voice, performance and audible environment. It is preserved even when a claim is incorrect.
SCRIPT
Recognized speech with a round-bracket delivery cue and duration before each timed segment. Free-form vocal descriptions appear at their approximate word-gap positions. Square brackets mark alignment gaps.
Illustrative syntax; not an evaluated example. Total: 4.20 s. The inline laugh is included in its speech budget.
Use the JSON for exact event starts/ends and overlaps. A printed duration is a formatting budget, not proof that a detector proposal is real. Segments may be shorter than grammatical sentences. Read the complete contract →
Audio-based evaluation of the final format
Gemini 3.8 Flash listened to all 50 original recordings and reviewed the emitted annotations. It assessed global emotion, local emotion, talking style, vocal descriptions and timing separately. Intended TTS scripts and target emotions were withheld. Higher is better on a 0–5 scale.
Criterion
Mean
95% clip interval
Weight
How to read it
Weighted composite
3.263 / 5
3.192–3.332
100%
Fixed weighted score, equal weight per recording.
Whole-recording caption fidelity
3.82 / 5
3.70–3.92
15%
Does GENERAL describe the audible voice, performance and environment accurately, without invented demographic/identity/psychological claims?
Whole-recording emotion fit
3.74 / 5
3.60–3.86
10%
Are the whole-utterance emotion descriptions supported by audible affect, including intensity and changes? Important omissions reduce usefulness; restrained descriptions are appropriate when affect is ambiguous.
Local speech emotion fit
2.78 / 5
2.62–2.92
20%
Do emotion cues fit their particular spoken segments and intensity? Consider both false emotion tags and important missing affect. No emotion tag is not a claim of neutrality: do not penalize appropriate restraint on ambiguous segments.
Whole-recording talking-style fit
3.86 / 5
3.76–3.94
15%
Do GENERAL's speaking-style claims match audible tempo, articulation, breathiness, pitch movement, register, emphasis and performance style?
Local speech delivery fit
3.10 / 5
3.02–3.20
10%
Do local delivery/voice tags fit the corresponding spoken segment rather than merely repeating a global trait?
Burst-description fit
3.02 / 5
2.86–3.16
10%
Does each free-form burst caption match the actual nonverbal vocal sound in its selected interval? Reward acoustic accuracy, not verbosity.
Burst-timing fit
2.52 / 5
2.36–2.68
10%
Do absolute event starts/ends cover the audible events without unnecessary unrelated speech?
Event precision rating
2.88 / 5
2.74–3.00
5%
Are proposed intervals real nonverbal vocal events rather than spoken syllables or spurious detections? This is a reviewer rating, not measured detection precision.
Event coverage rating
3.46 / 5
3.32–3.60
5%
Are salient audible nonverbal events covered? This is a reviewer rating, not measured detection recall.
Transcript fidelity (diagnostic)
4.30 / 5
4.14–4.46
Diagnostic
Does the recognized word sequence match what is heard? This is diagnostic, not part of the weighted composite.
Speech-timing fit (diagnostic)
3.42 / 5
3.28–3.56
Diagnostic
Do speech budgets and pause intervals fit the audio? This is diagnostic, separate from mechanical formatting validation.
Readability (diagnostic)
3.86 / 5
3.76–3.94
Diagnostic
Is the annotation understandable, concise and useful? This is diagnostic, not part of the weighted composite.
What the scores show. Global acoustic descriptions and talking style are stronger than local emotion specificity and event timing. Several fine-grained emotion labels are misplaced; some event intervals are redundant or include unrelated speech. All original outputs and reviewer notes remain visible below.
Single automated reviewer; small, reused synthetic listening set; no human adjudication. 95% intervals use 10,000 recording-level bootstrap resamples and exclude reviewer bias. Precision/coverage are ordinal ratings, not detection rates. New criteria/weights differ from earlier comparisons, so the composite is not a direct before/after number.
Formatting checks: all 50 pass. Every recognized word is retained once (2,703 total), all selected events are serialized once, and elapsed budgets agree. Acoustic accuracy is assessed separately. The 662 script segments derive from 443 source sentences and reuse source-sentence scores after splitting.
Estimated review list-price cost: $0.437 from returned usage. Annotation inference itself uses released local models and requires no Gemini API.
Listen and inspect the predictions
English examples come first, followed by German. The players contain the original synthetic acting audio. Voice name, source subset, timestamps and hashes remain available. You can seek through the timeline and inspect every local review.
A single young-to-mature adult masculine-coded speaker delivers an intensely intimate, whispered monologue in General American English. The vocal production is heavily unvoiced and breathy, characterized by low harmonicity, warm low-mid resonance, and close-mic proximity that borders on an ASMR or bedroom whisper style. The tempo is very slow and fragmented into short, cautious breath groups separated by frequent, audible inhalations and weary sighs. The emotional tone is saturated with deep affection, protective tenderness, and underlying distress or fear, as if trying to soothe a loved one in a dangerous or high-stakes situation. The recording is pristine with a dead-quiet background, allowing every subtle breath, lip parting, and vocal tremor to be clearly heard.
SCRIPT — timed speech, local cues and free-form burst descriptions
(very whispered, very fading, very having a crush, enjoyable; 2.72 seconds duration) ("A heavy, shuddering exhalation and breathy sigh precedes the opening line, establishing a state of acute distress and physical exhaustion.", 1.08 seconds) I won't let anything touch you.
[0.24 seconds pause]
(very non-narrative, choppy, very uncertainty; 1.12 seconds duration) Do you hear me?
[0.32 seconds pause]
(very darkening, very clean-content, very detachment; 0.43 seconds duration) Nothing.
[0.65 seconds pause]
(very full, very denasal, slightly schadenfreude, enjoyable; 2.47 seconds duration) I'll ("A quick, tense inhalation taken through the mouth between 'Nothing' and 'I'll tear the store down', showing physical bracing and elevated respiratory demand.", 0.66 seconds) tear this door down before they get close.
[0.21 seconds pause]
(very organic, very whispered, detachment; 0.03 seconds duration) Just
[0.85 seconds pause]
(very organic, very whispered, detachment; 0.87 seconds duration) stay behind
[0.25 seconds pause]
(very organic, very whispered, detachment; 0.56 seconds duration) me.
(very breathy, very denasal, very burnout, enjoyable; 2.44 seconds duration) ("A quick, tense inhalation taken through the mouth immediately before whispering 'We can hide in the crawlspace', reflecting heightened anxiety and physical breathlessness.", 0.66 seconds) We can hide in the crawl space or ("A sharp, constricted gasp of air taken in sudden panic right before whispering 'No, they already know.'", 0.80 seconds)
(very breathy, very denasal, very burnout, enjoyable; 2.28 seconds duration) ("A heavy, despondent sigh and breath release following the realization that the threat has been discovered, leading into a quiet, defeated whisper.", 1.00 seconds) No, they already know. ("A fragile, breathy whimper with a slight pitch tremor occurs in the pause before 'Please', signaling deep vulnerability and fear.", 0.68 seconds)
(very noisy, very breathy, very terror; 1.44 seconds duration) ("A quick, tense inhalation taken right before pleading 'Please', showing emotional vulnerability and physical breathlessness.", 0.60 seconds) Please ("A quick, breathy gasp of air taken right before pleading 'Please', showing desperate vulnerability.", 0.54 seconds)
(very noisy, very breathy, very terror; 1.12 seconds duration) take me instead.
(very meek, very asmr, saudade; 0.23 seconds duration) Just
[0.57 seconds pause]
(very meek, very asmr, saudade; 0.59 seconds duration) take me.
[2.21 seconds pause]
(very asmr, very organic, very trust; 0.90 seconds duration) It's okay, baby.
[0.54 seconds pause]
(very fading, very effortless, very comfort; 0.50 seconds duration) You're safe.
[0.94 seconds pause]
(very thin, non-instructional, very infatuation; 1.12 seconds duration) I've got you.
[0.32 seconds pause]
(very fading, very monotone, very boredom; 0.74 seconds duration) I promise.
[1.47 seconds pause]
Automated review note The global acoustic description accurately captures the close-mic ASMR-like whispered performance. However, segment-level emotion tags suffer from several bizarre choices (such as schadenfreude, burnout, saudade, and boredom) for what is clearly a frightened, urgent, protective monologue. Furthermore, several burst events have heavy overlapping timestamps with speech or each other.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
1.08 s duration · HVA_EVENT_0000
A heavy, shuddering exhalation and breathy sigh precedes the opening line, establishing a state of acute distress and physical exhaustion.
A quick, tense inhalation taken through the mouth between 'Nothing' and 'I'll tear the store down', showing physical bracing and elevated respiratory demand.
A quick, tense inhalation taken through the mouth immediately before whispering 'We can hide in the crawlspace', reflecting heightened anxiety and physical breathlessness.
Local emotion and delivery review — 17 speech segments
Absolute interval
Recognized speech and cue
Gemini ratings
Reviewer explanation
segment-001 Source sentence 1
I won't let anything touch you. very whispered, very fading, very having a crush, enjoyable
Emotion 2/5 Delivery 4/5
Tagged as 'very having a crush' which is a misleading label for protective distress.
segment-002 Source sentence 2
Do you hear me? very non-narrative, choppy, very uncertainty
Emotion 3/5 Delivery 3/5
Uncertainty fits reasonably well, but tone is urgent whisper.
segment-003 Source sentence 3
Nothing. very darkening, very clean-content, very detachment
Emotion 2/5 Delivery 3/5
Tagged as detachment, but it sounds tense and fearful rather than numb.
segment-004 Source sentence 4
I'll tear this door down before they get close. very full, very denasal, slightly schadenfreude, enjoyable
Emotion 1/5 Delivery 3/5
Schadenfreude is completely incorrect here; the line is protective and desperate.
segment-005 Source sentence 5
Just very organic, very whispered, detachment
Emotion 2/5 Delivery 3/5
Extremely short chunk; detachment tag does not fit.
segment-006 Source sentence 5
stay behind very organic, very whispered, detachment
Emotion 2/5 Delivery 3/5
Whispered and tense, not emotionally detached.
segment-007 Source sentence 5
me. very organic, very whispered, detachment
Emotion 2/5 Delivery 3/5
Tense whisper, not detached.
segment-008 Source sentence 6
We can hide in the crawl space or very breathy, very denasal, very burnout, enjoyable
Emotion 2/5 Delivery 3/5
Tagged as burnout; speaker is panicking and planning an escape.
segment-009 Source sentence 6
No, they already know. very breathy, very denasal, very burnout, enjoyable
Emotion 3/5 Delivery 3/5
Despairing defeat rather than general burnout.
segment-010 Source sentence 7
Please very noisy, very breathy, very terror
Emotion 4/5 Delivery 4/5
Terror/fear fits the desperate pleading.
segment-011 Source sentence 7
take me instead. very noisy, very breathy, very terror
Emotion 4/5 Delivery 4/5
Terror fits well.
segment-012 Source sentence 8
Just very meek, very asmr, saudade
Emotion 2/5 Delivery 3/5
Saudade does not fit the urgent pleading context.
segment-013 Source sentence 8
take me. very meek, very asmr, saudade
Emotion 2/5 Delivery 3/5
Desperate plea, not nostalgic longing (saudade).
segment-014 Source sentence 9
It's okay, baby. very asmr, very organic, very trust
Emotion 4/5 Delivery 4/5
Soothing, gentle reassurance fits well.
segment-015 Source sentence 10
You're safe. very fading, very effortless, very comfort
Emotion 4/5 Delivery 4/5
Comforting whisper fits well.
segment-016 Source sentence 11
I've got you. very thin, non-instructional, very infatuation
Emotion 2/5 Delivery 3/5
Tagged as infatuation; it is protective soothing.
segment-017 Source sentence 12
I promise. very fading, very monotone, very boredom
Emotion 1/5 Delivery 3/5
Boredom is entirely incorrect; it is a heartfelt whispered promise.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
4/5
Whole-recording emotion fit
4/5
Local speech emotion fit
2/5
Whole-recording talking-style fit
4/5
Local speech delivery fit
3/5
Burst-description fit
3/5
Burst-timing fit
2/5
Event precision rating
3/5
Event coverage rating
3/5
Transcript fidelity (diagnostic)
4/5
Speech-timing fit (diagnostic)
3/5
Readability (diagnostic)
4/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
53 words preserved; 8 selected events serialized; printed elapsed budget 28.130 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English6 selected intervals
The WeatherObsessed Neighbor
en-01167 · 15.74 s · Voice: Puck · laions_got_talent
2.85 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A young adult masculine-coded speaker delivers an engaging, conversational commentary in General American English about a fast-paced race. The vocal delivery is energetic, brisk, and articulate, characterized by forward oral resonance, clear diction, and moderate pitch variety. The tone conveys genuine interest, mild astonishment, and playful amusement, highlighted by a brief chuckle and a sharp preparatory inhale. The recording is studio-clean with minimal background noise and high intelligibility.
SCRIPT — timed speech, local cues and free-form burst descriptions
(very darkening, very dialogic, very mocking playfully, enjoyable; 0.94 seconds duration) Did you catch that bare
[0.26 seconds pause]
(very darkening, very dialogic, very mocking playfully, enjoyable; 0.95 seconds duration) metric drop around four?
[0.33 seconds pause]
(very clean-bg, very hi-fi, very shock, enjoyable; 2.64 seconds duration) The isobar spacing on the radar was tightening up
[0.24 seconds pause]
(very clean-bg, very hi-fi, very shock, enjoyable; 0.46 seconds duration) fast.
(fast, long-phrased, very suspicion, enjoyable; 2.30 seconds duration) ("A brief, low-pitched chuckle occurs in the pause between 'fast' and 'You probably just thought', featuring two quiet, breathy pulses that convey dry amusement and skepticism before smoothly transitioning into the next sentence.", 0.54 seconds) You probably just thought it was a passing breeze, right? ("A brief, low-pitched, breathy chuckle occurs right after 'right?', expressing cynical amusement and mild disbelief at the dismissive attitude of colleagues. It blends smoothly into the subsequent sentence without disrupting the conversational flow.", 0.46 seconds)
(very long-phrased, very clean-bg, very awkwardness, enjoyable; 3.92 seconds duration) ("A brief, low-pitched, breathy chuckle occurs in the pause between 'right?' and 'Everyone on the coldest', conveying dry, cynical amusement at the dismissive attitude of colleagues.", 0.58 seconds) Everyone ("A brief, quiet, low-pitched chuckle occurs in the pause between 'right?' and 'Everyone on the coldest', conveying dry amusement and mild cynicism about colleagues' dismissive reactions.", 0.54 seconds) on the cul de sac gave me that polite nod when I brought up the dew point.
(clean-bg, accelerating, very sourness, enjoyable; 3.47 seconds duration) ("A deliberate, audible inhalation occurs in the pause before delivering the final warning, grounding the speaker's resolve and transitioning smoothly into the firm, assertive delivery of 'But mark my words'.", 0.56 seconds) But ("A quick, audible preparatory intake of breath occurs right before the speaker delivers the emphatic warning 'But mark my words', naturally fueling the subsequent assertive phrase.", 0.46 seconds) mark my words, that cold front isn't messing around.
[0.23 seconds pause]
Automated review note Burst events are excessively duplicated with overlapping windows for single events.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
0.54 s duration · HVA_EVENT_0000
A brief, low-pitched chuckle occurs in the pause between 'fast' and 'You probably just thought', featuring two quiet, breathy pulses that convey dry amusement and skepticism before smoothly transitioning into the next sentence.
Top-three type softmax: Sharp Inhale: 34.5%; Chuckle: 17.2%; Breath: 8.2%
0.46 s duration · HVA_EVENT_0001
A brief, low-pitched, breathy chuckle occurs right after 'right?', expressing cynical amusement and mild disbelief at the dismissive attitude of colleagues. It blends smoothly into the subsequent sentence without disrupting the conversational flow.
A brief, low-pitched, breathy chuckle occurs in the pause between 'right?' and 'Everyone on the coldest', conveying dry, cynical amusement at the dismissive attitude of colleagues.
A brief, quiet, low-pitched chuckle occurs in the pause between 'right?' and 'Everyone on the coldest', conveying dry amusement and mild cynicism about colleagues' dismissive reactions.
A deliberate, audible inhalation occurs in the pause before delivering the final warning, grounding the speaker's resolve and transitioning smoothly into the firm, assertive delivery of 'But mark my words'.
A quick, audible preparatory intake of breath occurs right before the speaker delivers the emphatic warning 'But mark my words', naturally fueling the subsequent assertive phrase.
Local emotion and delivery review — 7 speech segments
Absolute interval
Recognized speech and cue
Gemini ratings
Reviewer explanation
segment-001 Source sentence 1
Did you catch that bare very darkening, very dialogic, very mocking playfully, enjoyable
Emotion 3/5 Delivery 3/5
Tone is moderately playful, split sentence is awkward.
segment-002 Source sentence 1
metric drop around four? very darkening, very dialogic, very mocking playfully, enjoyable
Emotion 3/5 Delivery 3/5
Labels are moderately aligned with the conversational delivery.
segment-003 Source sentence 2
The isobar spacing on the radar was tightening up very clean-bg, very hi-fi, very shock, enjoyable
Emotion 3/5 Delivery 3/5
Shock label slightly overstates mild engagement.
segment-004 Source sentence 2
fast. very clean-bg, very hi-fi, very shock, enjoyable
Emotion 3/5 Delivery 3/5
Isolated word marked with shock is excessive.
segment-005 Source sentence 3
You probably just thought it was a passing breeze, right? fast, long-phrased, very suspicion, enjoyable
Emotion 3/5 Delivery 3/5
Suspicion and tempo descriptions are partially supported.
segment-006 Source sentence 4
Everyone on the cul de sac gave me that polite nod when I brought up the dew point. very long-phrased, very clean-bg, very awkwardness, enjoyable
Emotion 3/5 Delivery 3/5
Awkwardness label moderately captures the narrative tone.
segment-007 Source sentence 5
But mark my words, that cold front isn't messing around. clean-bg, accelerating, very sourness, enjoyable
Emotion 3/5 Delivery 3/5
Sourness is slightly exaggerated for the emphatic finish.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
3/5
Whole-recording emotion fit
3/5
Local speech emotion fit
3/5
Whole-recording talking-style fit
3/5
Local speech delivery fit
3/5
Burst-description fit
3/5
Burst-timing fit
2/5
Event precision rating
2/5
Event coverage rating
3/5
Transcript fidelity (diagnostic)
3/5
Speech-timing fit (diagnostic)
3/5
Readability (diagnostic)
3/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
57 words preserved; 6 selected events serialized; printed elapsed budget 15.740 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English11 selected intervals
The Hollow Release
en-01199 · 33.26 s · Voice: Zephyr · situation_inspired
3.45 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A young adult feminine-coded speaker delivers an intimate, deeply fatigued, and contemplative monologue in General American English. The vocal delivery is exceptionally slow, quiet, and breathy, hovering between a half-whisper and soft-spoken speech with low pitch range and subdued dynamic energy. The timbre is warm yet thin, dominated by unvoiced air and pharyngeal resonance with minimal chest support. The speaker exhibits high vulnerability and inward focus, conveying profound physical exhaustion, numbness, and quiet resignation. Multiple audible sighs, weary groans, and labored breaths punctuate the fragmented phrasing, blending seamlessly into the hushed, close-mic acoustic environment.
SCRIPT — timed speech, local cues and free-form burst descriptions
[0.04 seconds pause]
("A sharp, audible intake of breath through the mouth and nose occurs right at the start of the recording before any speech begins, signaling physical exertion or preparation.", 0.90 seconds)
[0.28 seconds pause]
("A heavy, weary exhalation with slight vocalization that leads directly into the muttered expletive 'God', conveying deep fatigue and exasperation.", 0.92 seconds)
(very fading, very mumbled, very solace, less-enjoyable; 0.49 seconds duration) God.
[1.05 seconds pause]
(very whispered, very hi-fi, very lethargy, enjoyable; 1.99 seconds duration) Your body just keeps firing off,
[0.25 seconds pause]
(very whispered, very hi-fi, very lethargy, enjoyable; 1.18 seconds duration) doesn't it? ("A quick, audible gasp-like inhalation through the mouth between phrases, reflecting physical breathlessness and fatigue.", 0.56 seconds)
[0.26 seconds pause]
(very asmr, very whispered, very lust, enjoyable; 1.79 seconds duration) Twitches right on cue,
[0.29 seconds pause]
(very asmr, very whispered, very lust, enjoyable; 1.16 seconds duration) chest heaving like
[0.20 seconds pause]
(very asmr, very whispered, very lust, enjoyable; 1.58 seconds duration) it actually mattered.
(overlapping non-speech vocal events; 1.14 seconds duration) ("A brief, quiet, breathy chuckle occurs in the pause after 'like it actually mattered', expressing wry amusement at the situation.", 0.78 seconds) ("A brief, quiet breathy chuckle occurs in the pause before 'He didn't even notice', conveying mild bemused disbelief.", 0.72 seconds)
[0.56 seconds pause]
(hi-fi, oral-bright, very detachment, enjoyable; 1.20 seconds duration) He didn't even notice.
[0.56 seconds pause]
(very asmr, very thin, very apathy; 1.60 seconds duration) Just rolled right over.
[0.22 seconds pause]
(very fading, very whispered, very sympathy; 3.41 seconds duration) ("A soft, quiet breathy chuckle with a slight nasal release, expressing wry amusement at the situation.", 0.96 seconds) ("A slow, audible inhalation and soft exhalation marking a pause before continuing the quiet, intimate observation.", 0.98 seconds) ("A soft, breathy sigh of weary resignation and mild dark amusement before remarking that the situation is almost funny.", 0.70 seconds) It ("A soft, quiet breathy sigh of resignation and mild bemusement right before uttering 'It is almost funny'.", 0.74 seconds) is almost funny.
[0.37 seconds pause]
("A quiet, breathy sigh of resignation and mild bewilderment before continuing the observation.", 0.76 seconds)
(very whispered, very breathy, very weariness, enjoyable; 1.08 seconds duration) Brain is
[0.24 seconds pause]
(very whispered, very breathy, very weariness, enjoyable; 1.60 seconds duration) completely turned off,
[0.40 seconds pause]
(very whispered, very breathy, very weariness, enjoyable; 1.05 seconds duration) totally vacant,
[0.55 seconds pause]
(very whispered, very breathy, very weariness, enjoyable; 2.16 seconds duration) but the machinery just ("A quiet, breathy sighing release between 'just' and 'still goes through the motions', conveying a mix of exhaustion and detached fascination.", 0.56 seconds)
[0.24 seconds pause]
(very whispered, very breathy, very weariness, enjoyable; 1.68 seconds duration) still goes through the motions.
[0.64 seconds pause]
(very accelerating, very asmr, very indifference, enjoyable; 1.19 seconds duration) Even when nothing's there.
[0.23 seconds pause]
Automated review note The global description is strong, but several segments are inappropriately tagged with 'lust'. Additionally, the burst events feature heavily redundant overlapping duplicates (e.g. at 12-13.5s and 17.6-20.2s).
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
0.90 s duration · HVA_EVENT_0000
A sharp, audible intake of breath through the mouth and nose occurs right at the start of the recording before any speech begins, signaling physical exertion or preparation.
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
51 words preserved; 11 selected events serialized; printed elapsed budget 33.260 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English9 selected intervals
Ecstasy and Betrayal in the Same Breath
en-01608 · 26.66 s · Voice: Autonoe · eric_morris_inspired
3.75 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A young adult feminine-coded speaker delivers an intensely intimate, breathy, and emotionally vulnerable confession in General American English. The vocal delivery is predominantly whispered and half-voiced, characterized by high breathiness, low harmonicity, and prominent pharyngeal and head resonance. The tempo is slow and fragmented into short, breathless chunks separated by frequent audible gasps, sighs, and tremulous whimpers. The emotional tone is saturated with intense longing, infatuation, and romantic distress, shifting between tender affection and helpless heartache. The recording is pristine with a dead-quiet background, capturing every subtle lip parting, breath intake, and vocal tremor with high fidelity.
SCRIPT — timed speech, local cues and free-form burst descriptions
(very whispered, very vulnerable, very panic; 0.40 seconds duration) God,
(very whispered, very vulnerable, very panic; 2.00 seconds duration) ("A sharp, breathy intake of air occurs right at the start before speaking, signaling physical trembling and emotional vulnerability.", 0.88 seconds) my hands won't stop shaking.
(very breathy, very whispered, very fondness, enjoyable; 1.68 seconds duration) ("A soft, breathy exhalation occurs in the pause following 'shaking', conveying physical vulnerability and emotional overwhelm before the next phrase begins.", 1.10 seconds) It feels like
(very breathy, very whispered, very fondness, enjoyable; 3.14 seconds duration) ("A sharp, audible intake of air punctuates the pause between 'It feels like' and 'pure electricity', conveying physical breathlessness and heightened arousal.", 0.88 seconds) pure electricity just having you right here against
(very breathy, very whispered, very fondness, enjoyable; 1.14 seconds duration) me. ("A soft, breathy exhalation occurs in the pause after 'against me', conveying emotional vulnerability and physical longing.", 0.98 seconds)
(very breathy, very vulnerable, very powerlessness, enjoyable; 1.68 seconds duration) ("A soft, breathy sigh of longing and emotional release occurs in the pause after 'against me', blending smoothly into the subsequent phrase.", 1.26 seconds) I was ready to laugh
[0.28 seconds pause]
(very breathy, very vulnerable, very powerlessness, enjoyable; 1.67 seconds duration) to just ("A quick, breathy gasp-like inhalation interrupts the phrase between 'to just' and 'hold you', conveying emotional vulnerability and physical breathlessness.", 0.74 seconds) hold you
(very breathy, very vulnerable, very powerlessness, enjoyable; 1.30 seconds duration) and ("A quick, breathy gasp-like inhalation occurs right before 'and never let go', conveying emotional vulnerability and physical breathlessness.", 0.50 seconds) never let go.
[0.61 seconds pause]
(very mumbled, very scattered, romantic desire; 1.51 seconds duration) ("A soft, tremulous breathy sigh and release of air following 'and never let go', conveying bittersweet longing and emotional vulnerability.", 1.00 seconds) But you knew.
[0.83 seconds pause]
(very tense, very swelling, very cynicism; 3.37 seconds duration) The whole time you were kissing me, you already knew what you'd done to
[0.23 seconds pause]
(very tense, very swelling, very cynicism; 0.04 seconds duration) us.
[1.44 seconds pause]
("A heavy, despondent sigh and shuddering breath release following the sharp inhale, conveying deep emotional ache and helplessness.", 1.14 seconds)
[0.74 seconds pause]
(very vulnerable, very darkening, very wistfulness, enjoyable; 2.41 seconds duration) How can my chest hurt this much when I still
[0.23 seconds pause]
(very vulnerable, very darkening, very wistfulness, enjoyable; 0.40 seconds duration) want you?
[0.42 seconds pause]
Automated review note Strong overall description of vocal quality and emotional arc. However, some nonverbal burst event boundaries overlap significantly (e.g. EVENT_0003 and EVENT_0004 cover the same exhalation after 'me'), and EVENT_0000 placed an intake right over the spoken word 'God'.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
0.88 s duration · HVA_EVENT_0000
A sharp, breathy intake of air occurs right at the start before speaking, signaling physical trembling and emotional vulnerability.
A soft, breathy exhalation occurs in the pause following 'shaking', conveying physical vulnerability and emotional overwhelm before the next phrase begins.
A sharp, audible intake of air punctuates the pause between 'It feels like' and 'pure electricity', conveying physical breathlessness and heightened arousal.
A quick, breathy gasp-like inhalation interrupts the phrase between 'to just' and 'hold you', conveying emotional vulnerability and physical breathlessness.
Local emotion and delivery review — 13 speech segments
Absolute interval
Recognized speech and cue
Gemini ratings
Reviewer explanation
segment-001 Source sentence 1
God, very whispered, very vulnerable, very panic
Emotion 4/5 Delivery 4/5
Accurately identifies the whispered, vulnerable opening.
segment-002 Source sentence 1
my hands won't stop shaking. very whispered, very vulnerable, very panic
Emotion 4/5 Delivery 4/5
Good fit for the shaky whispered delivery.
segment-003 Source sentence 2
It feels like very breathy, very whispered, very fondness, enjoyable
Emotion 4/5 Delivery 4/5
Well described breathy tone and longing.
segment-004 Source sentence 2
pure electricity just having you right here against very breathy, very whispered, very fondness, enjoyable
Emotion 4/5 Delivery 4/5
Appropriate description of breathy infatuation.
segment-005 Source sentence 2
me. very breathy, very whispered, very fondness, enjoyable
Emotion 4/5 Delivery 4/5
Delivery matches the trailing off whisper.
segment-006 Source sentence 3
I was ready to laugh very breathy, very vulnerable, very powerlessness, enjoyable
Emotion 4/5 Delivery 4/5
Captures the emotional vulnerability accurately.
segment-007 Source sentence 3
to just hold you very breathy, very vulnerable, very powerlessness, enjoyable
Emotion 4/5 Delivery 4/5
Fits the breathy pause and continuation.
segment-008 Source sentence 3
and never let go. very breathy, very vulnerable, very powerlessness, enjoyable
Emotion 4/5 Delivery 4/5
Well characterized breathy delivery.
segment-009 Source sentence 4
But you knew. very mumbled, very scattered, romantic desire
Emotion 3/5 Delivery 3/5
Tagged as mumbled/scattered, but it is clear whispered speech with sorrowful tone.
segment-010 Source sentence 5
The whole time you were kissing me, you already knew what you'd done to very tense, very swelling, very cynicism
Emotion 4/5 Delivery 4/5
Tension and bitter disappointment are audible.
segment-011 Source sentence 5
us. very tense, very swelling, very cynicism
Emotion 4/5 Delivery 4/5
Short final word maintains the tense emotion.
segment-012 Source sentence 6
How can my chest hurt this much when I still very vulnerable, very darkening, very wistfulness, enjoyable
Emotion 4/5 Delivery 4/5
Wistful, hurt, and vulnerable tone is accurately identified.
segment-013 Source sentence 6
want you? very vulnerable, very darkening, very wistfulness, enjoyable
Emotion 4/5 Delivery 4/5
Quiet fading question matches the wistful tone.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
4/5
Whole-recording emotion fit
4/5
Local speech emotion fit
4/5
Whole-recording talking-style fit
4/5
Local speech delivery fit
4/5
Burst-description fit
3/5
Burst-timing fit
3/5
Event precision rating
3/5
Event coverage rating
4/5
Transcript fidelity (diagnostic)
5/5
Speech-timing fit (diagnostic)
4/5
Readability (diagnostic)
4/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
61 words preserved; 9 selected events serialized; printed elapsed budget 26.660 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English6 selected intervals
Dom/Sub Scene: Establishing the Dynamic
en-01865 · 36.65 s · Voice: Fenrir · voicenet_extension
3.20 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A single young-adult to mature-adult masculine speaker delivers an intensely intimate, whispered monologue in General American English. The vocal delivery is characterized by an extremely low pitch register, heavy chest and pharyngeal resonance, and pervasive breathiness that borders on an ASMR or close-mic bedroom style. The tempo is very slow and fragmented into short, deliberate phrases separated by long, quiet pauses. Articulation is soft yet clear, with gentle onsets and low muscular tension. The emotional tone is deeply affectionate, soothing, and protective, projecting calm authority and focused intimacy. The recording is pristine with a dead-quiet background and close proximity effect, capturing subtle respiratory details and natural vocal fry.
SCRIPT — timed speech, local cues and free-form burst descriptions
[0.24 seconds pause]
(scattered, choppy, detachment; 0.49 seconds duration) Kneel
[0.79 seconds pause]
(scattered, choppy, detachment; 0.23 seconds duration) here.
[1.85 seconds pause]
(very denasal, accelerating, very sexual lust, enjoyable; 1.44 seconds duration) Hands behind your back
[0.56 seconds pause]
(very denasal, accelerating, very sexual lust, enjoyable; 0.96 seconds duration) eyes on me.
[1.52 seconds pause]
(very choppy, smooth, very liberation; 0.55 seconds duration) Yes.
[0.25 seconds pause]
(very fading, scattered, very insensitivity; 2.58 seconds duration) Just like that. ("A slow, audible, close-mic inhalation and exhalation cycle occurs during a pause after 'just like that', reflecting physical exertion and controlled pacing.", 1.10 seconds) ("A slow, audible, close-mic inhalation and exhalation cycle occurs during a pause in speech, reflecting physical exertion and controlled pacing.", 1.06 seconds)
[0.54 seconds pause]
("A slow, audible nasal and oral inhalation occurs during a pause in the whispered commands, reflecting physical exertion and close-mic proximity.", 0.76 seconds)
[0.92 seconds pause]
(very dissociated, very non-narrative, numbness; 0.48 seconds duration) Good.
[1.52 seconds pause]
(very denasal, very naturalistic, very butterflies in the stomach; 1.16 seconds duration) You're doing perfectly.
[0.32 seconds pause]
("A quick, quiet sharp inhale through the mouth occurs right before the speaker interrupts with 'Wait', signaling a sudden shift in focus and physical preparation.", 0.60 seconds)
(organic, non-instructional, shyness; 0.42 seconds duration) Wait
[0.54 seconds pause]
(organic, non-instructional, shyness; 0.32 seconds duration) hey
[0.64 seconds pause]
(organic, non-instructional, shyness; 2.08 seconds duration) looking at me for real a second.
[0.40 seconds pause]
(very naturalistic, very hi-fi, very clemency, enjoyable; 1.27 seconds duration) You still good with this pace?
[0.38 seconds pause]
(very fading, very choppy, slightly existential void; 1.89 seconds duration) ("A quiet, controlled nasal and oral inhalation in the pause before confirming 'Yes', maintaining the close-mic, intimate, and focused atmosphere.", 0.90 seconds) ("A quiet, controlled nasal and oral inhalation taken during a pause before asking 'Yes?', maintaining the close-mic, intimate atmosphere.", 0.80 seconds) Yes?
[1.50 seconds pause]
(very fading, very explicit, very distrust, enjoyable; 0.53 seconds duration) Truly?
[2.03 seconds pause]
(very non-narrative, very scattered, very indifference, less-enjoyable; 0.24 seconds duration) Okay.
[1.36 seconds pause]
(very accelerating, very fading, very emotional blunting; 0.96 seconds duration) Then shin up.
[0.80 seconds pause]
(very steady, very hi-fi, slightly indifference, enjoyable; 1.28 seconds duration) Breathe for me.
[0.32 seconds pause]
(very naturalistic, very hi-fi, slightly caring, enjoyable; 1.51 seconds duration) And let me take control from here.
[0.42 seconds pause]
Automated review note A slow intimate whisper performance. Several transcript errors ('shin up' for 'chin up', 'looking at me for real a second' for 'look at me for a second'). Several segments have duplicative, heavily overlapping burst event intervals (e.g. around 10-11s and 24-25s), and local emotion tags repeatedly default to 'emotional numbness/detachment/indifference' which conflicts with the intimate, sensual context.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
1.10 s duration · HVA_EVENT_0000
A slow, audible, close-mic inhalation and exhalation cycle occurs during a pause after 'just like that', reflecting physical exertion and controlled pacing.
A quick, quiet sharp inhale through the mouth occurs right before the speaker interrupts with 'Wait', signaling a sudden shift in focus and physical preparation.
eyes on me. very denasal, accelerating, very sexual lust, enjoyable
Emotion 4/5 Delivery 3/5
Soft whispered direction, intimate.
segment-005 Source sentence 3
Yes. very choppy, smooth, very liberation
Emotion 2/5 Delivery 3/5
'Very liberation' overinterprets a simple quiet affirmation.
segment-006 Source sentence 4
Just like that. very fading, scattered, very insensitivity
Emotion 1/5 Delivery 2/5
The phrase ends around 10.0s, but the segment timeline is stretched to 11.46s encompassing pauses and breaths. 'Insensitivity' contradicts the affectionate whispered intimacy.
segment-007 Source sentence 5
Good. very dissociated, very non-narrative, numbness
Emotion 2/5 Delivery 2/5
Intimate praise, not 'numbness' or 'dissociated'.
segment-008 Source sentence 6
You're doing perfectly. very denasal, very naturalistic, very butterflies in the stomach
Emotion 3/5 Delivery 4/5
Warm whispered praise; 'butterflies in the stomach' is slightly exaggerated.
segment-009 Source sentence 7
Wait organic, non-instructional, shyness
Emotion 2/5 Delivery 3/5
Attention shift rather than 'shyness'.
segment-010 Source sentence 7
hey organic, non-instructional, shyness
Emotion 2/5 Delivery 3/5
Gentle call, not embarrassed or shy.
segment-011 Source sentence 7
looking at me for real a second. organic, non-instructional, shyness
Emotion 2/5 Delivery 2/5
Transcript error: speaker says 'look at me for a second', not 'looking at me for real a second'.
segment-012 Source sentence 8
You still good with this pace? very naturalistic, very hi-fi, very clemency, enjoyable
Emotion 4/5 Delivery 4/5
Attentive, checking in; fits gentle caring tone.
segment-013 Source sentence 9
Yes? very fading, very choppy, slightly existential void
Emotion 2/5 Delivery 2/5
Speech is just 'Yes?' around 25.3s to 25.7s; segment spans breath intervals. 'Existential void' is inappropriate.
segment-014 Source sentence 10
Truly? very fading, very explicit, very distrust, enjoyable
Emotion 2/5 Delivery 3/5
Quiet check-in rather than genuine 'distrust'.
segment-015 Source sentence 11
Okay. very non-narrative, very scattered, very indifference, less-enjoyable
Emotion 2/5 Delivery 3/5
Quiet consent/acknowledgment, not indifference.
segment-016 Source sentence 12
Then shin up. very accelerating, very fading, very emotional blunting
Emotion 2/5 Delivery 3/5
Transcript error: speaker says 'chin up', not 'shin up'. 'Emotional blunting' does not match.
segment-017 Source sentence 13
Breathe for me. very steady, very hi-fi, slightly indifference, enjoyable
Emotion 2/5 Delivery 3/5
Tender guidance, not 'indifference'.
segment-018 Source sentence 14
And let me take control from here. very naturalistic, very hi-fi, slightly caring, enjoyable
Emotion 4/5 Delivery 4/5
Affectionate/caring fits the intimate whisper well.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
4/5
Whole-recording emotion fit
4/5
Local speech emotion fit
2/5
Whole-recording talking-style fit
4/5
Local speech delivery fit
3/5
Burst-description fit
3/5
Burst-timing fit
3/5
Event precision rating
3/5
Event coverage rating
3/5
Transcript fidelity (diagnostic)
3/5
Speech-timing fit (diagnostic)
3/5
Readability (diagnostic)
3/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
48 words preserved; 6 selected events serialized; printed elapsed budget 36.650 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English10 selected intervals
Laughing at Your Own Memorial Service (You Survived When No One Thought You Would)
en-02220 · 36.53 s · Voice: Leda · existing_challenge_inspired
3.15 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A young adult feminine-coded speaker delivers an intensely intimate, whispered monologue in General American English. The performance is dominated by an ASMR-like, unvoiced whisper register with very low harmonicity, high breathiness, and close-mic proximity. The tempo is very slow and heavily fragmented into short, cautious breath groups separated by frequent pauses and audible respiratory events. The emotional tone is deeply vulnerable, shifting from initial fear and distress to a fragile, trembling sense of relief and quiet hope. The recording is pristine with a dead-quiet background, allowing subtle mouth sounds, breath catches, and delicate sighs to blend seamlessly into the spoken delivery.
SCRIPT — timed speech, local cues and free-form burst descriptions
(very fading, very soft-onset, very carnal desire; 2.42 seconds duration) Listen ("A sharp, audible intake of breath precedes the opening whispered phrase, setting an immediate tone of tension and vulnerability.", 0.60 seconds) to them. ("A sharp, audible intake of air following 'Listen to them', marked by throat constriction and breathiness, signaling emotional distress and physical tension.", 0.52 seconds) ("A sharp, audible intake of air through the mouth and nose occurs in the pause between 'Listen to them' and 'steady', conveying physical tension and emotional distress.", 0.56 seconds)
[0.30 seconds pause]
(very meek, very explicit, stoicism, less-enjoyable; 0.63 seconds duration) Steady.
[1.13 seconds pause]
(very asmr, very whispered, very carnal desire; 1.00 seconds duration) Fearless.
("A soft, breathy exhalation occurs right after the word 'fearless', conveying bitter irony and emotional exhaustion before continuing the thought.", 0.62 seconds)
[0.78 seconds pause]
(very whispered, very breathy, very disappointment, enjoyable; 3.74 seconds duration) God, they really bought the whole act, didn't they? ("A soft, tremulous breathy sigh and quiet exhale following 'didn't they?', conveying emotional exhaustion and lingering trauma.", 0.76 seconds) ("A soft, tremulous breathy sigh and quiet whimper-like release occurs in the pause after 'didn't they?', conveying deep vulnerability and emotional exhaustion.", 0.86 seconds)
(very explicit, very whispered, very carnal desire, enjoyable; 1.93 seconds duration) They're ("A quick, quiet gasp-like inhalation occurs in the pause before 'They're burying a statue', reflecting emotional tension and breathlessness.", 0.56 seconds) burying a statue,
[0.25 seconds pause]
(very explicit, very whispered, very carnal desire, enjoyable; 0.38 seconds duration) not me.
[1.78 seconds pause]
(very whispered, very vulnerable, very compassion, enjoyable; 0.56 seconds duration) But then
[1.28 seconds pause]
(very whispered, very vulnerable, very compassion, enjoyable; 3.52 seconds duration) hearing my kid up there talking about how strong I was
[0.33 seconds pause]
("A soft, tremulous exhalation and quiet sob-like release following 'how strong I was', conveying deep vulnerability and emotional exhaustion.", 1.68 seconds)
[0.39 seconds pause]
(very whispered, very vulnerable, very compassion, enjoyable; 2.33 seconds duration) Sweetheart, I was so terrified.
("A soft, tremulous breath intake and release between 'terrified' and 'I'm sorry', conveying lingering vulnerability and emotional exposure.", 0.90 seconds)
[0.53 seconds pause]
(very whispered, very decelerating, very discomfiture; 1.68 seconds duration) I'm sorry I hid it.
[1.52 seconds pause]
(very thin, very decelerating, alleviation; 0.73 seconds duration) I'm here.
[1.03 seconds pause]
(very whispered, very decelerating, very comfort; 0.85 seconds duration) I'm alive.
("A soft, deliberate intake of air between 'I'm alive' and 'and I'm finally ready', grounding the final vulnerable confession.", 0.68 seconds)
[0.71 seconds pause]
(very asmr, very organic, very alleviation, enjoyable; 2.55 seconds duration) And I'm finally ready to be seen.
[0.30 seconds pause]
Automated review note Several segments (001, 003, 005, 006) are inappropriately tagged with 'carnal desire' / sexual lust, which severely contradicts the solemn, emotionally vulnerable grief narrative. Burst events also suffer from overlapping duplicate intervals around 1.5-2.4s and 9.3-10.6s.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
0.60 s duration · HVA_EVENT_0000
A sharp, audible intake of breath precedes the opening whispered phrase, setting an immediate tone of tension and vulnerability.
A sharp, audible intake of air following 'Listen to them', marked by throat constriction and breathiness, signaling emotional distress and physical tension.
A sharp, audible intake of air through the mouth and nose occurs in the pause between 'Listen to them' and 'steady', conveying physical tension and emotional distress.
A soft, tremulous breathy sigh and quiet whimper-like release occurs in the pause after 'didn't they?', conveying deep vulnerability and emotional exhaustion.
Local emotion and delivery review — 13 speech segments
Absolute interval
Recognized speech and cue
Gemini ratings
Reviewer explanation
segment-001 Source sentence 1
Listen to them. very fading, very soft-onset, very carnal desire
Emotion 1/5 Delivery 3/5
Tagged with 'carnal desire' which is completely inaccurate for this somber, grief-stricken whisper.
segment-002 Source sentence 2
Steady. very meek, very explicit, stoicism, less-enjoyable
Emotion 3/5 Delivery 3/5
Tone is subdued and whispered; 'explicit' tag is unwarranted.
segment-003 Source sentence 3
Fearless. very asmr, very whispered, very carnal desire
Emotion 1/5 Delivery 4/5
Again incorrectly labeled with 'carnal desire'.
segment-004 Source sentence 4
God, they really bought the whole act, didn't they? very whispered, very breathy, very disappointment, enjoyable
Emotion 4/5 Delivery 4/5
Disappointment and whispered style match the audible bitter delivery.
segment-005 Source sentence 5
They're burying a statue, very explicit, very whispered, very carnal desire, enjoyable
Emotion 1/5 Delivery 3/5
Erroneously tagged with 'carnal desire' and 'explicit'.
segment-006 Source sentence 5
not me. very explicit, very whispered, very carnal desire, enjoyable
Emotion 1/5 Delivery 3/5
Continuation of previous phrase, similarly mislabeled with sexual lust/carnal desire.
segment-007 Source sentence 6
But then very whispered, very vulnerable, very compassion, enjoyable
Emotion 3/5 Delivery 4/5
Vulnerable whispered delivery is accurate, though affection/compassion is questionable here.
segment-008 Source sentence 6
hearing my kid up there talking about how strong I was very whispered, very vulnerable, very compassion, enjoyable
Emotion 4/5 Delivery 4/5
Vulnerable and sorrowful whisper fits well.
segment-009 Source sentence 6
Sweetheart, I was so terrified. very whispered, very vulnerable, very compassion, enjoyable
Emotion 4/5 Delivery 4/5
Vulnerable and tender confession matches well.
segment-010 Source sentence 7
I'm sorry I hid it. very whispered, very decelerating, very discomfiture
Emotion 3/5 Delivery 4/5
Apologetic, regretful whisper fits well.
segment-011 Source sentence 8
I'm here. very thin, very decelerating, alleviation
Emotion 4/5 Delivery 4/5
Quiet relief matches the spoken words.
segment-012 Source sentence 9
I'm alive. very whispered, very decelerating, very comfort
Emotion 4/5 Delivery 4/5
Soft relief and comfort fit this affirmation.
segment-013 Source sentence 10
And I'm finally ready to be seen. very asmr, very organic, very alleviation, enjoyable
Emotion 4/5 Delivery 4/5
Intimate ASMR-like whispered tone conveying quiet relief and resolution.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
4/5
Whole-recording emotion fit
4/5
Local speech emotion fit
2/5
Whole-recording talking-style fit
4/5
Local speech delivery fit
3/5
Burst-description fit
3/5
Burst-timing fit
2/5
Event precision rating
3/5
Event coverage rating
4/5
Transcript fidelity (diagnostic)
5/5
Speech-timing fit (diagnostic)
3/5
Readability (diagnostic)
4/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
54 words preserved; 10 selected events serialized; printed elapsed budget 36.530 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English9 selected intervals
Setting Boundary With Family Member
en-02261 · 23.09 s · Voice: Kore · haiku_batch
3.15 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A feminine-coded, young-adult speaker delivers an emotionally charged, confrontational monologue in General American English. The performance is marked by intense distress, bitterness, and exasperation, characterized by a tense, slightly breathy vocal quality with prominent oral and pharyngeal resonance. The delivery is choppy and fragmented, featuring frequent pauses, sharp inhalations, and an exasperated sigh that reflect high emotional volatility and cognitive strain. The speaker's stance is assertive and direct, projecting sharp interpersonal focus and mounting frustration as she demands an end to deception. The recording is studio-clean with minimal background noise and sharp transient clarity.
SCRIPT — timed speech, local cues and free-form burst descriptions
(clean-bg, non-instructional, very feeling horny, enjoyable; 2.90 seconds duration) ("A sharp, audible intake of breath precedes the opening line, setting an immediate tone of emotional tension and preparation before speaking.", 0.46 seconds) I need you to stop talking to me like
[0.30 seconds pause]
(clean-bg, non-instructional, very feeling horny, enjoyable; 0.32 seconds duration) that.
[0.88 seconds pause]
(very bright, very tonal, enjoyable; 0.23 seconds duration) No,
[0.41 seconds pause]
(very bright, very tonal, enjoyable; 1.50 seconds duration) listen to me for once.
[0.34 seconds pause]
("A sharp, audible intake of air through the mouth and nose occurs during a tense pause between 'for once' and 'This is not negotiable anymore', reflecting rising emotional distress and physical preparation for the next forceful statement.", 0.52 seconds)
[0.20 seconds pause]
(very clean-bg, very tonal, very dismay, enjoyable; 1.92 seconds duration) This is not negotiable anymore.
(overlapping non-speech vocal events; 1.18 seconds duration) ("A sharp, audible intake of air following 'This is not negotiable anymore', marking a tense respiratory reset before delivering the next confrontational line.", 0.76 seconds) ("A sharp, audible intake of air through the mouth and nose occurs during a tense pause before the speaker delivers the resigned line 'Go ahead then', reflecting emotional bracing and physical tension.", 0.62 seconds)
(very tonal, very non-narrative, fervor; 1.14 seconds duration) Go ahead, then.
(very bright, very oral-bright, very irritability, enjoyable; 2.19 seconds duration) ("A quick, audible gasp-like inhalation occurs in the brief pause before 'Cut me off', reflecting emotional bracing and physical tension.", 0.44 seconds) Cut me off if that's what you have to do,
[0.37 seconds pause]
(very bright, very oral-bright, very irritability, enjoyable; 0.90 seconds duration) okay? ("A sharp, audible intake of air following the confrontational question 'okay?', marking a brief pause before the speaker drops into a quieter, more vulnerable register.", 0.64 seconds)
(very dissociated, very mumbled, very numbness; 0.94 seconds duration) ("A soft, tremulous breathy release and quiet sob-like catch in the throat occurs right after the word 'okay', marking a sudden collapse of composure and emotional vulnerability.", 0.70 seconds) Okay.
[0.42 seconds pause]
("A soft, tremulous breathy release and quiet sob-like exhalation occurs in the pause after the whispered 'Okay', conveying deep emotional vulnerability and suppressed weeping before the final confession.", 0.86 seconds)
(very tense, very negative, very acrimony, enjoyable; 1.30 seconds duration) Because I'd honestly
[0.30 seconds pause]
(very tense, very negative, very acrimony, enjoyable; 1.65 seconds duration) rather have nothing from you
(very tense, very negative, very acrimony, enjoyable; 2.07 seconds duration) than ("A quick, tense inhalation taken right before the final clause 'than keep pretending', showing emotional strain and breath recovery under distress.", 0.60 seconds) keep pretending this doesn't hurt.
[0.25 seconds pause]
Automated review note Segments 1 and 2 suffer from a bizarrely hallucinated 'feeling horny' (sexual lust) emotion tag in an otherwise tense confrontation. Several breath/burst events overlap speech or are duplicated in the same pause gap (e.g., HVA_EVENT_0002 and HVA_EVENT_0003).
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
0.46 s duration · HVA_EVENT_0000
A sharp, audible intake of breath precedes the opening line, setting an immediate tone of emotional tension and preparation before speaking.
A sharp, audible intake of air through the mouth and nose occurs during a tense pause between 'for once' and 'This is not negotiable anymore', reflecting rising emotional distress and physical preparation for the next forceful statement.
A sharp, audible intake of air following 'This is not negotiable anymore', marking a tense respiratory reset before delivering the next confrontational line.
A sharp, audible intake of air through the mouth and nose occurs during a tense pause before the speaker delivers the resigned line 'Go ahead then', reflecting emotional bracing and physical tension.
A sharp, audible intake of air following the confrontational question 'okay?', marking a brief pause before the speaker drops into a quieter, more vulnerable register.
A soft, tremulous breathy release and quiet sob-like catch in the throat occurs right after the word 'okay', marking a sudden collapse of composure and emotional vulnerability.
A soft, tremulous breathy release and quiet sob-like exhalation occurs in the pause after the whispered 'Okay', conveying deep emotional vulnerability and suppressed weeping before the final confession.
Local emotion and delivery review — 12 speech segments
Absolute interval
Recognized speech and cue
Gemini ratings
Reviewer explanation
segment-001 Source sentence 1
I need you to stop talking to me like clean-bg, non-instructional, very feeling horny, enjoyable
Emotion 0/5 Delivery 3/5
Tagged with 'feeling horny' (sexual lust), which is entirely wrong and absurd for this tense, confrontational delivery.
segment-002 Source sentence 1
that. clean-bg, non-instructional, very feeling horny, enjoyable
Emotion 0/5 Delivery 3/5
Inherited the completely hallucinated 'feeling horny' tag for 'that.'
segment-003 Source sentence 2
No, very bright, very tonal, enjoyable
Emotion 3/5 Delivery 3/5
No emotion tag, which is acceptable for a brief interjection, though voice is sharp rather than tonal/bright.
segment-004 Source sentence 2
listen to me for once. very bright, very tonal, enjoyable
Emotion 3/5 Delivery 3/5
Omits specific anger/frustration emotion tags, but acoustic descriptors are acceptable.
segment-005 Source sentence 3
This is not negotiable anymore. very clean-bg, very tonal, very dismay, enjoyable
Emotion 4/5 Delivery 4/5
Dismay tag fits the exasperated, firm delivery decently.
segment-006 Source sentence 4
Go ahead, then. very tonal, very non-narrative, fervor
Emotion 1/5 Delivery 3/5
Tagged with 'fervor / hope/optimism', which contradicts the resigned, bitter tone.
segment-007 Source sentence 5
Cut me off if that's what you have to do, very bright, very oral-bright, very irritability, enjoyable
Emotion 4/5 Delivery 4/5
Irritability tag captures the sharp confrontational delivery well.
segment-008 Source sentence 5
okay? very bright, very oral-bright, very irritability, enjoyable
Emotion 4/5 Delivery 4/5
Irritability tag is suitable for the sharp questioning 'okay?'.
segment-009 Source sentence 6
Okay. very dissociated, very mumbled, very numbness
Emotion 4/5 Delivery 4/5
Dissociated/numbness captures the deflated, soft second 'okay'.
segment-010 Source sentence 7
Because I'd honestly very tense, very negative, very acrimony, enjoyable
Emotion 4/5 Delivery 4/5
Acrimony/bitterness matches the strained confession.
segment-011 Source sentence 7
rather have nothing from you very tense, very negative, very acrimony, enjoyable
Emotion 4/5 Delivery 4/5
Acrimony and tense delivery tags fit well.
segment-012 Source sentence 7
than keep pretending this doesn't hurt. very tense, very negative, very acrimony, enjoyable
Emotion 4/5 Delivery 4/5
Negative/acrimony matches the final vulnerable and hurt statement.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
4/5
Whole-recording emotion fit
4/5
Local speech emotion fit
2/5
Whole-recording talking-style fit
4/5
Local speech delivery fit
3/5
Burst-description fit
3/5
Burst-timing fit
2/5
Event precision rating
3/5
Event coverage rating
4/5
Transcript fidelity (diagnostic)
5/5
Speech-timing fit (diagnostic)
4/5
Readability (diagnostic)
4/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
50 words preserved; 9 selected events serialized; printed elapsed budget 23.090 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English6 selected intervals
Flipping the Deadweight
en-02600 · 20.50 s · Voice: Puck · extreme_physical
3.15 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A young adult masculine-coded speaker delivers an intensely dramatic, desperate plea in General American English. The performance begins in a hushed, strained whisper with low volume and close-mic intimacy, gradually escalating into a high-arousal, physically strained shout. The vocal timbre is characterized by prominent throat resonance, roughness, and breathiness, with noticeable vocal fry and constriction. The delivery is heavily fragmented into short, urgent breath groups punctuated by audible gasps, grunts of physical effort, and a sharp intake of breath. The emotional trajectory shifts from anxious reassurance to acute distress, pain, and desperate helplessness, maintaining high dramatic intensity throughout. The recording is dry and clean with minimal background interference.
SCRIPT — timed speech, local cues and free-form burst descriptions
(very brightening, very non-narrative, very relief, enjoyable; 0.47 seconds duration) Hey,
[0.25 seconds pause]
(very brightening, very non-narrative, very relief, enjoyable; 0.52 seconds duration) it's okay.
[0.36 seconds pause]
(fading, fading, awestruck, enjoyable; 1.44 seconds duration) I see you, I'm right here.
[0.24 seconds pause]
(asmr, darkening, very romantic desire, enjoyable; 1.20 seconds duration) Just keep looking at me.
(very clean-content, very accelerating, slightly mocking playfully; 1.36 seconds duration) Don't move your leg.
[0.64 seconds pause]
(very non-narrative, very scattered, slightly indifference, less-enjoyable; 0.31 seconds duration) Okay.
[1.13 seconds pause]
(very naturalistic, very whispered, slightly apathy, enjoyable; 1.65 seconds duration) I'm gonna brace right under the cage.
[0.75 seconds pause]
(hi-fi, clean-bg, burnout, enjoyable; 2.58 seconds duration) On three take a breath. ("A strained, guttural effort grunt occurs right after the speaker counts down 'on three', signaling physical bracing and exertion as they attempt to lift or stabilize something under pressure.", 0.68 seconds) ("A strained, guttural effort grunt occurs during a pause between instructions, characterized by a sudden glottal attack and pressed pharyngeal tension as the speaker physically braces or lifts something heavy.", 0.86 seconds)
[0.22 seconds pause]
(very non-narrative, very sharp-onset, less-enjoyable; 0.48 seconds duration) Come on
[0.24 seconds pause]
(very non-narrative, very sharp-onset, less-enjoyable; 0.32 seconds duration) up.
[0.64 seconds pause]
(very tense, very metallic, very irascibility, less-enjoyable; 1.59 seconds duration) Get up, damn it, please.
(very tense, very breathy, very short-temperedness; 1.49 seconds duration) ("A strained, guttural effort grunt occurs during a pause between pleading phrases, reflecting physical exertion and mounting frustration.", 0.64 seconds) ("A strained, guttural effort grunt and breath catch occurs right before the speaker pleads 'Just an inch', reflecting intense physical exertion and mounting desperation.", 0.64 seconds) Oh ("A strained, low-pitched vocalized groan of physical exertion and frustration ('Ugh') uttered right before pleading 'just an inch', blending directly into the desperate, breathless speech.", 0.58 seconds)
(very tense, very breathy, very short-temperedness; 1.80 seconds duration) just ("A sharp, audible intake of air following the grunt, signaling physical strain and breathlessness before pleading 'Just an inch'.", 0.54 seconds) an inch, come on, just give me an inch.
[0.82 seconds pause]
Automated review note The global description is fairly accurate regarding the vocal progression from whisper to strained shouting/groaning. However, the local sentence emotion tags heavily hallucinate romantic desire, teasing/mocking, relief, and apathy during an obvious emergency rescue scenario. Furthermore, burst events heavily overlap each other repeatedly around 12s and 17s.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
0.68 s duration · HVA_EVENT_0000
A strained, guttural effort grunt occurs right after the speaker counts down 'on three', signaling physical bracing and exertion as they attempt to lift or stabilize something under pressure.
A strained, guttural effort grunt occurs during a pause between instructions, characterized by a sudden glottal attack and pressed pharyngeal tension as the speaker physically braces or lifts something heavy.
A strained, guttural effort grunt and breath catch occurs right before the speaker pleads 'Just an inch', reflecting intense physical exertion and mounting desperation.
A strained, low-pitched vocalized groan of physical exertion and frustration ('Ugh') uttered right before pleading 'just an inch', blending directly into the desperate, breathless speech.
Local emotion and delivery review — 13 speech segments
Absolute interval
Recognized speech and cue
Gemini ratings
Reviewer explanation
segment-001 Source sentence 1
Hey, very brightening, very non-narrative, very relief, enjoyable
Emotion 1/5 Delivery 2/5
Tagged with 'relief' and 'brightening', but the speaker is speaking in a tense, anxious, calming whisper.
segment-002 Source sentence 1
it's okay. very brightening, very non-narrative, very relief, enjoyable
Emotion 1/5 Delivery 2/5
Reassuring whisper, not 'relief'.
segment-003 Source sentence 2
I see you, I'm right here. fading, fading, awestruck, enjoyable
Emotion 1/5 Delivery 3/5
Tagged as 'awestruck', which is inaccurate; the tone is supportive reassurance under intense stress.
segment-004 Source sentence 3
Just keep looking at me. asmr, darkening, very romantic desire, enjoyable
Emotion 0/5 Delivery 2/5
Tagged as 'romantic desire', which completely misinterprets the tense emergency situation.
segment-005 Source sentence 4
Don't move your leg. very clean-content, very accelerating, slightly mocking playfully
Emotion 0/5 Delivery 2/5
Tagged as 'mocking playfully', but this is an urgent instruction.
segment-006 Source sentence 5
Okay. very non-narrative, very scattered, slightly indifference, less-enjoyable
Emotion 1/5 Delivery 2/5
A tense, focused breathy 'Okay', not 'indifference' or 'scattered'.
segment-007 Source sentence 6
I'm gonna brace right under the cage. very naturalistic, very whispered, slightly apathy, enjoyable
Emotion 1/5 Delivery 3/5
Tagged with 'apathy', completely contrary to the focused emergency context.
segment-008 Source sentence 7
On three take a breath. hi-fi, clean-bg, burnout, enjoyable
Emotion 2/5 Delivery 3/5
Tagged as 'burnout'; speaker is bracing with physical exertion.
segment-009 Source sentence 8
Come on very non-narrative, very sharp-onset, less-enjoyable
Emotion 3/5 Delivery 3/5
A strained grunt/push ('Come on'), sharp-onset tag is reasonable.
segment-010 Source sentence 8
up. very non-narrative, very sharp-onset, less-enjoyable
Emotion 3/5 Delivery 3/5
Short exertion syllable ('up').
segment-011 Source sentence 9
Get up, damn it, please. very tense, very metallic, very irascibility, less-enjoyable
Emotion 4/5 Delivery 4/5
High physical tension and distress/desperation; 'irascibility'/'very tense' captures the strain reasonably well.
segment-012 Source sentence 10
Oh very tense, very breathy, very short-temperedness
Emotion 3/5 Delivery 3/5
A groaning exertion sound transcribed as 'Oh'.
segment-013 Source sentence 10
just an inch, come on, just give me an inch. very tense, very breathy, very short-temperedness
Emotion 4/5 Delivery 4/5
Very strained, desperate pleading under heavy physical effort.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
4/5
Whole-recording emotion fit
4/5
Local speech emotion fit
2/5
Whole-recording talking-style fit
4/5
Local speech delivery fit
3/5
Burst-description fit
3/5
Burst-timing fit
2/5
Event precision rating
3/5
Event coverage rating
4/5
Transcript fidelity (diagnostic)
4/5
Speech-timing fit (diagnostic)
3/5
Readability (diagnostic)
4/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
50 words preserved; 6 selected events serialized; printed elapsed budget 20.500 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English1 selected intervals
The Friendly Predator
en-02690 · 21.02 s · Voice: Leda · situation_inspired
3.50 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A single speaker with a young-adult feminine-coded voice delivers a playful, conspiratorial monologue in General American English. The timbre is clean, warm, and moderately bright, with prominent oral and mask resonance and a smooth, controlled delivery. The pacing is deliberate and relaxed, featuring clear articulation and natural conversational pauses. The performance carries an amused, teasing, and slightly mischievous tone, underscored by subtle breathy giggles and chuckles that blend seamlessly into the spoken phrases. The recording is pristine with a dead-quiet background and no audible distortion.
SCRIPT — timed speech, local cues and free-form burst descriptions
[0.24 seconds pause]
(very asmr, effortless, solace, enjoyable; 1.78 seconds duration) We could easily slip out the back,
[0.22 seconds pause]
(very asmr, effortless, solace, enjoyable; 0.76 seconds duration) you know. ("A brief, quiet, breathy chuckle occurs in the pause between 'you know' and 'My place', conveying playful conspiratorial amusement and blending smoothly into the subsequent phrase.", 0.48 seconds)
(very asmr, organic, pensiveness, enjoyable; 2.16 seconds duration) My place is only four blocks over,
[0.36 seconds pause]
(very asmr, organic, pensiveness, enjoyable; 2.20 seconds duration) and the drinks are a lot cheaper.
[1.16 seconds pause]
(very explicit, very heady, very astonishment; 0.80 seconds duration) Seriously?
[0.48 seconds pause]
(very clean-bg, very hi-fi, very stoicism, enjoyable; 1.87 seconds duration) Nobody here is even paying attention.
[1.01 seconds pause]
(very accelerating, very asmr, very compassion, enjoyable; 1.20 seconds duration) Which is convenient,
[0.32 seconds pause]
(very accelerating, very asmr, very compassion, enjoyable; 0.12 seconds duration) really,
[0.60 seconds pause]
(very accelerating, very asmr, very compassion, enjoyable; 2.33 seconds duration) because once we step outside into that alley,
[0.63 seconds pause]
(very accelerating, very asmr, very compassion, enjoyable; 2.56 seconds duration) you are going to do exactly what I tell you.
[0.22 seconds pause]
Automated review note Acoustic and style descriptions are solid, but emotion tags diverge heavily from the playful conspiratorial intent.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
0.48 s duration · HVA_EVENT_0000
A brief, quiet, breathy chuckle occurs in the pause between 'you know' and 'My place', conveying playful conspiratorial amusement and blending smoothly into the subsequent phrase.
Top-three type softmax: Snort: 27.2%; Sharp Inhale: 9.7%; Chuckle: 7.0%
Local emotion and delivery review — 10 speech segments
Absolute interval
Recognized speech and cue
Gemini ratings
Reviewer explanation
segment-001 Source sentence 1
We could easily slip out the back, very asmr, effortless, solace, enjoyable
Emotion 3/5 Delivery 3/5
Casual tone; tags like solace are inaccurate.
segment-002 Source sentence 1
you know. very asmr, effortless, solace, enjoyable
Emotion 3/5 Delivery 3/5
Short tag phrase; solace fits poorly.
segment-003 Source sentence 2
My place is only four blocks over, very asmr, organic, pensiveness, enjoyable
Emotion 3/5 Delivery 3/5
Tone is conspiratorial rather than pensiveness.
segment-004 Source sentence 2
and the drinks are a lot cheaper. very asmr, organic, pensiveness, enjoyable
Emotion 3/5 Delivery 3/5
Mischievous and relaxed rather than reflective.
segment-005 Source sentence 3
Seriously? very explicit, very heady, very astonishment
Emotion 3/5 Delivery 3/5
Rhetorical whisper rather than genuine astonishment.
segment-006 Source sentence 4
Nobody here is even paying attention. very clean-bg, very hi-fi, very stoicism, enjoyable
Emotion 2/5 Delivery 3/5
Stoicism poorly captures the playful delivery.
segment-007 Source sentence 5
Which is convenient, very accelerating, very asmr, very compassion, enjoyable
Emotion 2/5 Delivery 3/5
Compassion is a poor fit for the suggestive tone.
segment-008 Source sentence 5
really, very accelerating, very asmr, very compassion, enjoyable
Emotion 3/5 Delivery 3/5
Pacing tag does not match this brief single word.
segment-009 Source sentence 5
because once we step outside into that alley, very accelerating, very asmr, very compassion, enjoyable
Emotion 2/5 Delivery 3/5
Tone is teasing and dominant, not compassionate.
segment-010 Source sentence 5
you are going to do exactly what I tell you. very accelerating, very asmr, very compassion, enjoyable
Emotion 2/5 Delivery 3/5
Clear assertive edge is mislabeled as compassion.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
4/5
Whole-recording emotion fit
4/5
Local speech emotion fit
3/5
Whole-recording talking-style fit
4/5
Local speech delivery fit
3/5
Burst-description fit
4/5
Burst-timing fit
3/5
Event precision rating
3/5
Event coverage rating
3/5
Transcript fidelity (diagnostic)
4/5
Speech-timing fit (diagnostic)
4/5
Readability (diagnostic)
4/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
52 words preserved; 1 selected events serialized; printed elapsed budget 21.020 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English5 selected intervals
Baked in the Blacktop Glare
en-02741 · 30.46 s · Voice: Vindemiatrix · extreme_physical
3.40 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A single young-adult to mature-adult masculine-coded speaker delivers an intensely dramatic, menacing monologue in General American English. The vocal delivery is extremely low in pitch, situated in a deep baritone register with heavy chest and pharyngeal resonance, wrapped in a persistent, breathy half-whisper. The tempo is very slow and heavily chunked, with long, deliberate pauses between short, calculated phrases. Articulation is clear yet softened by the continuous whisper-talk style, creating an intimate, close-mic ASMR-like proximity. The emotional tone is saturated with cold malice, contempt, and simmering anger, maintaining a controlled, predatory stance rather than explosive rage. The recording is technically clean with minimal background noise and dry acoustics, allowing subtle respiratory details and vocal fry to remain prominent throughout.
SCRIPT — timed speech, local cues and free-form burst descriptions
(very whispered, very soft-onset, very fatigue, enjoyable; 1.91 seconds duration) Just ("A sharp, audible intake of air through the mouth and nose right before speaking, setting a tense, breathy, and physically strained tone.", 0.60 seconds) me and this road
[0.65 seconds pause]
(very whispered, very soft-onset, very fatigue, enjoyable; 0.73 seconds duration) all day.
[0.55 seconds pause]
(very meek, very whispered, very pining, enjoyable; 1.34 seconds duration) Just me and this road
[0.50 seconds pause]
(very meek, very whispered, very pining, enjoyable; 0.15 seconds duration) all
[0.33 seconds pause]
(very meek, very whispered, very pining, enjoyable; 0.02 seconds duration) day.
[0.94 seconds pause]
(very whispered, very soft-onset, very wistfulness, enjoyable; 1.32 seconds duration) I could live in this heat,
[0.36 seconds pause]
(very whispered, very soft-onset, very wistfulness, enjoyable; 0.35 seconds duration) honestly.
[0.85 seconds pause]
(very organic, very breathy, very warmth, enjoyable; 1.15 seconds duration) The way it holds you,
[0.29 seconds pause]
(very organic, very breathy, very warmth, enjoyable; 0.88 seconds duration) right to the bone
[0.80 seconds pause]
("A slow, deep audible inhalation and exhalation cycle occurs in the silence before 'Yeah', functioning as a deliberate, menacing physical reset that grounds the speaker's vocal tract and sets up the aggressive delivery of the following line.", 1.34 seconds)
[0.26 seconds pause]
(very organic, very breathy, very warmth, enjoyable; 0.63 seconds duration) Yeah.
[0.41 seconds pause]
(very thin-mask, very slow, very sadism; 0.63 seconds duration) Burn ya
[0.25 seconds pause]
(very thin-mask, very slow, very sadism; 0.31 seconds duration) bitch.
[0.49 seconds pause]
(very whispered, very steady, very saudade; 0.99 seconds duration) Cook me alive.
[0.61 seconds pause]
(very denasal, very hi-fi, very sadism, enjoyable; 2.85 seconds duration) I'll still lay this asphalt flat through your skull.
[0.99 seconds pause]
(very whispered, very noisy, very malice, enjoyable; 0.32 seconds duration) Gonna
("A sharp, percussive bilabial puff of air ('pff') expelled with contempt and derision, functioning as a dismissive scoff before the final challenge.", 0.68 seconds)
(very whispered, very noisy, very malice, enjoyable; 2.32 seconds duration) gonna pound it black until you scream.
[0.90 seconds pause]
(very breathy, very noisy, very fatigue; 1.14 seconds duration) ("A sharp, percussive bilabial puff of air ('pff') expelled with contempt and dismissive irritation right before the final challenge 'Just try and stop me.'", 0.68 seconds) Pff ("A quick, sharp preparatory intake of breath right before the final defiant challenge 'Just try and stop me', showing physical readiness and tight throat constriction.", 0.56 seconds)
[0.32 seconds pause]
(very breathy, very noisy, very fatigue; 0.71 seconds duration) just try and
[0.25 seconds pause]
(very breathy, very noisy, very fatigue; 0.64 seconds duration) stop me.
[0.30 seconds pause]
Automated review note The global description is strong, accurately capturing the menacing half-whisper delivery. However, segment-017 duplicates the scoff sound ('Pff') as both spoken text and an event, and an earlier event (HVA_EVENT_0002) hallucinates a scoff between 'Gonna' and 'gonna pound it black' where none occurred. Some emotional labels (meek, warmth) conflict with the overarching menacing tone.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
0.60 s duration · HVA_EVENT_0000
A sharp, audible intake of air through the mouth and nose right before speaking, setting a tense, breathy, and physically strained tone.
A slow, deep audible inhalation and exhalation cycle occurs in the silence before 'Yeah', functioning as a deliberate, menacing physical reset that grounds the speaker's vocal tract and sets up the aggressive delivery of the following line.
A sharp, percussive bilabial puff of air ('pff') expelled with contempt and dismissive irritation right before the final challenge 'Just try and stop me.'
A quick, sharp preparatory intake of breath right before the final defiant challenge 'Just try and stop me', showing physical readiness and tight throat constriction.
Transcript treats nonverbal scoff 'Pff' as spoken text while also annotating it as an event.
segment-018 Source sentence 9
just try and very breathy, very noisy, very fatigue
Emotion 3/5 Delivery 4/5
Breathy challenge delivery matches.
segment-019 Source sentence 9
stop me. very breathy, very noisy, very fatigue
Emotion 3/5 Delivery 4/5
Low quiet ending matches.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
4/5
Whole-recording emotion fit
4/5
Local speech emotion fit
3/5
Whole-recording talking-style fit
4/5
Local speech delivery fit
3/5
Burst-description fit
3/5
Burst-timing fit
3/5
Event precision rating
3/5
Event coverage rating
3/5
Transcript fidelity (diagnostic)
4/5
Speech-timing fit (diagnostic)
4/5
Readability (diagnostic)
4/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
60 words preserved; 5 selected events serialized; printed elapsed budget 30.460 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English1 selected intervals
Hearing Your Child Get Bullied and Intervening
en-04819 · 24.74 s · Voice: Leda · voicenet_extension
2.80 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A single speaker with a young-adult feminine-coded voice speaks in a close-mic, intimate, and deeply soothing whisper-talk style in General American English. The delivery is extremely gentle, slow, and deliberate, characterized by low vocal effort, soft attacks, and warm, breathy harmonicity with minimal metallic edge. The speaker maintains an intensely focused, protective, and affectionate stance, directly addressing a child in a comforting, reassuring manner. The acoustic environment is completely clean and dry with negligible background noise, highlighting the delicate breath textures and subtle mouth sounds. A soft, contented sigh occurs midway through the clip, naturally bridging the initial hushed instructions with the subsequent tender promises of safety.
SCRIPT — timed speech, local cues and free-form burst descriptions
(very darkening, very non-instructional, very homesickness, enjoyable; 1.66 seconds duration) What did you just say to my child?
[1.14 seconds pause]
(very non-narrative, very thin, butterflies in the stomach; 1.12 seconds duration) Look at me.
[0.40 seconds pause]
(very darkening, non-narrative; 0.85 seconds duration) Don't look away.
[0.59 seconds pause]
(choppy, smooth, very having a crush; 1.10 seconds duration) Look in my eyes.
[1.06 seconds pause]
(very choppy, thin-throat, apathy; 1.20 seconds duration) We're done here.
[0.40 seconds pause]
(very accelerating, very tonal, slightly doubt; 0.59 seconds duration) Understand?
[1.01 seconds pause]
(very cold, very meek, very boredom, less-enjoyable; 0.19 seconds duration) Good.
[1.17 seconds pause]
(very breathy, very whispered, very lust, enjoyable; 0.24 seconds duration) Hey
[0.30 seconds pause]
(very breathy, very whispered, very lust, enjoyable; 2.36 seconds duration) ("A deep, audible exhalation of relief and calming release occurs in the pause between the stern confrontation and the gentle reassurance, transitioning the vocal tract from tight tension to warm, open comfort.", 1.52 seconds) Come here, baby.
[0.78 seconds pause]
(smooth, clean-bg, very carnal desire; 1.09 seconds duration) Give me your backpack.
[1.15 seconds pause]
(very heady, casual, very lust, enjoyable; 1.80 seconds duration) Let's go home and get some ice cream, okay?
[1.08 seconds pause]
(very asmr, very whispered, very devotion, enjoyable; 1.90 seconds duration) You're completely safe now, I promise.
[0.50 seconds pause]
(very thin, very choppy, very having a crush; 1.06 seconds duration) I've got you.
Automated review note Acoustic and stylistic descriptions are broadly adequate, but segment emotion tags repeatedly assign romantic or sexual labels to protective and reassuring speech.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
1.52 s duration · HVA_EVENT_0000
A deep, audible exhalation of relief and calming release occurs in the pause between the stern confrontation and the gentle reassurance, transitioning the vocal tract from tight tension to warm, open comfort.
Look in my eyes. choppy, smooth, very having a crush
Emotion 2/5 Delivery 3/5
Infatuation tag does not match the firm confrontation.
segment-005 Source sentence 5
We're done here. very choppy, thin-throat, apathy
Emotion 3/5 Delivery 3/5
Apathy is somewhat off for firm finality, but delivery is controlled.
segment-006 Source sentence 6
Understand? very accelerating, very tonal, slightly doubt
Emotion 3/5 Delivery 3/5
Short rhetorical check; doubt is only partially appropriate.
segment-007 Source sentence 7
Good. very cold, very meek, very boredom, less-enjoyable
Emotion 3/5 Delivery 3/5
Tersely spoken; cold tag fits better than boredom.
segment-008 Source sentence 8
Hey very breathy, very whispered, very lust, enjoyable
Emotion 2/5 Delivery 3/5
Shift to comfort tagged inappropriately as lust.
segment-009 Source sentence 8
Come here, baby. very breathy, very whispered, very lust, enjoyable
Emotion 2/5 Delivery 3/5
Gentle maternal comfort mischaracterized as lust.
segment-010 Source sentence 9
Give me your backpack. smooth, clean-bg, very carnal desire
Emotion 2/5 Delivery 3/5
Carnal desire label conflicts with soothing parental speech.
segment-011 Source sentence 10
Let's go home and get some ice cream, okay? very heady, casual, very lust, enjoyable
Emotion 2/5 Delivery 3/5
Reassuring offer tagged as lust inappropriately.
segment-012 Source sentence 11
You're completely safe now, I promise. very asmr, very whispered, very devotion, enjoyable
Emotion 4/5 Delivery 4/5
Affection and devotion fit the soothing promise well.
segment-013 Source sentence 12
I've got you. very thin, very choppy, very having a crush
Emotion 2/5 Delivery 3/5
Reassuring closure mislabeled with crush/infatuation.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
3/5
Whole-recording emotion fit
3/5
Local speech emotion fit
2/5
Whole-recording talking-style fit
3/5
Local speech delivery fit
3/5
Burst-description fit
3/5
Burst-timing fit
3/5
Event precision rating
3/5
Event coverage rating
3/5
Transcript fidelity (diagnostic)
4/5
Speech-timing fit (diagnostic)
4/5
Readability (diagnostic)
4/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
49 words preserved; 1 selected events serialized; printed elapsed budget 24.740 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English3 selected intervals
Confronting the Parent Who Made You Invisible
en-04841 · 33.86 s · Voice: Zephyr · existing_challenge_inspired
3.35 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A feminine-coded, young-adult speaker delivers an intensely dramatic, confrontational monologue in General American English. The performance is marked by a slow, deliberate tempo, heavy chunking, and prominent pauses that heighten the sense of emotional tension and bitterness. The vocal timbre is clear, slightly breathy, and forward in oral and mask resonance, with a controlled, slightly pressed throat tension that conveys simmering anger and disappointment. The delivery is highly focused and direct, addressing an interlocutor with sharp, accusatory emphasis on key words like 'meeting', 'phone', and 'phone'. A prominent exasperated sigh occurs midway through the clip, naturally bridging the initial grievance with the escalating confrontation. The recording is studio-clean with minimal background noise and high intelligibility, capturing a theatrical yet emotionally grounded performance.
SCRIPT — timed speech, local cues and free-form burst descriptions
[0.24 seconds pause]
(very clean-bg, very hi-fi, very craving, enjoyable; 1.40 seconds duration) I need you to hear something.
[1.64 seconds pause]
(very hi-fi, smooth, very acrimony, enjoyable; 1.30 seconds duration) When I broke my wrist
[0.38 seconds pause]
(very hi-fi, smooth, very acrimony, enjoyable; 1.52 seconds duration) in seventh grade,
[0.80 seconds pause]
(very hi-fi, smooth, very acrimony, enjoyable; 2.00 seconds duration) you asked the nurse if we were done
[0.24 seconds pause]
(very hi-fi, smooth, very acrimony, enjoyable; 1.04 seconds duration) because you had a meeting.
[0.28 seconds pause]
(overlapping non-speech vocal events; 1.52 seconds duration) ("A sharp, derisive nasal snort and breath expulsion occurs in the pause following 'you had a meeting', conveying bitter disbelief and contempt before the speaker continues.", 0.96 seconds) ("A sharp, derisive nasal snort and breath expulsion occurs in the pause following 'you had a meeting', conveying bitter disbelief and contempt before the speaker directly challenges the listener.", 0.96 seconds)
[0.60 seconds pause]
(very darkening, very fading, very disdain, enjoyable; 2.00 seconds duration) You don't even remember that, do you?
[2.20 seconds pause]
("A heavy, weary exhalation and breath release occurs in the silence before 'It's sickening', conveying deep emotional exhaustion and lingering hurt.", 1.10 seconds)
[0.78 seconds pause]
(very whispered, very asmr, very abhorrence; 1.07 seconds duration) It's sickening.
[1.17 seconds pause]
(very asmr, very clean-bg, very sexual lust, enjoyable; 0.90 seconds duration) Standing here
[0.70 seconds pause]
(very asmr, very clean-bg, very sexual lust, enjoyable; 0.39 seconds duration) still
[0.25 seconds pause]
(very asmr, very clean-bg, very sexual lust, enjoyable; 2.36 seconds duration) wanting you to look up from your phone
[0.28 seconds pause]
(very asmr, very clean-bg, very sexual lust, enjoyable; 0.98 seconds duration) and just see me.
[2.06 seconds pause]
(very thin, very hi-fi, very having a crush, enjoyable; 1.12 seconds duration) Look at me.
[1.44 seconds pause]
(very accelerating, very steady, very pride, enjoyable; 1.75 seconds duration) Acknowledge what you took.
[0.35 seconds pause]
Automated review note The global description captures the monologue's tone, pacing, and sigh well. However, duplicate overlapping events were proposed for the chuckle/snort around 11-12s, and segments 8 through 12 have completely hallucinated romantic/sexual emotion tags ('sexual lust', 'having a crush') that conflict sharply with the bitter confrontational performance.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
0.96 s duration · HVA_EVENT_0000
A sharp, derisive nasal snort and breath expulsion occurs in the pause following 'you had a meeting', conveying bitter disbelief and contempt before the speaker continues.
A sharp, derisive nasal snort and breath expulsion occurs in the pause following 'you had a meeting', conveying bitter disbelief and contempt before the speaker directly challenges the listener.
Top-three type softmax: Snort: 19.9%; Cough: 13.9%; Displeased Grunt: 6.5%
1.10 s duration · HVA_EVENT_0002
A heavy, weary exhalation and breath release occurs in the silence before 'It's sickening', conveying deep emotional exhaustion and lingering hurt.
Local emotion and delivery review — 13 speech segments
Absolute interval
Recognized speech and cue
Gemini ratings
Reviewer explanation
segment-001 Source sentence 1
I need you to hear something. very clean-bg, very hi-fi, very craving, enjoyable
Emotion 3/5 Delivery 4/5
The tone is tense and confrontational rather than strongly characterized as craving.
segment-002 Source sentence 2
When I broke my wrist very hi-fi, smooth, very acrimony, enjoyable
Emotion 4/5 Delivery 4/5
Bitterness fits the narrative setup well.
segment-003 Source sentence 2
in seventh grade, very hi-fi, smooth, very acrimony, enjoyable
Emotion 4/5 Delivery 4/5
Consistent quiet bitter delivery.
segment-004 Source sentence 2
you asked the nurse if we were done very hi-fi, smooth, very acrimony, enjoyable
Emotion 4/5 Delivery 4/5
Accusatory and bitter delivery matches.
segment-005 Source sentence 2
because you had a meeting. very hi-fi, smooth, very acrimony, enjoyable
Emotion 4/5 Delivery 4/5
Delivers the punchline of the anecdote with bitter restraint.
segment-006 Source sentence 3
You don't even remember that, do you? very darkening, very fading, very disdain, enjoyable
Emotion 4/5 Delivery 4/5
Disdain and contempt fit well.
segment-007 Source sentence 4
It's sickening. very whispered, very asmr, very abhorrence
Emotion 4/5 Delivery 4/5
Whispered disgust/abhorrence is accurately captured.
segment-008 Source sentence 5
Standing here very asmr, very clean-bg, very sexual lust, enjoyable
Emotion 1/5 Delivery 3/5
Labeling this as sexual lust is completely inaccurate for this dramatic, hurt monologue.
segment-009 Source sentence 5
still very asmr, very clean-bg, very sexual lust, enjoyable
Emotion 1/5 Delivery 3/5
Sexual lust is inappropriate for this quiet pause/word.
segment-010 Source sentence 5
wanting you to look up from your phone very asmr, very clean-bg, very sexual lust, enjoyable
Emotion 1/5 Delivery 3/5
Severe mismatch with 'sexual lust'; the speaker expresses vulnerable hurt and resentment.
segment-011 Source sentence 5
and just see me. very asmr, very clean-bg, very sexual lust, enjoyable
Emotion 1/5 Delivery 3/5
Severe mismatch with sexual lust.
segment-012 Source sentence 6
Look at me. very thin, very hi-fi, very having a crush, enjoyable
Emotion 1/5 Delivery 3/5
'Having a crush' / infatuation is completely dissonant with the confrontational demand.
segment-013 Source sentence 7
Acknowledge what you took. very accelerating, very steady, very pride, enjoyable
Emotion 2/5 Delivery 3/5
'Pride' does not match the bitter, demanding resolution.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
4/5
Whole-recording emotion fit
4/5
Local speech emotion fit
2/5
Whole-recording talking-style fit
4/5
Local speech delivery fit
3/5
Burst-description fit
4/5
Burst-timing fit
3/5
Event precision rating
3/5
Event coverage rating
4/5
Transcript fidelity (diagnostic)
5/5
Speech-timing fit (diagnostic)
4/5
Readability (diagnostic)
4/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
58 words preserved; 3 selected events serialized; printed elapsed budget 33.860 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English7 selected intervals
Two Strangers Stuck in an Elevator: Entire Relationship Arc
en-05245 · 25.82 s · Voice: Sadachbia · voicenet_extension
3.30 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A young adult masculine-coded speaker delivers an intimate, playful, and flirtatious monologue in General American English. The vocal delivery is grounded in a warm, resonant baritone register with noticeable chest resonance and a relaxed, slightly breathy quality. The tempo is leisurely and conversational, punctuated by natural pauses, soft chuckles, and an audible sigh of relief. The speaker maintains a close, direct interpersonal focus, blending teasing amusement with romantic interest and subtle vulnerability. The recording is clean and dry with negligible background noise, capturing natural vocal micro-textures and organic respiratory transitions.
SCRIPT — timed speech, local cues and free-form burst descriptions
("A soft, quiet exhale of wistful reflection opens the recording right before the speaker begins to talk, blending smoothly into the initial phrase.", 0.46 seconds)
(choppy, naturalistic, very startlement; 0.92 seconds duration) Crazy to think.
[0.22 seconds pause]
(very denasal, full, very amazement, enjoyable; 2.23 seconds duration) Three hours ago we didn't even know each other's names.
[0.39 seconds pause]
(very chesty, very full, very mocking playfully, enjoyable; 3.54 seconds duration) ("A brief, quiet chuckle with subtle breathy pulses occurs in the pause before 'sharing that crushed Snickers', reflecting fond amusement at the memory.", 0.74 seconds) ("A brief, quiet chuckle with subtle breathy pulses occurs right before 'sharing that crushed Snickers', conveying fond amusement and nostalgic intimacy.", 0.68 seconds) Sharing that crushed Snickers on the floor. ("A soft, quiet chuckle with gentle breath pulses occurs right after 'floor', expressing fond amusement and nostalgic warmth before continuing the reflective narrative.", 0.74 seconds)
[1.04 seconds pause]
(very denasal, very full, very mournfulness, enjoyable; 1.07 seconds duration) Sitting in the dark,
[0.53 seconds pause]
(very denasal, very full, very mournfulness, enjoyable; 1.45 seconds duration) talking about our regrets.
[1.59 seconds pause]
(very full, very naturalistic, enjoyable; 1.88 seconds duration) Then the door's actually pry open.
("A soft, audible inhalation taken in the pause before asking 'So, what happens now?', functioning as a contemplative reset.", 0.58 seconds)
[0.34 seconds pause]
(clean-content, scattered, very doubt; 0.31 seconds duration) So,
[0.65 seconds pause]
(clean-content, scattered, very doubt; 2.98 seconds duration) what happens now? ("A slow, audible inhalation and soft exhalation during a long pause after 'what happens now?', capturing contemplative hesitation and emotional vulnerability.", 1.22 seconds) ("A slow, prolonged, breathy sigh of contemplation and mild romantic hesitation occurs in the pause before the final question, blending smoothly into the quiet, intimate delivery.", 1.34 seconds)
[0.78 seconds pause]
(very full, darkening, very suspicion, enjoyable; 1.93 seconds duration) Do we pretend this whole thing never happened,
[0.39 seconds pause]
(very full, darkening, very suspicion, enjoyable; 0.17 seconds duration) or
[0.79 seconds pause]
(very full, darkening, very suspicion, enjoyable; 1.14 seconds duration) do you want to give me your number?
[0.44 seconds pause]
Automated review note Acoustic style descriptions are solid, but several burst events are duplicated across overlapping intervals. Emotional labeling skews overly dramatic with mournfulness and suspicion.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
0.46 s duration · HVA_EVENT_0000
A soft, quiet exhale of wistful reflection opens the recording right before the speaker begins to talk, blending smoothly into the initial phrase.
A brief, quiet chuckle with subtle breathy pulses occurs in the pause before 'sharing that crushed Snickers', reflecting fond amusement at the memory.
Top-three type softmax: Exasperated Sigh: 15.6%; Chuckle: 10.7%; Snort: 9.4%
0.68 s duration · HVA_EVENT_0002
A brief, quiet chuckle with subtle breathy pulses occurs right before 'sharing that crushed Snickers', conveying fond amusement and nostalgic intimacy.
Top-three type softmax: Chuckle: 20.7%; Breathy Giggle: 13.7%; Snort: 8.8%
0.74 s duration · HVA_EVENT_0003
A soft, quiet chuckle with gentle breath pulses occurs right after 'floor', expressing fond amusement and nostalgic warmth before continuing the reflective narrative.
Top-three type softmax: Chuckle: 71.9%; Snicker: 7.4%; Breathy Giggle: 5.9%
0.58 s duration · HVA_EVENT_0004
A soft, audible inhalation taken in the pause before asking 'So, what happens now?', functioning as a contemplative reset.
A slow, audible inhalation and soft exhalation during a long pause after 'what happens now?', capturing contemplative hesitation and emotional vulnerability.
A slow, prolonged, breathy sigh of contemplation and mild romantic hesitation occurs in the pause before the final question, blending smoothly into the quiet, intimate delivery.
do you want to give me your number? very full, darkening, very suspicion, enjoyable
Emotion 3/5 Delivery 3/5
Flirtatious delivery mismatches the negative suspicion cue.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
4/5
Whole-recording emotion fit
3/5
Local speech emotion fit
3/5
Whole-recording talking-style fit
4/5
Local speech delivery fit
3/5
Burst-description fit
3/5
Burst-timing fit
3/5
Event precision rating
3/5
Event coverage rating
3/5
Transcript fidelity (diagnostic)
4/5
Speech-timing fit (diagnostic)
3/5
Readability (diagnostic)
4/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
55 words preserved; 7 selected events serialized; printed elapsed budget 25.820 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English13 selected intervals
Stroke of Desire
en-06071 · 33.89 s · Voice: Erinome · situation_inspired
3.60 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A young adult feminine-coded speaker delivers an intensely intimate, breathless, and sexually charged monologue in General American English. The performance is dominated by heavy respiration, frequent panting, and close-mic ASMR-like proximity, creating a high level of vulnerability and physical arousal. The vocal timbre is breathy, warm, and slightly strained, with low harmonicity and prominent oral and pharyngeal resonance. Phrasing is fragmented and choppy, broken up by frequent gasps, sighs, and moans of pleasure and longing. The emotional tone shifts between sensual ecstasy, physical exhaustion, and desperate pleading, maintaining a high degree of dramatic intimacy throughout.
SCRIPT — timed speech, local cues and free-form burst descriptions
("A sharp, breathy intake of air occurs right at the start of the recording before any speech begins, establishing an immediate state of physical exertion and breathlessness.", 0.68 seconds)
[2.20 seconds pause]
(very oral-bright, very thin-chest, slightly apathy, enjoyable; 2.20 seconds duration) Just another two thousand meters.
[0.44 seconds pause]
(very whispered, very organic, very comfort, enjoyable; 1.19 seconds duration) Nothing crazy today.
[1.05 seconds pause]
(very meek, very asmr, very emotional blunting; 1.12 seconds duration) Steady pace.
[0.56 seconds pause]
(very whispered, very breathy, very solace, enjoyable; 1.80 seconds duration) Keep the split clean. ("A quick, audible gasp-like inhalation taken in the pause before 'Except', showing physical breathlessness.", 0.64 seconds) ("A sharp, breathy intake of air punctuated by a slight vocalized catch, occurring right before the exclamation 'God', signaling sudden physical or emotional arousal.", 1.08 seconds)
(very breathy, very whispered, very fatigue, enjoyable; 2.05 seconds duration) ("A heavy, weary exhalation that transitions into a sharp intake of air, marking a pause in the running cadence.", 0.92 seconds) Except ("A sharp, audible intake of air interrupting the sentence right before 'God', signaling sudden physical or emotional arousal.", 0.76 seconds)
(very breathy, very whispered, very fatigue, enjoyable; 3.99 seconds duration) God, ("A sharp, constricted gasp of air taken right before the exclamation 'God', reflecting sudden physical or emotional arousal.", 0.66 seconds) the way your ("A quick, breathy gasp interrupting the phrase 'the way your hands', showing physical breathlessness and sensual tension.", 0.80 seconds) hands pulled me in.
[0.24 seconds pause]
(very breathy, very whispered, very unease, enjoyable; 4.45 seconds duration) You didn't give me any room to breathe, did you? ("A heavy, rapid double-breath cycle of panting in the pause after 'did you', showing escalating physical arousal and breathlessness.", 1.30 seconds)
[1.31 seconds pause]
(very whispered, very darkening, very weariness; 0.63 seconds duration) Harder
[0.33 seconds pause]
(very whispered, very darkening, very weariness; 0.03 seconds duration) now.
[0.61 seconds pause]
(very fading, very thin, very respite; 1.99 seconds duration) Just like that. ("A strained, breathy vocalized moan of physical pleasure and exertion following the whispered command 'Just like that'.", 1.08 seconds)
("A strained, breathy vocalized moan of physical pleasure and exertion.", 0.84 seconds)
[0.68 seconds pause]
("A quick, shallow panting breath taken in isolation during a pause, showing continued physical exertion.", 0.76 seconds)
[1.17 seconds pause]
(very breathy, very scattered, helplessness, enjoyable; 0.89 seconds duration) Don't stop.
(very non-narrative, very meek, very powerlessness; 1.16 seconds duration) ("A quick, desperate gasp for air between 'Don't stop' and 'Please', showing heightened vulnerability and physical exhaustion.", 0.70 seconds) Please. ("A quick, desperate gasp for air between 'Don't stop' and 'Please', showing heightened vulnerability.", 0.50 seconds)
[0.27 seconds pause]
(very breathy, very whispered, very carnal desire, enjoyable; 0.91 seconds duration) Don't let me go.
[0.34 seconds pause]
Automated review note The speaker delivers a breathy, intimate, ASMR-style whispered performance with heavy panting throughout. Several annotated burst events heavily overlap each other (e.g. HVA_EVENT_0001 through 0005, and 0011 with 0012) creating redundant detections over single breath instances, while segment-009 has an artificially compressed duration budget.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
0.68 s duration · HVA_EVENT_0000
A sharp, breathy intake of air occurs right at the start of the recording before any speech begins, establishing an immediate state of physical exertion and breathlessness.
Top-three type softmax: Exasperated Sigh: 52.7%; Breath: 7.1%; Other / rare burst: 6.9%
0.64 s duration · HVA_EVENT_0001
A quick, audible gasp-like inhalation taken in the pause before 'Except', showing physical breathlessness.
A sharp, breathy intake of air punctuated by a slight vocalized catch, occurring right before the exclamation 'God', signaling sudden physical or emotional arousal.
Local emotion and delivery review — 13 speech segments
Absolute interval
Recognized speech and cue
Gemini ratings
Reviewer explanation
segment-001 Source sentence 1
Just another two thousand meters. very oral-bright, very thin-chest, slightly apathy, enjoyable
Emotion 4/5 Delivery 4/5
Whispered, breathless athletic setup.
segment-002 Source sentence 2
Nothing crazy today. very whispered, very organic, very comfort, enjoyable
Emotion 4/5 Delivery 4/5
Quiet, steady whisper.
segment-003 Source sentence 3
Steady pace. very meek, very asmr, very emotional blunting
Emotion 4/5 Delivery 4/5
Breath-led pacing, accurate ASMR whisper tone.
segment-004 Source sentence 4
Keep the split clean. very whispered, very breathy, very solace, enjoyable
Emotion 4/5 Delivery 4/5
Soft and whispered delivery.
segment-005 Source sentence 5
Except very breathy, very whispered, very fatigue, enjoyable
Emotion 4/5 Delivery 3/5
Isolated single word framed by heavy breaths; timing overlaps with burst boundaries.
segment-006 Source sentence 5
God, the way your hands pulled me in. very breathy, very whispered, very fatigue, enjoyable
Emotion 4/5 Delivery 4/5
Breathy, intimate delivery with sensual inflection.
segment-007 Source sentence 6
You didn't give me any room to breathe, did you? very breathy, very whispered, very unease, enjoyable
Emotion 4/5 Delivery 4/5
Quiet breathless questioning.
segment-008 Source sentence 7
Harder very whispered, very darkening, very weariness
Emotion 4/5 Delivery 4/5
Brief whisper.
segment-009 Source sentence 7
now. very whispered, very darkening, very weariness
Emotion 3/5 Delivery 2/5
Duration assigned (0.03s) is unrealistically narrow and poorly segmented relative to speech.
segment-010 Source sentence 8
Just like that. very fading, very thin, very respite
Emotion 4/5 Delivery 4/5
Soft whisper followed immediately by breath/sigh.
segment-011 Source sentence 9
Don't stop. very breathy, very scattered, helplessness, enjoyable
Emotion 4/5 Delivery 4/5
Sensual whisper.
segment-012 Source sentence 10
Please. very non-narrative, very meek, very powerlessness
Emotion 4/5 Delivery 3/5
Overlaps with breath tags around 'Please'.
segment-013 Source sentence 11
Don't let me go. very breathy, very whispered, very carnal desire, enjoyable
Emotion 4/5 Delivery 4/5
Faint breathy whisper ending the phrase.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
4/5
Whole-recording emotion fit
4/5
Local speech emotion fit
4/5
Whole-recording talking-style fit
4/5
Local speech delivery fit
4/5
Burst-description fit
3/5
Burst-timing fit
2/5
Event precision rating
3/5
Event coverage rating
3/5
Transcript fidelity (diagnostic)
5/5
Speech-timing fit (diagnostic)
3/5
Readability (diagnostic)
4/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
45 words preserved; 13 selected events serialized; printed elapsed budget 33.890 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English6 selected intervals
The Twins with Opposite Personalities
en-06195 · 20.62 s · Voice: Algieba · laions_got_talent
2.70 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A single young-adult to mature-adult masculine-coded speaker delivers an exasperated, conversational monologue in General American English. The vocal timbre is grounded with balanced chest and oral resonance, moderate warmth, and a clean harmonic profile with slight natural roughness. The delivery is brisk and dynamic, characterized by punchy emphasis on key words like 'actually', 'packed', and 'entirely', conveying mounting irritation, impatience, and mild theatrical distress. The speaker's stance is assertive and direct, addressing an unseen interlocutor with a mix of disbelief and weary resignation. Two distinct exasperated sighs punctuate the speech, naturally integrated into the phrasing to emphasize the speaker's frustration. The recording is studio-clean with minimal background noise and sharp acoustic definition.
SCRIPT — timed speech, local cues and free-form burst descriptions
(very bright, very blended-resonance, very startlement, enjoyable; 0.54 seconds duration) Come ("A quick, audible preparatory intake of breath occurs right before the speaker launches into the opening phrase 'Come on', setting an urgent, slightly exasperated tone.", 0.50 seconds) on,
(very bright, very blended-resonance, very startlement, enjoyable; 3.62 seconds duration) signing the lease today means our cafe actually happens.
(very clean-bg, hi-fi, very optimism, enjoyable; 2.21 seconds duration) It's gonna be packed every morning. ("A brief, quiet chuckle occurs in the pause between 'every morning' and 'Just picture the patio lights', featuring a soft, breathy vocal release that naturally reflects the speaker's wry amusement before resuming speech.", 0.52 seconds)
(very playful, very casual, very bantering; 1.38 seconds duration) ("A brief, quiet chuckle occurs right before the speaker says 'Just picture the patio lights', expressing wry amusement and blending smoothly into the following phrase.", 0.40 seconds) Just picture the patio
[0.25 seconds pause]
(very playful, very casual, very bantering; 0.74 seconds duration) lights. ("A soft, amused chuckle occurs in the pause between 'patio lights' and 'or picture us broke', featuring gentle breathy pulses that naturally reflect the speaker's playful, teasing tone.", 0.70 seconds)
[0.22 seconds pause]
(very full, very clean-content, slightly ease, enjoyable; 0.32 seconds duration) Or
[0.24 seconds pause]
(very full, very clean-content, slightly ease, enjoyable; 2.88 seconds duration) picture us broke by Christmas when nobody shows up.
[0.24 seconds pause]
(full, even-rhythm, awestruck, enjoyable; 3.77 seconds duration) Same face in the mirror, sure, but your head is entirely in the clouds.
[0.23 seconds pause]
(clean-bg, emphatic, very shock, enjoyable; 1.50 seconds duration) We're rushing right off a cliff.
(very full, fading, honor; 2.16 seconds duration) ("A soft, breathy exasperated sigh occurs in the pause before the final resigned statement, releasing built-up tension and blending smoothly into the weary, self-critical tone of the closing words.", 0.66 seconds) ("A soft, breathy exasperated sigh occurs in the pause before the final resigned statement 'We always do this', releasing built-up tension and blending naturally into the weary, self-deprecating tone.", 0.52 seconds) We always do this.
[0.32 seconds pause]
Automated review note Acoustic and style descriptions are largely grounded, but burst events include duplicated/overlapping intervals and local emotion tags mismatch speaker cynicism.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
0.50 s duration · HVA_EVENT_0000
A quick, audible preparatory intake of breath occurs right before the speaker launches into the opening phrase 'Come on', setting an urgent, slightly exasperated tone.
A brief, quiet chuckle occurs in the pause between 'every morning' and 'Just picture the patio lights', featuring a soft, breathy vocal release that naturally reflects the speaker's wry amusement before resuming speech.
Top-three type softmax: Chuckle: 50.3%; Breathy Giggle: 6.6%; Snort: 5.8%
0.40 s duration · HVA_EVENT_0002
A brief, quiet chuckle occurs right before the speaker says 'Just picture the patio lights', expressing wry amusement and blending smoothly into the following phrase.
Top-three type softmax: Chuckle: 51.7%; Snort: 7.0%; Snicker: 5.7%
0.70 s duration · HVA_EVENT_0003
A soft, amused chuckle occurs in the pause between 'patio lights' and 'or picture us broke', featuring gentle breathy pulses that naturally reflect the speaker's playful, teasing tone.
A soft, breathy exasperated sigh occurs in the pause before the final resigned statement, releasing built-up tension and blending smoothly into the weary, self-critical tone of the closing words.
A soft, breathy exasperated sigh occurs in the pause before the final resigned statement 'We always do this', releasing built-up tension and blending naturally into the weary, self-deprecating tone.
Local emotion and delivery review — 10 speech segments
Absolute interval
Recognized speech and cue
Gemini ratings
Reviewer explanation
segment-001 Source sentence 1
Come on, very bright, very blended-resonance, very startlement, enjoyable
Emotion 2/5 Delivery 3/5
Prompt startlement emotion does not capture the mildly exasperated pleading.
segment-002 Source sentence 1
signing the lease today means our cafe actually happens. very bright, very blended-resonance, very startlement, enjoyable
Emotion 2/5 Delivery 3/5
Carries forward startlement instead of urgent, persuasive reasoning.
segment-003 Source sentence 2
It's gonna be packed every morning. very clean-bg, hi-fi, very optimism, enjoyable
Emotion 3/5 Delivery 3/5
Optimism tag aligns reasonably well with the hopeful projection.
segment-004 Source sentence 3
Just picture the patio very playful, very casual, very bantering
Emotion 3/5 Delivery 3/5
Playful and bantering tags reflect the conversational pivot.
segment-005 Source sentence 3
lights. very playful, very casual, very bantering
Emotion 3/5 Delivery 3/5
Short remnant carries acceptable playfully casual cues.
segment-006 Source sentence 4
Or very full, very clean-content, slightly ease, enjoyable
Emotion 2/5 Delivery 3/5
Single word conjunction; contentment/ease does not fit the cynical pivot.
segment-007 Source sentence 4
picture us broke by Christmas when nobody shows up. very full, very clean-content, slightly ease, enjoyable
Emotion 2/5 Delivery 3/5
Ease is inappropriate for the pessimistic counter-argument.
segment-008 Source sentence 5
Same face in the mirror, sure, but your head is entirely in the clouds. full, even-rhythm, awestruck, enjoyable
Emotion 2/5 Delivery 3/5
Awestruck label misses the critical and skeptical tone.
segment-009 Source sentence 6
We're rushing right off a cliff. clean-bg, emphatic, very shock, enjoyable
Emotion 3/5 Delivery 3/5
Shock emotion loosely maps to the alarmed, urgent warning.
segment-010 Source sentence 7
We always do this. very full, fading, honor
Emotion 2/5 Delivery 3/5
Honor/pride contradicts the weary, resigned conclusion.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
3/5
Whole-recording emotion fit
3/5
Local speech emotion fit
2/5
Whole-recording talking-style fit
3/5
Local speech delivery fit
3/5
Burst-description fit
3/5
Burst-timing fit
2/5
Event precision rating
3/5
Event coverage rating
3/5
Transcript fidelity (diagnostic)
4/5
Speech-timing fit (diagnostic)
3/5
Readability (diagnostic)
3/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
56 words preserved; 6 selected events serialized; printed elapsed budget 20.620 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English8 selected intervals
Congested Call
en-06251 · 30.70 s · Voice: Achernar · situation_inspired
3.55 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A young adult female speaker delivers an informal, highly conversational monologue in General American English, characterized by a relaxed, low-arousal state and prominent fatigue. Her vocal timbre is warm and slightly breathy, with a balanced oral and pharyngeal resonance and mild vocal fry creeping in during low-energy phrases. The tempo is slow and hesitant, broken up by frequent pauses, audible breaths, and self-interruptions as she struggles to maintain focus. The performance is rich in paralinguistic vocal bursts, including an exasperated sigh, a weary groan, a suppressed snicker, and a sharp intake of breath, all of which blend seamlessly into the spoken delivery. The recording is clean and dry with negligible background noise, capturing an authentic, unscripted moment of someone talking through a heavy head cold.
SCRIPT — timed speech, local cues and free-form burst descriptions
[0.32 seconds pause]
(very fading, very thin, very emotional blunting; 0.48 seconds duration) Hey.
[0.72 seconds pause]
(very cold, very non-narrative, slightly serenity, less-enjoyable; 0.50 seconds duration) Yeah.
[0.38 seconds pause]
(very clean-bg, darkening, very boredom; 1.14 seconds duration) Still buried in
[0.22 seconds pause]
(very clean-bg, darkening, very boredom; 0.74 seconds duration) blankets.
[0.56 seconds pause]
(very soft-onset, very vulnerable, very fatigue, enjoyable; 3.40 seconds duration) Honestly ("A low-pitched, breathy vocalized groan of physical discomfort and mental fog occurs during a pause between phrases, blending naturally into the sluggish, half-asleep delivery.", 0.68 seconds) feel like my brain is full of wet cotton.
[0.38 seconds pause]
("A slow, audible, weary inhalation and exhalation cycle occurs in the pause after 'cotton', reflecting physical fatigue and labored respiration.", 1.06 seconds)
[0.58 seconds pause]
(very non-narrative, very scattered, very discomfiture, less-enjoyable; 0.31 seconds duration) Wait,
[0.33 seconds pause]
(very non-narrative, very scattered, very discomfiture, less-enjoyable; 0.56 seconds duration) shut up.
(very explicit, very submissive, very ribbing, enjoyable; 2.10 seconds duration) Did he really say that to her? ("A brief, self-deprecating breathy chuckle occurs right after questioning whether she really said that to her, blending smoothly into the exasperated 'Oh, god.'", 0.60 seconds)
[0.26 seconds pause]
(very non-narrative, very non-narrative, fatigue, less-enjoyable; 0.58 seconds duration) Oh ("A brief, self-deprecating breathy chuckle escapes right before saying 'God', blending naturally with the speaker's amused embarrassment.", 0.58 seconds)
(very non-narrative, very non-narrative, fatigue, less-enjoyable; 0.31 seconds duration) God.
[0.59 seconds pause]
(very darkening, very cold, very revulsion; 0.91 seconds duration) You're terrible.
[0.81 seconds pause]
("A soft, quiet breathy chuckle with gentle nasal and oral airflow, expressing self-deprecating amusement at their own silly comment.", 0.90 seconds)
[0.58 seconds pause]
("A quick, audible preparatory intake of breath through the mouth immediately preceding the word 'Listen', signaling an effort to refocus.", 0.42 seconds)
[0.22 seconds pause]
(oral-bright, thin-chest, very exhaustion, enjoyable; 0.55 seconds duration) Listen,
[0.25 seconds pause]
(oral-bright, thin-chest, very exhaustion, enjoyable; 1.18 seconds duration) once my fever breaks,
[0.34 seconds pause]
(oral-bright, thin-chest, very exhaustion, enjoyable; 1.34 seconds duration) we have to get smoothies.
[0.34 seconds pause]
(dialogic, non-narrative, slightly teasing; 2.32 seconds duration) We can sit outside so I don't give you plague and
[0.24 seconds pause]
(dialogic, non-narrative, slightly teasing; 0.08 seconds duration) hang
[0.32 seconds pause]
(dialogic, non-narrative, slightly teasing; 0.24 seconds duration) on.
[0.24 seconds pause]
(very flat, very fading, very mortification, less-enjoyable; 0.24 seconds duration) Sorry.
[0.20 seconds pause]
(clean-bg, heady, slightly skepticism; 1.81 seconds duration) ("A brief, quiet, breathy chuckle escapes right after saying 'sorry', blending naturally with the speaker's self-conscious amusement.", 0.72 seconds) ("A quick, breathy gasp-like inhalation taken right before resuming speech with 'Anyway, Saturday?'.", 0.68 seconds) Anyway,
[0.55 seconds pause]
(clean-bg, heady, slightly skepticism; 0.54 seconds duration) Saturday?
[0.56 seconds pause]
Automated review note The speaker is audibly sick and tired, and the global monologue description captures this context nicely. However, several local emotion labels dramatically overstate negative affect (e.g., revulsion, mortification) when the tone is playfully conversational. In the latter section (around 27-28s), the speaker clearly coughs, but the annotation mischaracterizes it as a chuckle followed by a gasp.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
0.68 s duration · HVA_EVENT_0000
A low-pitched, breathy vocalized groan of physical discomfort and mental fog occurs during a pause between phrases, blending naturally into the sluggish, half-asleep delivery.
A brief, self-deprecating breathy chuckle occurs right after questioning whether she really said that to her, blending smoothly into the exasperated 'Oh, god.'
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
56 words preserved; 8 selected events serialized; printed elapsed budget 30.700 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English7 selected intervals
Hiding Disappointment When Plans Get Canceled
en-06532 · 19.13 s · Voice: Rasalgethi · laions_got_talent
3.40 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A young adult masculine-coded speaker delivers a casual, conversational monologue in General American English, expressing a relaxed and contented state of mind. The vocal timbre is warm, grounded in a baritone register with noticeable chest resonance and clean harmonicity. The delivery is unhurried and natural, featuring comfortable phrasing, clear articulation, and subtle conversational emphasis. The speaker conveys genuine relief and mild amusement, punctuated by a soft chuckle and a contented sigh that blend seamlessly into the spoken flow. The recording is studio-clean with negligible background noise and balanced acoustic presence.
SCRIPT — timed speech, local cues and free-form burst descriptions
(very emphatic, very bright, irritation, enjoyable; 2.56 seconds duration) Oh ("A soft, breathy chuckle occurs right at the start before the speaker utters 'Oh, no', setting a warm, amused, and reassuring tone.", 0.56 seconds) no, hey, don't worry about it at all.
(very playful, effortless, very provoking lightly, enjoyable; 1.28 seconds duration) Seriously. ("A brief, quiet chuckle occurs right after 'Seriously,' featuring a soft, breathy vocal release that naturally bridges the polite refusal into the explanation of being tired.", 0.44 seconds) ("A brief, quiet chuckle occurs right after 'Seriously,' featuring a soft, breathy vocal release that naturally bridges the polite refusal into the explanation of being tired.", 0.42 seconds)
(very non-instructional, very clean-content, very alleviation, enjoyable; 5.04 seconds duration) Honestly, my week was pretty exhausting anyway, so an open evening sounds kinda nice.
[0.48 seconds pause]
(very casual, organic, fervor, enjoyable; 2.52 seconds duration) I can just catch up on some laundry, maybe
[0.28 seconds pause]
(very casual, organic, fervor, enjoyable; 1.26 seconds duration) finish that book.
(very accelerating, very brightening, very respite, enjoyable; 1.19 seconds duration) ("A soft, quiet chuckle with gentle breath release occurs right after 'maybe finish that book', naturally blending into the reassuring phrase 'It's totally fine'.", 0.64 seconds) It's ("A soft, quiet chuckle with gentle breath release occurs right after 'finish that book', blending naturally into the reassuring phrase 'It's totally fine'.", 0.44 seconds) totally fine.
[0.27 seconds pause]
(very clean-content, very non-narrative, slightly courage, enjoyable; 1.69 seconds duration) We'll just figure out another time down the road.
[0.47 seconds pause]
(very effortless, very clean-content, very shock; 0.23 seconds duration) Really.
(very non-narrative, very fading, apathy, enjoyable; 1.43 seconds duration) ("A soft, relaxed exhalation of relief and mild fatigue occurs right before the final phrase 'no big deal', naturally releasing residual tension and blending smoothly into the concluding words.", 0.58 seconds) ("A soft, quiet exhale of relief and mild fatigue occurs right before the final phrase 'no big deal', naturally releasing residual tension and blending smoothly into the relaxed, low-energy delivery.", 0.52 seconds) No big deal.
[0.43 seconds pause]
Automated review note Annotation captures overall acoustic qualities well, but includes several duplicate burst events and over-interprets mild conversational turns as extreme emotional states.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
0.56 s duration · HVA_EVENT_0000
A soft, breathy chuckle occurs right at the start before the speaker utters 'Oh, no', setting a warm, amused, and reassuring tone.
A brief, quiet chuckle occurs right after 'Seriously,' featuring a soft, breathy vocal release that naturally bridges the polite refusal into the explanation of being tired.
Top-three type softmax: Chuckle: 67.5%; Breathy Giggle: 7.7%; Snicker: 5.7%
0.42 s duration · HVA_EVENT_0002
A brief, quiet chuckle occurs right after 'Seriously,' featuring a soft, breathy vocal release that naturally bridges the polite refusal into the explanation of being tired.
Top-three type softmax: Chuckle: 56.9%; Breathy Giggle: 9.1%; Snicker: 7.2%
0.64 s duration · HVA_EVENT_0003
A soft, quiet chuckle with gentle breath release occurs right after 'maybe finish that book', naturally blending into the reassuring phrase 'It's totally fine'.
Top-three type softmax: Chuckle: 53.1%; Breathy Giggle: 16.4%; Snicker: 5.6%
0.44 s duration · HVA_EVENT_0004
A soft, quiet chuckle with gentle breath release occurs right after 'finish that book', blending naturally into the reassuring phrase 'It's totally fine'.
Top-three type softmax: Chuckle: 55.0%; Breathy Giggle: 15.2%; Snicker: 5.1%
0.58 s duration · HVA_EVENT_0005
A soft, relaxed exhalation of relief and mild fatigue occurs right before the final phrase 'no big deal', naturally releasing residual tension and blending smoothly into the concluding words.
A soft, quiet exhale of relief and mild fatigue occurs right before the final phrase 'no big deal', naturally releasing residual tension and blending smoothly into the relaxed, low-energy delivery.
Local emotion and delivery review — 9 speech segments
Absolute interval
Recognized speech and cue
Gemini ratings
Reviewer explanation
segment-001 Source sentence 1
Oh no, hey, don't worry about it at all. very emphatic, very bright, irritation, enjoyable
Emotion 2/5 Delivery 3/5
Tone is friendly reassurance rather than irritation.
segment-002 Source sentence 2
Seriously. very playful, effortless, very provoking lightly, enjoyable
Emotion 3/5 Delivery 3/5
Fairly casual and reassuring rather than strongly provoking.
segment-003 Source sentence 3
Honestly, my week was pretty exhausting anyway, so an open evening sounds kinda nice. very non-instructional, very clean-content, very alleviation, enjoyable
Emotion 4/5 Delivery 3/5
Fits the relaxed, relieved conversational tone well.
segment-004 Source sentence 4
I can just catch up on some laundry, maybe very casual, organic, fervor, enjoyable
Emotion 3/5 Delivery 3/5
Fervor is overly dramatic for a casual thought.
segment-005 Source sentence 4
finish that book. very casual, organic, fervor, enjoyable
Emotion 3/5 Delivery 3/5
Simple continuation, low intensity.
segment-006 Source sentence 5
It's totally fine. very accelerating, very brightening, very respite, enjoyable
Emotion 3/5 Delivery 3/5
Light and warm, tags are somewhat exaggerated.
segment-007 Source sentence 6
We'll just figure out another time down the road. very clean-content, very non-narrative, slightly courage, enjoyable
Emotion 3/5 Delivery 3/5
Moderate fit; calm forward-looking statement.
segment-008 Source sentence 7
Really. very effortless, very clean-content, very shock
Emotion 2/5 Delivery 3/5
Affirming reassurance rather than shock.
segment-009 Source sentence 8
No big deal. very non-narrative, very fading, apathy, enjoyable
Emotion 3/5 Delivery 3/5
Apathy does not quite match the gentle, friendly conclusion.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
4/5
Whole-recording emotion fit
4/5
Local speech emotion fit
3/5
Whole-recording talking-style fit
4/5
Local speech delivery fit
3/5
Burst-description fit
3/5
Burst-timing fit
3/5
Event precision rating
3/5
Event coverage rating
3/5
Transcript fidelity (diagnostic)
4/5
Speech-timing fit (diagnostic)
4/5
Readability (diagnostic)
4/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
52 words preserved; 7 selected events serialized; printed elapsed budget 19.130 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English5 selected intervals
Telling Someone They Have a Habit That Hurts You
en-06597 · 22.34 s · Voice: Algieba · haiku_batch
3.00 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A single young-adult to mature-adult masculine-coded speaker delivers an emotionally charged, confrontational monologue in General American English. The vocal delivery is grounded in a baritone register with noticeable chest resonance, moderate roughness, and a slightly breathy, strained quality that conveys deep distress, disappointment, and simmering anger. The tempo is deliberate and slow, broken into short, heavy breath groups with clear, punchy emphasis on key grievances. The speaker's stance is direct, assertive, and intensely focused on the listener, projecting a sense of wounded pride and exasperation. Two distinct negative vocal bursts—a heavy exasperated sigh and a sharp, frustrated groan—punctuate the pauses between sentences, seamlessly integrating into the speaker's respiratory rhythm and emotional state. The recording is technically clean, dry, and close-mic'd with minimal room reflection and no background noise.
SCRIPT — timed speech, local cues and free-form burst descriptions
(very warm, very denasal, very sympathy, enjoyable; 1.17 seconds duration) When you do that,
[0.27 seconds pause]
(very warm, very denasal, very sympathy, enjoyable; 1.94 seconds duration) it makes me feel completely invisible.
[0.70 seconds pause]
(very full, very denasal, very fatigue, enjoyable; 1.43 seconds duration) I brought it up last week,
[0.25 seconds pause]
(very full, very denasal, very fatigue, enjoyable; 0.36 seconds duration) gently,
[0.44 seconds pause]
(very full, very denasal, very fatigue, enjoyable; 2.50 seconds duration) and nothing shifted. ("A heavy, weary exhalation occurs in the pause following 'and nothing shifted', releasing built-up tension and signaling deep disappointment before the speaker attempts to regain control.", 0.76 seconds) ("A low-pitched, breathy exasperated sigh occurs during a pause after 'and nothing shifted', releasing pent-up frustration and fatigue before the speaker firmly interrupts the listener.", 0.94 seconds)
(very darkening, very fading, slightly stoicism; 1.35 seconds duration) No, don't interrupt
[0.23 seconds pause]
(very darkening, very fading, slightly stoicism; 0.32 seconds duration) me.
[0.24 seconds pause]
(very denasal, very full, enjoyable; 1.44 seconds duration) I'm asking you to listen
[0.48 seconds pause]
(very denasal, very full, enjoyable; 0.84 seconds duration) right now. ("A quick, audible gasp-like inhalation occurs in the pause before 'I shouldn't have to fight', reflecting emotional strain and physical preparation to speak.", 0.62 seconds)
[0.44 seconds pause]
(very fading, very negative, letdown, enjoyable; 4.16 seconds duration) I shouldn't have to fight just to be heard on this. ("A heavy, weary exhalation and breath release occurs in the pause before 'It hurts', conveying deep emotional exhaustion and physical deflation.", 0.90 seconds) ("A heavy, weary exhalation of breath occurs in the pause before 'It hurts', conveying deep emotional exhaustion and physical deflation.", 0.64 seconds)
(very negative, very whispered, very physical pain, enjoyable; 0.94 seconds duration) It hurts,
[0.34 seconds pause]
(very negative, very whispered, very physical pain, enjoyable; 1.96 seconds duration) and I need you to actually stop.
[0.54 seconds pause]
Automated review note Global summary is largely accurate, but several emotion tags mismatch the confrontational delivery, and breath events overlap redundantly.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
0.76 s duration · HVA_EVENT_0000
A heavy, weary exhalation occurs in the pause following 'and nothing shifted', releasing built-up tension and signaling deep disappointment before the speaker attempts to regain control.
A low-pitched, breathy exasperated sigh occurs during a pause after 'and nothing shifted', releasing pent-up frustration and fatigue before the speaker firmly interrupts the listener.
A quick, audible gasp-like inhalation occurs in the pause before 'I shouldn't have to fight', reflecting emotional strain and physical preparation to speak.
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
51 words preserved; 5 selected events serialized; printed elapsed budget 22.340 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English5 selected intervals
Confidence Presence
en-07521 · 25.01 s · Voice: Sulafat · haiku_batch
3.20 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A single speaker with a young-adult feminine-coded voice delivers an intensely intimate, comforting monologue in General American English. The acoustic presentation is characterized by a very low, hushed volume bordering on an ASMR whisper, with prominent breathiness, soft attacks, and warm, close-mic oral resonance. The tempo is slow and measured, broken into short, deliberate breath groups with gentle pauses. The emotional tone is deeply centered on reassurance, affection, and protective calm, aimed directly at soothing an anxious listener. Audible, controlled inhalations and soft sighs punctuate the speech naturally, blending seamlessly into the quiet, studio-clean recording environment.
SCRIPT — timed speech, local cues and free-form burst descriptions
(very relaxed, very steady, very appreciation, enjoyable; 1.19 seconds duration) My name's David.
(very explicit, very whispered, very trust, enjoyable; 1.29 seconds duration) I'm right here to help
[0.24 seconds pause]
(very explicit, very whispered, very trust, enjoyable; 0.48 seconds duration) you.
(very whispered, very asmr, very terror, enjoyable; 4.28 seconds duration) ("A soft, deliberate inhalation and exhalation occurs during the pause between 'help you' and 'I know', serving as a calming respiratory reset that naturally grounds the speaker's soothing tone.", 0.66 seconds) I know your hands are shaking, and I know you're terrified right now. ("A soft, deliberate inhalation occurs in the pause before 'That is okay', functioning as a calming reset that naturally leads into the gentle, reassuring tone of the following phrase.", 0.58 seconds)
[0.28 seconds pause]
(very relaxed, very hi-fi, very alleviation; 1.03 seconds duration) That is okay.
[0.81 seconds pause]
(even-rhythm, smooth, emotional blunting, enjoyable; 1.36 seconds duration) I'm not walking away,
[0.24 seconds pause]
(even-rhythm, smooth, emotional blunting, enjoyable; 2.42 seconds duration) and nobody is rushing that door. ("A soft, deliberate inhalation and gentle exhalation in the pause after 'rushing that door', serving as a calming reset before offering reassurance.", 0.80 seconds)
(very even-rhythm, very effortless, very inspiration, enjoyable; 1.62 seconds duration) We're gonna figure this out together.
[0.84 seconds pause]
(very clean-bg, very choppy, insensitivity, enjoyable; 0.38 seconds duration) Step
[0.26 seconds pause]
(very clean-bg, very choppy, insensitivity, enjoyable; 0.55 seconds duration) by step.
[0.65 seconds pause]
(very whispered, very soft-onset, very romantic desire; 0.31 seconds duration) Just
[0.25 seconds pause]
(very whispered, very soft-onset, very romantic desire; 3.00 seconds duration) keep listening to my voice. ("A soft, prolonged, breathy exhalation occurs in the pause following 'Just keep listening to my voice', functioning as a calming, grounding release that naturally leads into the final reassuring phrase.", 1.08 seconds) ("A slow, deep audible inhalation and soft exhalation occurs in the silence following 'Just keep listening to my voice', serving as a calming physiological reset before delivering the final reassuring statements.", 1.08 seconds)
[0.60 seconds pause]
(very relaxed, very organic, very liberation, enjoyable; 1.36 seconds duration) You're safe now.
[0.40 seconds pause]
(very thin, very choppy, very infatuation; 0.96 seconds duration) I've got you.
[0.21 seconds pause]
Automated review note The speaker delivers a soft, ASMR-style comforting monologue. The annotation frequently confuses words mentioned in the script (e.g. 'terrified' leading to a label of terror) with the speaker's actual calming affect, and adds ungrounded labels like romantic desire, infatuation, and emotional blunting. The burst events between 19.7s and 21.5s are duplicate/overlapping entries for the single prolonged breath.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
0.66 s duration · HVA_EVENT_0000
A soft, deliberate inhalation and exhalation occurs during the pause between 'help you' and 'I know', serving as a calming respiratory reset that naturally grounds the speaker's soothing tone.
A soft, deliberate inhalation occurs in the pause before 'That is okay', functioning as a calming reset that naturally leads into the gentle, reassuring tone of the following phrase.
A soft, prolonged, breathy exhalation occurs in the pause following 'Just keep listening to my voice', functioning as a calming, grounding release that naturally leads into the final reassuring phrase.
A slow, deep audible inhalation and soft exhalation occurs in the silence following 'Just keep listening to my voice', serving as a calming physiological reset before delivering the final reassuring statements.
Step very clean-bg, very choppy, insensitivity, enjoyable
Emotion 2/5 Delivery 3/5
'Insensitivity' is contradictory to the delicate, soothing delivery.
segment-010 Source sentence 7
by step. very clean-bg, very choppy, insensitivity, enjoyable
Emotion 2/5 Delivery 3/5
Gentle pacing, not insensitive.
segment-011 Source sentence 8
Just very whispered, very soft-onset, very romantic desire
Emotion 2/5 Delivery 4/5
'Romantic desire' is an over-interpretation of ASMR/whisper intimacy.
segment-012 Source sentence 8
keep listening to my voice. very whispered, very soft-onset, very romantic desire
Emotion 2/5 Delivery 4/5
Intended to comfort/soothe, not clearly romantic desire.
segment-013 Source sentence 9
You're safe now. very relaxed, very organic, very liberation, enjoyable
Emotion 4/5 Delivery 4/5
Warm reassurance and relief.
segment-014 Source sentence 10
I've got you. very thin, very choppy, very infatuation
Emotion 2/5 Delivery 3/5
Infatuation is inaccurate; delivery is quiet and reassuring.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
4/5
Whole-recording emotion fit
4/5
Local speech emotion fit
2/5
Whole-recording talking-style fit
4/5
Local speech delivery fit
3/5
Burst-description fit
3/5
Burst-timing fit
3/5
Event precision rating
3/5
Event coverage rating
3/5
Transcript fidelity (diagnostic)
5/5
Speech-timing fit (diagnostic)
4/5
Readability (diagnostic)
4/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
56 words preserved; 5 selected events serialized; printed elapsed budget 25.010 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English10 selected intervals
Steam to Cryo Scream
en-07644 · 25.97 s · Voice: Enceladus · extreme_physical
3.30 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A young adult masculine-coded speaker delivers a highly dramatic, whispered monologue in General American English, depicting extreme physical cold and exhaustion. The vocal production is heavily unvoiced and breathy, characterized by low harmonicity, prominent throat resonance, and audible shivering breaths and gasps. The tempo is slow and fragmented into short, labored breath groups with frequent pauses. The emotional tone is dominated by distress, physical pain, and helplessness, with an intimate, close-mic ASMR-like quality against a clean, silent background.
SCRIPT — timed speech, local cues and free-form burst descriptions
(very fading, very naturalistic, lightness, enjoyable; 0.47 seconds duration) Yeah,
[0.41 seconds pause]
(very fading, very naturalistic, lightness, enjoyable; 1.20 seconds duration) this one's mine.
(very denasal, breathy, lighthearted fun; 2.16 seconds duration) Let ("A sharp, audible intake of air through constricted vocal cords occurs in the pause between 'Yeah' and 'Let the rookies watch', signaling physical exertion and cold-induced breathlessness.", 0.70 seconds) the rookies watch. ("A sharp, audible intake of air through constricted vocal cords occurs in the pause between 'Let the rookies watch' and 'This is how work gets done', reflecting physical exertion and cold-induced breathlessness.", 0.50 seconds)
[0.56 seconds pause]
(very noisy, fading, fatigue; 0.32 seconds duration) This
[0.32 seconds pause]
(very noisy, fading, fatigue; 1.84 seconds duration) is how work gets done.
[1.20 seconds pause]
("A sharp, shivering intake of breath through the teeth and mouth, demonstrating physical cold and respiratory strain.", 0.98 seconds)
(very decelerating, very flat-resonance, emotional blunting, less-enjoyable; 0.45 seconds duration) God.
[0.31 seconds pause]
(very fading, very non-narrative, very burnout, less-enjoyable; 0.70 seconds duration) Oh. ("A low-pitched, breathy vocalized groan of physical discomfort and cold-induced strain, trailing off into a shuddering exhale.", 0.70 seconds)
[0.60 seconds pause]
(very breathy, very throaty, very burnout; 0.86 seconds duration) I can't even
[0.42 seconds pause]
(very breathy, very throaty, very burnout; 2.40 seconds duration) draw full breath. ("A low-pitched, breathy vocalized groan of physical discomfort and cold-induced strain, trailing off into a shuddering exhale.", 0.84 seconds) ("A low-pitched, strained groan of physical discomfort and cold-induced constriction, blending directly into the onset of 'It's so stupid'.", 0.54 seconds)
(very tense, very throaty, very disgust; 1.34 seconds duration) It's so stupid
[0.58 seconds pause]
(very tense, very throaty, very disgust; 0.42 seconds duration) stupid
[0.22 seconds pause]
(very tense, very throaty, very disgust; 1.78 seconds duration) cold out here. ("A sharp, constricted gasp-like inhalation through the teeth and throat, demonstrating acute cold-induced breathlessness.", 0.74 seconds) ("A quick, shallow panting breath cycle occurs in the pause after 'out here', showing rapid respiratory recovery under physical strain.", 0.76 seconds)
[0.68 seconds pause]
(very whispered, very noisy, very exhaustion; 2.33 seconds duration) ("A quick, breathy gasp-like inhalation immediately preceding the phrase 'My chest is just locked up tight', demonstrating physical constriction.", 0.52 seconds) My chest is just locked up tight,
[0.33 seconds pause]
(very whispered, very noisy, very exhaustion; 2.16 seconds duration) freezing before the sweat even tries.
("A final trailing, shivering exhalation and breath release at the very end of the recording, fading out into silence.", 0.80 seconds)
[0.13 seconds pause]
Automated review note The global caption captures the cold, breathy, shivering acting performance accurately. However, earlier segments are mistakenly tagged with positive emotions (lightness, lighthearted fun), and several burst events heavily overlap directly with spoken words rather than strictly marking isolated nonverbal vocalizations.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
0.70 s duration · HVA_EVENT_0000
A sharp, audible intake of air through constricted vocal cords occurs in the pause between 'Yeah' and 'Let the rookies watch', signaling physical exertion and cold-induced breathlessness.
Top-three type softmax: Sharp Inhale: 45.5%; Breath: 13.3%; Sniff: 2.8%
0.50 s duration · HVA_EVENT_0001
A sharp, audible intake of air through constricted vocal cords occurs in the pause between 'Let the rookies watch' and 'This is how work gets done', reflecting physical exertion and cold-induced breathlessness.
Top-three type softmax: Sharp Inhale: 31.9%; Breath: 9.7%; Snort: 6.1%
0.98 s duration · HVA_EVENT_0002
A sharp, shivering intake of breath through the teeth and mouth, demonstrating physical cold and respiratory strain.
Top-three type softmax: Breath: 15.4%; Other / rare burst: 15.2%; Sharp Inhale: 12.2%
0.70 s duration · HVA_EVENT_0003
A low-pitched, breathy vocalized groan of physical discomfort and cold-induced strain, trailing off into a shuddering exhale.
Top-three type softmax: Frustrated Groan: 18.5%; Other / rare burst: 16.8%; Exasperated Sigh: 15.9%
0.84 s duration · HVA_EVENT_0004
A low-pitched, breathy vocalized groan of physical discomfort and cold-induced strain, trailing off into a shuddering exhale.
Top-three type softmax: Other / rare burst: 11.9%; Exasperated Sigh: 10.3%; Trembling Whimper: 9.0%
0.54 s duration · HVA_EVENT_0005
A low-pitched, strained groan of physical discomfort and cold-induced constriction, blending directly into the onset of 'It's so stupid'.
draw full breath. very breathy, very throaty, very burnout
Emotion 4/5 Delivery 4/5
Matches labored breathing and vocal fatigue.
segment-010 Source sentence 7
It's so stupid very tense, very throaty, very disgust
Emotion 3/5 Delivery 4/5
More frustrated/shivering than pure disgust.
segment-011 Source sentence 7
stupid very tense, very throaty, very disgust
Emotion 3/5 Delivery 3/5
Shivering repetition with cold distress.
segment-012 Source sentence 7
cold out here. very tense, very throaty, very disgust
Emotion 4/5 Delivery 4/5
Tense and throaty shivering fits well.
segment-013 Source sentence 8
My chest is just locked up tight, very whispered, very noisy, very exhaustion
Emotion 4/5 Delivery 4/5
Whispered exhaustion matches the audible shivering whisper.
segment-014 Source sentence 8
freezing before the sweat even tries. very whispered, very noisy, very exhaustion
Emotion 4/5 Delivery 4/5
Shivering and exhausted whisper matches well.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
4/5
Whole-recording emotion fit
4/5
Local speech emotion fit
3/5
Whole-recording talking-style fit
4/5
Local speech delivery fit
3/5
Burst-description fit
3/5
Burst-timing fit
2/5
Event precision rating
3/5
Event coverage rating
3/5
Transcript fidelity (diagnostic)
4/5
Speech-timing fit (diagnostic)
3/5
Readability (diagnostic)
4/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
42 words preserved; 10 selected events serialized; printed elapsed budget 25.970 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English2 selected intervals
Supervisor Discipline
en-07790 · 24.77 s · Voice: Achird · haiku_batch
2.85 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A single young-adult to mature-adult masculine-coded speaker delivers a serious, professional reprimand in General American English. The vocal delivery is grounded in a baritone register with noticeable chest resonance, balanced warmth, and a clean, controlled timbre. The pacing is deliberate and measured, characterized by short, segmented phrasing and clear, crisp articulation. The speaker maintains an authoritative, direct interpersonal stance, projecting firmness and controlled disappointment rather than explosive anger. Affectively, the performance is dominated by sternness, disappointment, and mild exasperation, with low-to-moderate arousal and a slightly negative valence. A single sharp preparatory inhale is audible between phrases, integrating naturally into the disciplined, formal cadence. The recording is studio-clean with minimal background noise and high intelligibility.
SCRIPT — timed speech, local cues and free-form burst descriptions
[0.32 seconds pause]
(very news-anchor, very formal, insensitivity, enjoyable; 3.37 seconds duration) This discrepancy resulted in a direct financial loss.
[0.39 seconds pause]
(very full, warm, slightly mocking playfully, enjoyable; 0.16 seconds duration) So
[0.24 seconds pause]
(very full, warm, slightly mocking playfully, enjoyable; 1.98 seconds duration) the formal reprimand goes in your file.
[0.42 seconds pause]
(very accelerating, very clean-content, slightly existential void; 0.75 seconds duration) That's policy.
[1.45 seconds pause]
(very accelerating, very non-narrative, very clemency, enjoyable; 0.70 seconds duration) ("A sharp, audible intake of breath occurs right before the speaker utters 'Hey', functioning as a sudden preparatory reset to regain the listener's attention during an intense confrontation.", 0.52 seconds) Hey, ("A quick, audible preparatory intake of breath occurs right before the speaker sharply commands 'Pull it together', signaling rising tension and a shift into direct confrontation.", 0.36 seconds)
[0.38 seconds pause]
(very accelerating, very non-narrative, very clemency, enjoyable; 0.72 seconds duration) pull it together.
[1.28 seconds pause]
(very non-narrative, very accelerating, slightly feeling horny, enjoyable; 1.04 seconds duration) Look at me.
[0.32 seconds pause]
(very negative, very tense, very exasperation, enjoyable; 2.07 seconds duration) If Marcus messed up the count,
[0.41 seconds pause]
(very negative, very tense, very exasperation, enjoyable; 1.68 seconds duration) you don't take the hit for him.
[0.88 seconds pause]
(very accelerating, very effortless, very thankfulness; 1.30 seconds duration) I appreciate the loyalty.
[0.22 seconds pause]
(very fading, very flat, slightly existential void; 0.58 seconds duration) I really do.
[0.70 seconds pause]
(very denasal, very long-phrased, very reverence, enjoyable; 3.05 seconds duration) But you cannot set yourself on fire to keep someone else warm.
[0.36 seconds pause]
Automated review note Acoustic baseline and pacing are well-captured, but multiple segment-level emotion tags introduce spurious affective states.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
0.52 s duration · HVA_EVENT_0000
A sharp, audible intake of breath occurs right before the speaker utters 'Hey', functioning as a sudden preparatory reset to regain the listener's attention during an intense confrontation.
A quick, audible preparatory intake of breath occurs right before the speaker sharply commands 'Pull it together', signaling rising tension and a shift into direct confrontation.
Local emotion and delivery review — 12 speech segments
Absolute interval
Recognized speech and cue
Gemini ratings
Reviewer explanation
segment-001 Source sentence 1
This discrepancy resulted in a direct financial loss. very news-anchor, very formal, insensitivity, enjoyable
Emotion 3/5 Delivery 4/5
Formal delivery matches well; numb tag is overstated.
segment-002 Source sentence 2
So very full, warm, slightly mocking playfully, enjoyable
Emotion 2/5 Delivery 3/5
Short transition; playful mocking is inappropriate.
segment-003 Source sentence 2
the formal reprimand goes in your file. very full, warm, slightly mocking playfully, enjoyable
Emotion 2/5 Delivery 3/5
Tone remains firm rather than teasing.
segment-004 Source sentence 3
That's policy. very accelerating, very clean-content, slightly existential void
Emotion 3/5 Delivery 3/5
Flat and matter-of-fact; existential label is excessive.
segment-005 Source sentence 4
Hey, very accelerating, very non-narrative, very clemency, enjoyable
Emotion 2/5 Delivery 3/5
Direct interjection; clemency does not fit.
segment-006 Source sentence 4
pull it together. very accelerating, very non-narrative, very clemency, enjoyable
Emotion 3/5 Delivery 3/5
Firm command; affection tag is inaccurate.
segment-007 Source sentence 5
Look at me. very non-narrative, very accelerating, slightly feeling horny, enjoyable
Emotion 2/5 Delivery 3/5
Serious instruction; sexual lust tag is incorrect.
segment-008 Source sentence 6
If Marcus messed up the count, very negative, very tense, very exasperation, enjoyable
Emotion 4/5 Delivery 4/5
Exasperation fits the audible tension reasonably well.
segment-009 Source sentence 6
you don't take the hit for him. very negative, very tense, very exasperation, enjoyable
Emotion 4/5 Delivery 4/5
Firm delivery aligns with exasperated descriptors.
segment-010 Source sentence 7
I appreciate the loyalty. very accelerating, very effortless, very thankfulness
Emotion 3/5 Delivery 3/5
Gratitude tag matches text but delivery is subdued.
segment-011 Source sentence 8
I really do. very fading, very flat, slightly existential void
Emotion 3/5 Delivery 3/5
Quiet reinforcement; void tag is exaggerated.
segment-012 Source sentence 9
But you cannot set yourself on fire to keep someone else warm. very denasal, very long-phrased, very reverence, enjoyable
Emotion 3/5 Delivery 3/5
Earnest advice; reverence tag is somewhat misplaced.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
3/5
Whole-recording emotion fit
3/5
Local speech emotion fit
2/5
Whole-recording talking-style fit
4/5
Local speech delivery fit
3/5
Burst-description fit
3/5
Burst-timing fit
2/5
Event precision rating
3/5
Event coverage rating
3/5
Transcript fidelity (diagnostic)
4/5
Speech-timing fit (diagnostic)
3/5
Readability (diagnostic)
4/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
57 words preserved; 2 selected events serialized; printed elapsed budget 24.770 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English5 selected intervals
The Vampire and the Innocent Villager
en-07830 · 27.02 s · Voice: Despina · laions_got_talent
2.95 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A single speaker with a young-adult feminine-coded voice speaks in a gentle, soothing, and deeply affectionate tone in General American English. The delivery is slow, intimate, and quiet, bordering on an ASMR or lullaby style, characterized by soft attacks, warm low-mid resonance, and a breathy, half-whispered quality. The speaker exhibits high vulnerability and deep emotional care, offering comfort and reassurance to a child. The pacing is deliberate with generous pauses, punctuated by soft, audible sighs of relief and tenderness. The recording is pristine with a dead-quiet background and close-mic intimacy, creating an exceptionally genuine and comforting performance.
SCRIPT — timed speech, local cues and free-form burst descriptions
[0.32 seconds pause]
(very asmr, very hi-fi, very sexual lust, enjoyable; 1.11 seconds duration) Step inside,
[0.49 seconds pause]
(very asmr, very hi-fi, very sexual lust, enjoyable; 0.89 seconds duration) out of the fog,
[0.31 seconds pause]
(very asmr, very hi-fi, very sexual lust, enjoyable; 0.21 seconds duration) child.
[0.51 seconds pause]
(very hi-fi, very asmr, very warmth, enjoyable; 2.28 seconds duration) It is warm by the hearth. ("A soft, breathy vocalized sigh of relief and comfort occurs in the pause before the second speaker responds, blending smoothly into the gentle, warm tone of the dialogue.", 0.90 seconds)
(very organic, very thin, very thankfulness; 0.80 seconds duration) Oh ("A soft, breathy vocalized sigh of relief and gratitude ('Oh') uttered by the second speaker right before thanking the host, blending smoothly into the spoken dialogue.", 0.92 seconds)
(very organic, very thin, very thankfulness; 0.91 seconds duration) thank you, sir.
(very fading, fading, stoicism, enjoyable; 1.39 seconds duration) My hands were nearly numb.
[0.86 seconds pause]
(very clean-bg, very smooth, very feeling turned on, enjoyable; 1.52 seconds duration) Let me take your coat.
(overlapping non-speech vocal events; 2.14 seconds duration) ("A deep, warm, and prolonged exhalation of relief and comfort occurs in the pause after offering to take the coat, releasing breath with a soft, relaxed vocalized sigh that naturally bridges the two speakers' dialogue.", 1.30 seconds) ("A soft, prolonged, breathy sigh of relief and comfort occurs in the pause before 'You are trembling', releasing gentle air with warm, relaxed vocal tract resonance.", 1.10 seconds)
[0.34 seconds pause]
(very whispered, very asmr, very fondness, enjoyable; 1.32 seconds duration) You are trembling.
[0.52 seconds pause]
(very hi-fi, very explicit, very lust, enjoyable; 2.28 seconds duration) Yet your pulse runs so fast.
(very whispered, very clean-bg, very pining, enjoyable; 2.87 seconds duration) ("A soft, audible inhalation taken in the pause before 'The woods were just so dark tonight', reflecting a moment of quiet emotional grounding and concern.", 0.80 seconds) The woods were just so dark to night.
[0.69 seconds pause]
(very non-narrative, asmr, slightly solace; 0.96 seconds duration) Hush now.
[0.32 seconds pause]
(very asmr, very clean-bg, very carnal desire, enjoyable; 1.44 seconds duration) Rest your head.
[0.64 seconds pause]
(very relaxed, very asmr, very respite, enjoyable; 1.49 seconds duration) You are quite safe here.
[0.41 seconds pause]
Automated review note The speaker performs a gentle, breathy dialogue with ASMR/whispered intimacy. However, multiple local segments are severely mislabeled with tags related to sexual lust, carnal desire, and explicit sexual arousal that are completely unsupported by the audible comforting/caretaking tone.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
0.90 s duration · HVA_EVENT_0000
A soft, breathy vocalized sigh of relief and comfort occurs in the pause before the second speaker responds, blending smoothly into the gentle, warm tone of the dialogue.
A soft, breathy vocalized sigh of relief and gratitude ('Oh') uttered by the second speaker right before thanking the host, blending smoothly into the spoken dialogue.
A deep, warm, and prolonged exhalation of relief and comfort occurs in the pause after offering to take the coat, releasing breath with a soft, relaxed vocalized sigh that naturally bridges the two speakers' dialogue.
A soft, prolonged, breathy sigh of relief and comfort occurs in the pause before 'You are trembling', releasing gentle air with warm, relaxed vocal tract resonance.
A soft, audible inhalation taken in the pause before 'The woods were just so dark tonight', reflecting a moment of quiet emotional grounding and concern.
Local emotion and delivery review — 14 speech segments
Absolute interval
Recognized speech and cue
Gemini ratings
Reviewer explanation
segment-001 Source sentence 1
Step inside, very asmr, very hi-fi, very sexual lust, enjoyable
Emotion 1/5 Delivery 4/5
Tagged as 'sexual lust', but the delivery is gentle, warm, and soothing caretaking.
segment-002 Source sentence 1
out of the fog, very asmr, very hi-fi, very sexual lust, enjoyable
Emotion 1/5 Delivery 4/5
Tagging as sexual lust is completely ungrounded in the gentle, protective delivery.
segment-003 Source sentence 1
child. very asmr, very hi-fi, very sexual lust, enjoyable
Emotion 1/5 Delivery 3/5
Labeling addressing a child with sexual lust is completely inaccurate.
segment-004 Source sentence 2
It is warm by the hearth. very hi-fi, very asmr, very warmth, enjoyable
Emotion 4/5 Delivery 4/5
Warmth and ASMR intimate delivery are well supported.
segment-005 Source sentence 3
Oh very organic, very thin, very thankfulness
Emotion 3/5 Delivery 3/5
'Oh' is vocalized breathily; 'thin' delivery matches the light head voice.
segment-006 Source sentence 3
thank you, sir. very organic, very thin, very thankfulness
Emotion 4/5 Delivery 4/5
Gratitude and soft tone are well captured.
segment-007 Source sentence 4
My hands were nearly numb. very fading, fading, stoicism, enjoyable
Emotion 3/5 Delivery 3/5
Vulnerable shivering tone, stoicism is somewhat misrepresentative of the cold/shivering affect.
segment-008 Source sentence 5
Let me take your coat. very clean-bg, very smooth, very feeling turned on, enjoyable
Emotion 1/5 Delivery 4/5
'Feeling turned on' is an inappropriate projection onto a gentle maternal/caretaking delivery.
segment-009 Source sentence 6
You are trembling. very whispered, very asmr, very fondness, enjoyable
Emotion 3/5 Delivery 4/5
Fondness/concern is audible; whispered ASMR delivery is accurate.
segment-010 Source sentence 7
Yet your pulse runs so fast. very hi-fi, very explicit, very lust, enjoyable
Emotion 1/5 Delivery 3/5
Labeling as 'explicit lust' is unsupported by the acoustic delivery, which remains soft and gentle.
segment-011 Source sentence 8
The woods were just so dark to night. very whispered, very clean-bg, very pining, enjoyable
Emotion 3/5 Delivery 4/5
Soft, vulnerable, and frightened tone; longing/pining is somewhat off.
segment-012 Source sentence 9
Hush now. very non-narrative, asmr, slightly solace
Emotion 4/5 Delivery 4/5
Solace/soothing hushed tone is accurate.
segment-013 Source sentence 10
Rest your head. very asmr, very clean-bg, very carnal desire, enjoyable
Emotion 1/5 Delivery 3/5
'Carnal desire' completely misinterprets the calming lullaby cadence.
segment-014 Source sentence 11
You are quite safe here. very relaxed, very asmr, very respite, enjoyable
Emotion 4/5 Delivery 4/5
Reassurance, relief, and relaxed delivery fit well.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
3/5
Whole-recording emotion fit
3/5
Local speech emotion fit
2/5
Whole-recording talking-style fit
4/5
Local speech delivery fit
3/5
Burst-description fit
3/5
Burst-timing fit
3/5
Event precision rating
3/5
Event coverage rating
3/5
Transcript fidelity (diagnostic)
4/5
Speech-timing fit (diagnostic)
4/5
Readability (diagnostic)
4/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
54 words preserved; 5 selected events serialized; printed elapsed budget 27.020 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English2 selected intervals
Voice Actor Recording a Happy Commercial While Devastated
en-07903 · 23.42 s · Voice: Autonoe · voicenet_extension
3.00 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A feminine-coded, young-adult speaker delivers a commercial-style promotional script in clear General American English. The vocal delivery is bright, warm, and highly polished, featuring crisp oral and mask resonance with smooth legato phrasing. The tempo is steady and conversational, punctuated by deliberate pauses and a single audible breath reset. The emotional tone is consistently upbeat, friendly, and optimistic, projecting an inviting, professional commercial presence with pristine studio recording quality and no background noise.
SCRIPT — timed speech, local cues and free-form burst descriptions
[0.24 seconds pause]
(very bright, very clean-bg, ease, enjoyable; 0.79 seconds duration) Summer fun
[0.25 seconds pause]
(very bright, very clean-bg, ease, enjoyable; 1.74 seconds duration) starts at Sandy's Water Park.
[0.34 seconds pause]
(very clean-bg, very tonal, very sympathy, enjoyable; 1.26 seconds duration) Bring the whole family.
[0.58 seconds pause]
(very tonal, very clean-bg, very reverence, enjoyable; 1.23 seconds duration) More energy on family?
[0.61 seconds pause]
(very accelerating, very non-narrative, lightness; 1.12 seconds duration) Yeah, copy that.
(non-narrative, explicit, slightly fulfillment; 0.28 seconds duration) Just
[0.28 seconds pause]
(non-narrative, explicit, slightly fulfillment; 0.96 seconds duration) give me one second.
[0.24 seconds pause]
(very explicit, very fading, slightly indifference; 1.04 seconds duration) Reset the cans.
("A soft, audible exhalation of breath occurs during a pause between taking notes and preparing to reread the script, functioning as a natural reset and release of tension.", 1.04 seconds)
[0.02 seconds pause]
("A soft, audible exhalation of relief and reset occurs during the pause between taking a break and restarting the take, naturally bridging the two spoken segments.", 0.86 seconds)
[0.72 seconds pause]
(very bright, very thin-throat, serenity, enjoyable; 0.79 seconds duration) Summer fun
[0.25 seconds pause]
(very bright, very thin-throat, serenity, enjoyable; 1.75 seconds duration) starts at Sandy's Water Park.
[0.33 seconds pause]
(very bright, very tonal, very tenderness, enjoyable; 1.98 seconds duration) Bring the whole family for slides,
[0.26 seconds pause]
(very bright, very tonal, very tenderness, enjoyable; 2.36 seconds duration) sunshine, and memories that last forever.
[0.60 seconds pause]
(very clean-content, very tonal, very self-consciousness, enjoyable; 1.15 seconds duration) Rolling whenever you are.
[0.35 seconds pause]
Automated review note Script captures both promotional takes and conversational outtakes, but misinterprets technical interjections as exaggerated emotional states.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
1.04 s duration · HVA_EVENT_0000
A soft, audible exhalation of breath occurs during a pause between taking notes and preparing to reread the script, functioning as a natural reset and release of tension.
A soft, audible exhalation of relief and reset occurs during the pause between taking a break and restarting the take, naturally bridging the two spoken segments.
Local emotion and delivery review — 13 speech segments
Absolute interval
Recognized speech and cue
Gemini ratings
Reviewer explanation
segment-001 Source sentence 1
Summer fun very bright, very clean-bg, ease, enjoyable
Emotion 3/5 Delivery 3/5
Commercial read starts with cheerful delivery.
segment-002 Source sentence 1
starts at Sandy's Water Park. very bright, very clean-bg, ease, enjoyable
Emotion 3/5 Delivery 3/5
Maintains cheerful, bright commercial tone.
segment-003 Source sentence 2
Bring the whole family. very clean-bg, very tonal, very sympathy, enjoyable
Emotion 3/5 Delivery 3/5
Tags overstate affection; tone is standard promo read.
segment-004 Source sentence 3
More energy on family? very tonal, very clean-bg, very reverence, enjoyable
Emotion 2/5 Delivery 3/5
Spoken aside asking for feedback, not reverence.
segment-005 Source sentence 4
Yeah, copy that. very accelerating, very non-narrative, lightness
Emotion 3/5 Delivery 3/5
Casual studio acknowledgment.
segment-006 Source sentence 5
Just non-narrative, explicit, slightly fulfillment
Emotion 3/5 Delivery 3/5
Brief transitional filler, neutral tone.
segment-007 Source sentence 5
give me one second. non-narrative, explicit, slightly fulfillment
Emotion 3/5 Delivery 3/5
Standard conversational pause for headphones.
segment-008 Source sentence 6
Reset the cans. very explicit, very fading, slightly indifference
Emotion 3/5 Delivery 3/5
Indifference label is mildly unfitting for functional aside.
segment-009 Source sentence 7
Summer fun very bright, very thin-throat, serenity, enjoyable
Emotion 4/5 Delivery 3/5
Restart of the read with renewed energy.
segment-010 Source sentence 7
starts at Sandy's Water Park. very bright, very thin-throat, serenity, enjoyable
Emotion 4/5 Delivery 3/5
Bright commercial performance.
segment-011 Source sentence 8
Bring the whole family for slides, very bright, very tonal, very tenderness, enjoyable
Emotion 3/5 Delivery 3/5
Energetic promotional style rather than deep tenderness.
segment-012 Source sentence 8
sunshine, and memories that last forever. very bright, very tonal, very tenderness, enjoyable
Emotion 3/5 Delivery 3/5
Warm commercial finish with good energy.
segment-013 Source sentence 9
Rolling whenever you are. very clean-content, very tonal, very self-consciousness, enjoyable
Emotion 2/5 Delivery 3/5
Spoken aside to engineer, not self-conscious.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
3/5
Whole-recording emotion fit
3/5
Local speech emotion fit
3/5
Whole-recording talking-style fit
3/5
Local speech delivery fit
3/5
Burst-description fit
3/5
Burst-timing fit
3/5
Event precision rating
3/5
Event coverage rating
3/5
Transcript fidelity (diagnostic)
4/5
Speech-timing fit (diagnostic)
4/5
Readability (diagnostic)
4/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
49 words preserved; 2 selected events serialized; printed elapsed budget 23.420 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English3 selected intervals
The Secret Ingredient
en-08566 · 25.54 s · Voice: Charon · situation_inspired
2.85 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A mature adult male speaker delivers an instructional, mildly exasperated cooking tip in General American English. The vocal delivery is grounded in a deep baritone register with noticeable chest resonance, balanced warmth, and a touch of natural vocal fry. The pacing is deliberate and conversational, marked by clear articulation and punchy emphasis on key words like 'cooking', 'magic', and 'nutty butter'. A prominent exasperated sigh occurs midway through the clip, naturally bridging the initial reprimand with the subsequent practical advice. The recording is studio-clean with negligible background noise and balanced acoustic presence.
SCRIPT — timed speech, local cues and free-form burst descriptions
[0.24 seconds pause]
(very denasal, very full, very amazement, enjoyable; 1.27 seconds duration) Because it's cooking,
[0.25 seconds pause]
(very denasal, very full, very amazement, enjoyable; 0.22 seconds duration) kid,
[0.50 seconds pause]
(very denasal, very full, very amazement, enjoyable; 0.15 seconds duration) not
[0.25 seconds pause]
(very denasal, very full, very amazement, enjoyable; 0.45 seconds duration) magic.
[0.51 seconds pause]
(very clean-content, very full, slightly exasperation, enjoyable; 1.73 seconds duration) Stop asking why the butter foams.
[0.41 seconds pause]
(overlapping non-speech vocal events; 1.54 seconds duration) ("A deep, audible exasperated sigh occurs during a pause between instructional phrases, featuring a heavy unvoiced exhalation that conveys mild impatience and fatigue before resuming speech.", 1.18 seconds) ("A soft, audible exhalation of breath occurs during a pause between phrases, conveying mild exasperation and fatigue before resuming instruction.", 0.64 seconds)
(chesty, non-narrative, slightly fulfillment, enjoyable; 1.68 seconds duration) Look, if you actually want to
[0.32 seconds pause]
(chesty, non-narrative, slightly fulfillment, enjoyable; 0.58 seconds duration) know, ("A quick, audible preparatory intake of breath occurs during a pause between phrases, serving as a natural respiratory reset before delivering the next instruction.", 0.58 seconds)
[0.54 seconds pause]
(chesty, non-narrative, slightly fulfillment, enjoyable; 1.92 seconds duration) don't turn the heat down yet.
[1.12 seconds pause]
(naturalistic, non-narrative, slightly relaxation, enjoyable; 1.62 seconds duration) See those brown flecks around the edge?
[0.70 seconds pause]
(very news-anchor, very even-rhythm, enjoyable; 3.20 seconds duration) Your grandfather used to toss the crushed garlic in right then.
[0.40 seconds pause]
(very accelerating, very non-narrative, very apathy; 0.79 seconds duration) Just ten seconds.
[1.05 seconds pause]
(very full, very denasal, very shock, enjoyable; 3.74 seconds duration) It hits the nutty butter and suddenly you smell the whole room change.
[0.36 seconds pause]
Automated review note Acoustic style descriptions are reasonable, but many segment emotion tags like amazement or shock do not match the dry instructional delivery.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
1.18 s duration · HVA_EVENT_0000
A deep, audible exasperated sigh occurs during a pause between instructional phrases, featuring a heavy unvoiced exhalation that conveys mild impatience and fatigue before resuming speech.
A quick, audible preparatory intake of breath occurs during a pause between phrases, serving as a natural respiratory reset before delivering the next instruction.
don't turn the heat down yet. chesty, non-narrative, slightly fulfillment, enjoyable
Emotion 3/5 Delivery 3/5
Instructional tone rather than fulfillment.
segment-009 Source sentence 4
See those brown flecks around the edge? naturalistic, non-narrative, slightly relaxation, enjoyable
Emotion 3/5 Delivery 4/5
Attentive question, fits relaxed delivery.
segment-010 Source sentence 5
Your grandfather used to toss the crushed garlic in right then. very news-anchor, very even-rhythm, enjoyable
Emotion 3/5 Delivery 2/5
Nostalgic instructional delivery, not news-anchor.
segment-011 Source sentence 6
Just ten seconds. very accelerating, very non-narrative, very apathy
Emotion 2/5 Delivery 3/5
Deliberate emphasis rather than apathy or rushing.
segment-012 Source sentence 7
It hits the nutty butter and suddenly you smell the whole room change. very full, very denasal, very shock, enjoyable
Emotion 2/5 Delivery 3/5
Evocative closing line, not shock.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
4/5
Whole-recording emotion fit
3/5
Local speech emotion fit
2/5
Whole-recording talking-style fit
3/5
Local speech delivery fit
3/5
Burst-description fit
3/5
Burst-timing fit
2/5
Event precision rating
3/5
Event coverage rating
3/5
Transcript fidelity (diagnostic)
4/5
Speech-timing fit (diagnostic)
3/5
Readability (diagnostic)
4/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
59 words preserved; 3 selected events serialized; printed elapsed budget 25.540 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English5 selected intervals
Laughing Through Tears as You Toast Your Ex at Their Wedding
en-09242 · 24.94 s · Voice: Aoede · existing_challenge_inspired
3.05 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A young adult female speaker delivers an intimate, emotionally vulnerable confession in General American English. Her voice is warm, breathy, and moderately soft, characterized by a relaxed tempo, gentle attacks, and natural conversational pauses. The performance begins with a wry, affectionate smile and a soft chuckle as she recalls an old argument, before shifting into a deeply tender, melancholic, and bittersweet tone. The acoustic environment is clean and dry with negligible background noise, highlighting the close-mic, confessional quality of the delivery.
SCRIPT — timed speech, local cues and free-form burst descriptions
(clean-bg, dialogic, tartness, enjoyable; 2.81 seconds duration) I still take full credit for that awful trivia night,
[0.39 seconds pause]
(clean-bg, dialogic, tartness, enjoyable; 1.08 seconds duration) you know? ("A soft, quiet chuckle occurs in the pause following 'you know?', featuring gentle breathy pulses that reflect fond amusement and affectionate teasing before continuing the sentence.", 0.80 seconds)
[0.52 seconds pause]
(very bright, very oral-bright, very ribbing, enjoyable; 3.76 seconds duration) Seeing you two argue over geography, I just knew. ("A brief, quiet chuckle escapes right after 'I just knew', featuring a soft breathy release that blends naturally into the reflective pause before continuing the sentence.", 0.60 seconds) ("A soft, quiet chuckle with gentle breathy pulses occurs right after 'I just knew', expressing fond amusement and affectionate teasing before continuing into the next sentence.", 0.88 seconds)
(very decelerating, very heady, jesting, enjoyable; 2.12 seconds duration) It was so obvious right away.
[0.84 seconds pause]
(very heady, oral-bright, very awe, enjoyable; 2.43 seconds duration) And watching you light up around them ever since,
[1.01 seconds pause]
(very heady, oral-bright, very awe, enjoyable; 0.16 seconds duration) God
[0.32 seconds pause]
(very heady, oral-bright, very awe, enjoyable; 0.32 seconds duration) has been
[0.24 seconds pause]
(very heady, oral-bright, very awe, enjoyable; 0.30 seconds duration) something.
[0.98 seconds pause]
(very decelerating, very organic, very compassion, enjoyable; 1.92 seconds duration) I really am happy for you.
[1.36 seconds pause]
(very vulnerable, very non-instructional, very fondness, enjoyable; 4.06 seconds duration) I just hope they love you the way I never could. ("A soft, trailing breathy sigh of resignation and bittersweet release follows the final phrase 'the way I never could', fading out into silence.", 0.78 seconds) ("A soft, trailing exhalation of wistful resignation and bittersweet release follows the final phrase 'the way I never could', fading out into silence.", 1.00 seconds)
[0.32 seconds pause]
Automated review note Annotation covers the recording reasonably well, though burst events contain duplicates and some segments are over-fragmented.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
0.80 s duration · HVA_EVENT_0000
A soft, quiet chuckle occurs in the pause following 'you know?', featuring gentle breathy pulses that reflect fond amusement and affectionate teasing before continuing the sentence.
Top-three type softmax: Chuckle: 18.6%; Snort: 13.4%; Breathy Giggle: 10.7%
0.60 s duration · HVA_EVENT_0001
A brief, quiet chuckle escapes right after 'I just knew', featuring a soft breathy release that blends naturally into the reflective pause before continuing the sentence.
Top-three type softmax: Chuckle: 39.6%; Breathy Giggle: 13.4%; Snicker: 11.4%
0.88 s duration · HVA_EVENT_0002
A soft, quiet chuckle with gentle breathy pulses occurs right after 'I just knew', expressing fond amusement and affectionate teasing before continuing into the next sentence.
Top-three type softmax: Chuckle: 29.8%; Breathy Giggle: 27.9%; Snicker: 10.2%
0.78 s duration · HVA_EVENT_0003
A soft, trailing breathy sigh of resignation and bittersweet release follows the final phrase 'the way I never could', fading out into silence.
Local emotion and delivery review — 10 speech segments
Absolute interval
Recognized speech and cue
Gemini ratings
Reviewer explanation
segment-001 Source sentence 1
I still take full credit for that awful trivia night, clean-bg, dialogic, tartness, enjoyable
Emotion 3/5 Delivery 3/5
Tone is friendly rather than displaying tartness.
segment-002 Source sentence 1
you know? clean-bg, dialogic, tartness, enjoyable
Emotion 3/5 Delivery 3/5
Short question followed by a light chuckle; emotion label is slightly mismatched.
segment-003 Source sentence 2
Seeing you two argue over geography, I just knew. very bright, very oral-bright, very ribbing, enjoyable
Emotion 3/5 Delivery 3/5
Playful teasing fits moderately well.
segment-004 Source sentence 3
It was so obvious right away. very decelerating, very heady, jesting, enjoyable
Emotion 3/5 Delivery 3/5
Amusement tag reasonably captures the tone.
segment-005 Source sentence 4
And watching you light up around them ever since, very heady, oral-bright, very awe, enjoyable
Emotion 3/5 Delivery 3/5
Awe is somewhat overstated for this tender phrase.
segment-006 Source sentence 4
God very heady, oral-bright, very awe, enjoyable
Emotion 3/5 Delivery 3/5
Single word segment split artificially; emotion fits broad context.
segment-007 Source sentence 4
has been very heady, oral-bright, very awe, enjoyable
Emotion 3/5 Delivery 3/5
Short fragmented segment carries minimal standalone affect.
segment-008 Source sentence 4
something. very heady, oral-bright, very awe, enjoyable
Emotion 3/5 Delivery 3/5
Fragmented word segment with moderate delivery fit.
segment-009 Source sentence 5
I really am happy for you. very decelerating, very organic, very compassion, enjoyable
Emotion 3/5 Delivery 3/5
Sincere affection matches the spoken sentiment.
segment-010 Source sentence 6
I just hope they love you the way I never could. very vulnerable, very non-instructional, very fondness, enjoyable
Emotion 3/5 Delivery 3/5
Vulnerable and bittersweet tone is generally captured.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
4/5
Whole-recording emotion fit
3/5
Local speech emotion fit
3/5
Whole-recording talking-style fit
3/5
Local speech delivery fit
3/5
Burst-description fit
3/5
Burst-timing fit
2/5
Event precision rating
3/5
Event coverage rating
3/5
Transcript fidelity (diagnostic)
4/5
Speech-timing fit (diagnostic)
3/5
Readability (diagnostic)
3/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
57 words preserved; 5 selected events serialized; printed elapsed budget 24.940 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English9 selected intervals
The Diagnosis Morning: Dancing While the World Ends
en-09554 · 34.46 s · Voice: Callirrhoe · existing_challenge_inspired
3.45 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A young adult feminine-coded speaker delivers a casual, playful, and slightly exasperated monologue in General American English. The vocal delivery is warm, intimate, and conversational, characterized by clear oral resonance, moderate pitch variety, and natural pacing. The speaker alternates between lighthearted teasing and mild, humorous frustration regarding a simple cup of coffee. The recording is studio-clean with virtually no background noise, capturing subtle breaths, soft chuckles, and an exasperated sigh that blend seamlessly into the spoken performance.
SCRIPT — timed speech, local cues and free-form burst descriptions
[0.32 seconds pause]
(very thin-throat, very hi-fi, very awe, enjoyable; 1.51 seconds duration) Look at that perfect
[0.25 seconds pause]
(very thin-throat, very hi-fi, very awe, enjoyable; 0.86 seconds duration) golden crust.
[0.80 seconds pause]
("A soft, amused chuckle occurs in the pause between 'golden crust' and 'Don't start', featuring gentle breathy pulses that naturally lead into the smiling, teasing tone of the following sentence.", 0.62 seconds)
(very bright, tonal, enjoyable; 1.96 seconds duration) Don't start about the sourdough crumbs.
(very non-instructional, very effortless, very having a crush, enjoyable; 2.32 seconds duration) You know you love the mess. ("A soft, quiet chuckle with gentle breathy pulses follows 'you know you love the mess', expressing fond amusement and affectionate teasing.", 0.80 seconds)
[0.24 seconds pause]
(very non-narrative, very oral-bright, very provoking lightly; 0.88 seconds duration) Two sugars,
[0.24 seconds pause]
(very non-narrative, very oral-bright, very provoking lightly; 0.32 seconds duration) right?
[0.56 seconds pause]
(very clean-bg, very hi-fi, slightly calmness, enjoyable; 1.03 seconds duration) Exactly forty
[0.25 seconds pause]
(very clean-bg, very hi-fi, slightly calmness, enjoyable; 0.30 seconds duration) stirs.
[1.86 seconds pause]
(very asmr, very vulnerable, very thoughtfulness, enjoyable; 0.46 seconds duration) You know,
[0.66 seconds pause]
(very asmr, very vulnerable, very thoughtfulness, enjoyable; 0.16 seconds duration) if
[0.24 seconds pause]
(very asmr, very vulnerable, very thoughtfulness, enjoyable; 1.98 seconds duration) someone offered me all of history,
[0.74 seconds pause]
(very asmr, very vulnerable, very thoughtfulness, enjoyable; 1.96 seconds duration) every sunrise that ever happened,
[1.24 seconds pause]
(very asmr, very vulnerable, very thoughtfulness, enjoyable; 1.76 seconds duration) I'd stay right here. ("A soft, quiet exhale of contentment and peaceful surrender right after 'I'd stay right here', blending seamlessly into the reflective pause.", 0.54 seconds)
[0.34 seconds pause]
(overlapping non-speech vocal events; 1.60 seconds duration) ("A soft, contented sigh of relief and peaceful surrender occurs in the pause before 'Just this counter', releasing breath with warm, relaxed vocal fold relaxation.", 0.94 seconds) ("A soft, contented sigh of relief and peaceful surrender occurs in the pause before 'Just this counter', releasing breath with warm, relaxed vocal fold relaxation.", 0.68 seconds)
[1.10 seconds pause]
(very fading, very whispered, very indifference; 1.01 seconds duration) Just this counter.
[0.67 seconds pause]
(very hi-fi, very clean-bg, very revulsion, enjoyable; 0.24 seconds duration) This
[0.24 seconds pause]
(very hi-fi, very clean-bg, very revulsion, enjoyable; 1.90 seconds duration) dumb chipped mug. ("A tiny, suppressed nasal snort or quiet chuckle right before 'Forever', conveying self-deprecating amusement at the mundane objects being cherished.", 0.54 seconds) ("A tiny, suppressed nasal chuckle right before 'Forever', conveying affectionate amusement and cozy intimacy.", 0.50 seconds)
[0.34 seconds pause]
(very non-narrative, very choppy, slightly boredom, less-enjoyable; 0.56 seconds duration) Forever.
(very hi-fi, oral-bright, slightly enthusiasm; 2.50 seconds duration) ("A soft, quiet inhalation taken in the pause before the final practical reminder 'Drink it before it gets cold.'", 0.78 seconds) ("A quick, quiet preparatory breath taken right before the final practical instruction 'Drink it before it gets cold.'", 0.62 seconds) Drink it before it gets cold.
[0.44 seconds pause]
Automated review note Several burst events are duplicated/overlapping in the script annotations (such as double sighs and double prep breaths), inflating the event list. Additionally, segments 15-16 are mislabeled as 'revulsion/disgust' despite being clearly tender and affectionate.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
0.62 s duration · HVA_EVENT_0000
A soft, amused chuckle occurs in the pause between 'golden crust' and 'Don't start', featuring gentle breathy pulses that naturally lead into the smiling, teasing tone of the following sentence.
A soft, contented sigh of relief and peaceful surrender occurs in the pause before 'Just this counter', releasing breath with warm, relaxed vocal fold relaxation.
A soft, contented sigh of relief and peaceful surrender occurs in the pause before 'Just this counter', releasing breath with warm, relaxed vocal fold relaxation.
Local emotion and delivery review — 18 speech segments
Absolute interval
Recognized speech and cue
Gemini ratings
Reviewer explanation
segment-001 Source sentence 1
Look at that perfect very thin-throat, very hi-fi, very awe, enjoyable
Emotion 4/5 Delivery 4/5
Tone is appreciative and gentle.
segment-002 Source sentence 1
golden crust. very thin-throat, very hi-fi, very awe, enjoyable
Emotion 4/5 Delivery 4/5
Maintains appreciative tone.
segment-003 Source sentence 2
Don't start about the sourdough crumbs. very bright, tonal, enjoyable
Emotion 4/5 Delivery 4/5
Playful and lightly teasing.
segment-004 Source sentence 3
You know you love the mess. very non-instructional, very effortless, very having a crush, enjoyable
Emotion 4/5 Delivery 4/5
Affectionate and teasing with an audible laugh at the end.
segment-005 Source sentence 4
Two sugars, very non-narrative, very oral-bright, very provoking lightly
Emotion 4/5 Delivery 4/5
Casual questioning.
segment-006 Source sentence 4
right? very non-narrative, very oral-bright, very provoking lightly
Emotion 4/5 Delivery 4/5
Light tag question.
segment-007 Source sentence 5
Exactly forty very clean-bg, very hi-fi, slightly calmness, enjoyable
Emotion 4/5 Delivery 4/5
Direct, calm remark.
segment-008 Source sentence 5
stirs. very clean-bg, very hi-fi, slightly calmness, enjoyable
Emotion 4/5 Delivery 4/5
Direct conclusion to the phrase.
segment-009 Source sentence 6
You know, very asmr, very vulnerable, very thoughtfulness, enjoyable
Emotion 4/5 Delivery 4/5
Intimate, thoughtful opening.
segment-010 Source sentence 6
if very asmr, very vulnerable, very thoughtfulness, enjoyable
Emotion 3/5 Delivery 4/5
Very short conjunction, thoughtful context.
segment-011 Source sentence 6
someone offered me all of history, very asmr, very vulnerable, very thoughtfulness, enjoyable
Emotion 4/5 Delivery 4/5
Soft, reflective, tender tone.
segment-012 Source sentence 6
every sunrise that ever happened, very asmr, very vulnerable, very thoughtfulness, enjoyable
Emotion 4/5 Delivery 4/5
Soft, intimate delivery.
segment-013 Source sentence 6
I'd stay right here. very asmr, very vulnerable, very thoughtfulness, enjoyable
Emotion 4/5 Delivery 4/5
Affectionate and grounded.
segment-014 Source sentence 7
Just this counter. very fading, very whispered, very indifference
Emotion 2/5 Delivery 3/5
Tagged as indifference, but delivery is tender and fond rather than numb.
segment-015 Source sentence 8
This very hi-fi, very clean-bg, very revulsion, enjoyable
Emotion 1/5 Delivery 3/5
Tagged as revulsion/disgust, but tone is warmly affectionate and teasing.
segment-016 Source sentence 8
dumb chipped mug. very hi-fi, very clean-bg, very revulsion, enjoyable
Emotion 1/5 Delivery 3/5
Tagged as revulsion/disgust, which completely misidentifies the fond sentiment.
segment-017 Source sentence 9
Forever. very non-narrative, very choppy, slightly boredom, less-enjoyable
Emotion 2/5 Delivery 3/5
Tagged as boredom, but tone is wistful and devoted.
segment-018 Source sentence 10
Drink it before it gets cold. very hi-fi, oral-bright, slightly enthusiasm
Emotion 4/5 Delivery 4/5
Gently teasing and grounded practical delivery.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
4/5
Whole-recording emotion fit
4/5
Local speech emotion fit
3/5
Whole-recording talking-style fit
4/5
Local speech delivery fit
3/5
Burst-description fit
3/5
Burst-timing fit
3/5
Event precision rating
3/5
Event coverage rating
4/5
Transcript fidelity (diagnostic)
5/5
Speech-timing fit (diagnostic)
4/5
Readability (diagnostic)
4/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
56 words preserved; 9 selected events serialized; printed elapsed budget 34.460 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English9 selected intervals
Megaphone Protest Leader: Voice Giving Out
en-10100 · 20.09 s · Voice: Schedar · voicenet_extension
3.35 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A masculine-coded young adult speaker delivers an impassioned, escalating motivational speech in General American English. The performance begins with an agitated, questioning tone marked by sharp, punchy attacks and moderate vocal tension, before building into a commanding, high-arousal rallying cry. The vocal timbre is grounded in chest and throat resonance with a gritty, pressed edge that intensifies as volume and dynamic force surge. Phrasing is segmented and emphatic, punctuated by an exasperated sigh and a sharp preparatory inhale that seamlessly feed into the escalating intensity. The recording is studio-clean with minimal background interference, capturing a dramatic, highly focused, and theatrical delivery.
SCRIPT — timed speech, local cues and free-form burst descriptions
(very metallic, very swelling, very perplexity; 0.95 seconds duration) What do we want?
[0.33 seconds pause]
(very clean-bg, very clean-content; 0.48 seconds duration) Tell 'em.
[0.32 seconds pause]
(very sharp-onset, very scattered, very confusion; 0.62 seconds duration) What do we
("A sharp, audible intake of breath occurs right after the speaker stumbles on 'What do we...', signaling sudden realization and preparation before pausing to address the dead battery.", 0.54 seconds)
(very sharp-onset, very scattered, very confusion; 0.84 seconds duration) ("A sharp, audible intake of breath occurs right after the speaker stumbles on 'What do we...', signaling sudden realization and frustration before pausing to say 'Wait'.", 0.60 seconds) wait.
(overlapping non-speech vocal events; 0.88 seconds duration) ("A sharp, audible intake of breath through the mouth occurs right after the speaker says 'Wait', signaling sudden realization and physical preparation before muttering about the dead battery.", 0.56 seconds) ("A sharp, audible intake of breath through the mouth and nose occurs right after the speaker says 'Wait', signaling sudden realization and physical preparation before muttering about the dead battery.", 0.60 seconds)
(very darkening, very thin-mask, very repulsion; 0.46 seconds duration) Damn
[0.26 seconds pause]
(very darkening, very thin-mask, very repulsion; 0.24 seconds duration) battery.
[0.40 seconds pause]
(very flat, very cold, very sullenness; 0.44 seconds duration) It's dead.
(very non-narrative, very accelerating, enjoyable; 1.56 seconds duration) ("A sharp, audible intake of breath through the mouth immediately preceding the urgent command 'Listen to me', signaling physical preparation and heightened arousal.", 0.48 seconds) Listen ("A quick, audible preparatory intake of breath through the mouth immediately preceding the urgent command 'Listen to me', signaling a sudden surge of focus and physical readiness.", 0.36 seconds) to me.
(very clean-bg, clean-content, insensitivity, enjoyable; 1.39 seconds duration) We don't need the speaker.
(very clean-bg, very sharp-onset, very honor; 0.92 seconds duration) We don't need the power.
(very swelling, very energetic, very disorientation; 1.87 seconds duration) ("A quick, audible gasp-like inhalation taken through the mouth immediately before shouting 'What do we want?', fueling the sudden surge in vocal projection.", 0.44 seconds) What do we want? ("A sharp, audible intake of air taken right before shouting 'Justice!', fueling the vocal projection and demonstrating physical exertion.", 0.48 seconds)
(very effortless, very sharp-onset, very disapproval; 0.85 seconds duration) Justice.
(very swelling, very energetic, very skepticism; 1.69 seconds duration) When do we want it? ("A quick, audible gasp-like inhalation taken between the shouted questions 'When do we want it?' and 'Now!', fueling the escalating vocal projection.", 0.46 seconds)
(very metallic, very commanding, slightly honor; 0.58 seconds duration) Now!
[0.30 seconds pause]
(very sharp-onset, very metallic, slightly rage; 0.56 seconds duration) Louder.
(very clean-content, very tonal, very dignity; 0.79 seconds duration) Everyone together.
(very clean-bg, very commanding, enjoyable; 1.73 seconds duration) Let them hear us across the whole street.
[0.20 seconds pause]
(very commanding, very metallic, very honor; 0.55 seconds duration) Right now.
[0.34 seconds pause]
Automated review note The proposed burst events contain several duplicates covering identical or overlapping time intervals (e.g., 3.26-3.86 overlaps with the word 'wait' at 3.24-4.08, and 4.08-4.64 overlaps with 4.36-4.96). In reality, the sound after 'wait' is an exasperated sigh/exhale rather than two separate sharp intakes. In the chants, emotions like confusion and skepticism are erroneously tagged for confident rhetorical questions.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
0.54 s duration · HVA_EVENT_0000
A sharp, audible intake of breath occurs right after the speaker stumbles on 'What do we...', signaling sudden realization and preparation before pausing to address the dead battery.
A sharp, audible intake of breath occurs right after the speaker stumbles on 'What do we...', signaling sudden realization and frustration before pausing to say 'Wait'.
A sharp, audible intake of breath through the mouth occurs right after the speaker says 'Wait', signaling sudden realization and physical preparation before muttering about the dead battery.
A sharp, audible intake of breath through the mouth and nose occurs right after the speaker says 'Wait', signaling sudden realization and physical preparation before muttering about the dead battery.
A sharp, audible intake of breath through the mouth immediately preceding the urgent command 'Listen to me', signaling physical preparation and heightened arousal.
A quick, audible preparatory intake of breath through the mouth immediately preceding the urgent command 'Listen to me', signaling a sudden surge of focus and physical readiness.
A quick, audible gasp-like inhalation taken through the mouth immediately before shouting 'What do we want?', fueling the sudden surge in vocal projection.
Local emotion and delivery review — 18 speech segments
Absolute interval
Recognized speech and cue
Gemini ratings
Reviewer explanation
segment-001 Source sentence 1
What do we want? very metallic, very swelling, very perplexity
Emotion 2/5 Delivery 3/5
The speaker shouts with rallying determination, not extreme confusion or perplexity.
segment-002 Source sentence 2
Tell 'em. very clean-bg, very clean-content
Emotion 4/5 Delivery 4/5
Short energetic call-out; appropriate restrained tag.
segment-003 Source sentence 3
What do we very sharp-onset, very scattered, very confusion
Emotion 4/5 Delivery 4/5
Trailing off abruptly due to megaphone failure, confusion tag fits.
segment-004 Source sentence 3
wait. very sharp-onset, very scattered, very confusion
Emotion 4/5 Delivery 3/5
Realization/interruption 'Wait', fits.
segment-005 Source sentence 4
Damn very darkening, very thin-mask, very repulsion
Emotion 4/5 Delivery 4/5
Muttered frustration fits repulsion/disgust tag.
segment-006 Source sentence 4
battery. very darkening, very thin-mask, very repulsion
Emotion 4/5 Delivery 4/5
Continuation of muttering.
segment-007 Source sentence 5
It's dead. very flat, very cold, very sullenness
Emotion 4/5 Delivery 4/5
Flat, sullen realization fits well.
segment-008 Source sentence 6
Listen to me. very non-narrative, very accelerating, enjoyable
Emotion 4/5 Delivery 3/5
Direct, rallying command.
segment-009 Source sentence 7
We don't need the speaker. very clean-bg, clean-content, insensitivity, enjoyable
Emotion 3/5 Delivery 3/5
Determined and firm, insensitivity tag is slightly off.
segment-010 Source sentence 8
We don't need the power. very clean-bg, very sharp-onset, very honor
Emotion 4/5 Delivery 4/5
Rising conviction and defiance.
segment-011 Source sentence 9
What do we want? very swelling, very energetic, very disorientation
Emotion 2/5 Delivery 4/5
Chanting rallying question, disorientation/confusion is inaccurate.
segment-012 Source sentence 10
Justice. very effortless, very sharp-onset, very disapproval
Emotion 3/5 Delivery 4/5
Firm response, not really contempt/disapproval towards the listener.
segment-013 Source sentence 11
When do we want it? very swelling, very energetic, very skepticism
Emotion 2/5 Delivery 4/5
Rallying call, skepticism tag is mismatched.
segment-014 Source sentence 12
Now! very metallic, very commanding, slightly honor
Emotion 4/5 Delivery 4/5
Shouted imperative 'Now!'.
segment-015 Source sentence 13
Louder. very sharp-onset, very metallic, slightly rage
Emotion 4/5 Delivery 4/5
Urgent command to the crowd.
segment-016 Source sentence 14
Everyone together. very clean-content, very tonal, very dignity
Emotion 4/5 Delivery 4/5
Rallying call, dignified/proud tone fits.
segment-017 Source sentence 15
Let them hear us across the whole street. very clean-bg, very commanding, enjoyable
Emotion 4/5 Delivery 4/5
Commanding and inspiring delivery.
segment-018 Source sentence 16
Right now. very commanding, very metallic, very honor
Emotion 4/5 Delivery 4/5
Punchy concluding command.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
4/5
Whole-recording emotion fit
4/5
Local speech emotion fit
3/5
Whole-recording talking-style fit
4/5
Local speech delivery fit
3/5
Burst-description fit
3/5
Burst-timing fit
2/5
Event precision rating
3/5
Event coverage rating
4/5
Transcript fidelity (diagnostic)
5/5
Speech-timing fit (diagnostic)
4/5
Readability (diagnostic)
4/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
51 words preserved; 9 selected events serialized; printed elapsed budget 20.090 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English8 selected intervals
Yellow Cucumber's Poison Kiss
en-10289 · 24.38 s · Voice: Sadaltager · extreme_physical
3.45 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A young adult masculine-coded speaker delivers an intensely dramatic, visceral reaction to an overwhelmingly sour and bitter food or drink. The performance is marked by heavy pharyngeal constriction, frequent breath catches, and audible physical discomfort. The vocal timbre is moderately rough, breathy, and strained, with prominent throat resonance and reduced harmonicity due to constant swallowing and gasping. The delivery is highly fragmented, featuring short, choppy phrases separated by sharp inhales, groans, and exasperated sighs. The emotional tone is dominated by acute sourness, disgust, distress, and physical pain, escalating in intensity as the speaker struggles to articulate their burning mouth. The recording is dry and clean with minimal background interference, capturing intimate mouth and breath textures that blend naturally with the actor's physical performance.
SCRIPT — timed speech, local cues and free-form burst descriptions
(very noisy, very breathy, very burnout, enjoyable; 1.12 seconds duration) Oh ("A low-pitched, breathy vocalized groan of intense craving and anticipation opens the recording, transitioning directly into the spoken phrase 'God, I want this' with natural physiological continuity.", 1.08 seconds)
(very noisy, very breathy, very burnout, enjoyable; 1.82 seconds duration) god, I want this ("A sharp, audible intake of air through the mouth and nose occurs during a pause between phrases, reflecting sensory anticipation and physical preparation.", 0.70 seconds)
[0.26 seconds pause]
(very noisy, very breathy, very burnout, enjoyable; 0.16 seconds duration) just
[0.24 seconds pause]
(very noisy, very breathy, very burnout, enjoyable; 2.28 seconds duration) the cold and the crunch
[0.68 seconds pause]
(very noisy, very breathy, very burnout, enjoyable; 1.30 seconds duration) summer in my mouth.
(very whispered, very organic, very butterflies in the stomach; 2.19 seconds duration) ("A heavy, weary exhale with slight vocalization occurs in the pause after 'mouth', conveying sensory overload and physical exhaustion.", 0.90 seconds) ("A heavy, breathy exhalation with a slight vocalized groan of sensory overload and anticipation before saying 'That looks so sweet.'", 0.84 seconds) It looks so sweet.
[1.97 seconds pause]
(very scattered, very tense, very astonishment; 1.53 seconds duration) ("A sharp, constricted gasp of sudden shock and physical distress right before exclaiming 'Oh no', marking an abrupt shift from anticipation to revulsion.", 0.80 seconds) Oh no, no, no, no.
[0.29 seconds pause]
(very non-narrative, very clean-content, very perplexity, less-enjoyable; 1.20 seconds duration) What is that?
(non-narrative, sharp-onset, very loathing; 0.84 seconds duration) ("A sharp, constricted double coughing spasm erupts in the pause between 'What is that?' and 'It tastes like straight poison', driven by sudden throat irritation and physical shock.", 0.96 seconds) Oh ("A short, sharp, constricted gasp of air taken in sudden shock and physical distress right before uttering 'It tastes like straight poison.'", 0.46 seconds)
(non-narrative, sharp-onset, very loathing; 1.66 seconds duration) tastes like straight poison.
[0.78 seconds pause]
(very clean-content, darkening, very ache; 1.16 seconds duration) My throat is burning.
[0.44 seconds pause]
(very clean-content, non-whispered; 0.70 seconds duration) I need water
[0.34 seconds pause]
(very clean-content, non-whispered; 0.92 seconds duration) right now.
(very breathy, very darkening, very exhaustion; 2.17 seconds duration) ("A sharp, audible intake of air through the mouth and throat, signaling physical shock and respiratory recovery from the intense sour taste.", 0.92 seconds) So bitter I can't even swallow.
[0.33 seconds pause]
Automated review note The general description accurately portrays the transition from anticipation to visceral revulsion and burning throat discomfort. However, the first half of the script is repeatedly mislabeled with 'burnout' (fatigue), and several proposed burst events overlap directly with each other or with spoken words.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
1.08 s duration · HVA_EVENT_0000
A low-pitched, breathy vocalized groan of intense craving and anticipation opens the recording, transitioning directly into the spoken phrase 'God, I want this' with natural physiological continuity.
A sharp, audible intake of air through the mouth and nose occurs during a pause between phrases, reflecting sensory anticipation and physical preparation.
A sharp, constricted gasp of sudden shock and physical distress right before exclaiming 'Oh no', marking an abrupt shift from anticipation to revulsion.
A sharp, constricted double coughing spasm erupts in the pause between 'What is that?' and 'It tastes like straight poison', driven by sudden throat irritation and physical shock.
tastes like straight poison. non-narrative, sharp-onset, very loathing
Emotion 4/5 Delivery 4/5
Tense disgust and pain are well captured by loathing.
segment-011 Source sentence 6
My throat is burning. very clean-content, darkening, very ache
Emotion 4/5 Delivery 4/5
Strain and pain are accurately captured.
segment-012 Source sentence 7
I need water very clean-content, non-whispered
Emotion 3/5 Delivery 3/5
Urgent and distressed request, missing emotion tag.
segment-013 Source sentence 7
right now. very clean-content, non-whispered
Emotion 3/5 Delivery 3/5
Tense and distressed, left untagged for emotion.
segment-014 Source sentence 8
So bitter I can't even swallow. very breathy, very darkening, very exhaustion
Emotion 3/5 Delivery 3/5
Expresses disgust, pain, and constriction; 'exhaustion' is an approximate fit.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
4/5
Whole-recording emotion fit
4/5
Local speech emotion fit
3/5
Whole-recording talking-style fit
4/5
Local speech delivery fit
3/5
Burst-description fit
3/5
Burst-timing fit
3/5
Event precision rating
3/5
Event coverage rating
4/5
Transcript fidelity (diagnostic)
4/5
Speech-timing fit (diagnostic)
4/5
Readability (diagnostic)
4/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
47 words preserved; 8 selected events serialized; printed elapsed budget 24.380 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English7 selected intervals
The Farmer and the Talking Crow
en-10969 · 30.89 s · Voice: Algenib · laions_got_talent
3.35 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A mature adult masculine-coded voice delivers a slow, heavily stylized, and gravelly character performance in General American English. The timbre is exceptionally deep, chest-resonant, and raspy, characterized by prominent vocal fry and low-frequency weight. The delivery is deliberately paced with short, fragmented phrasing and heavy pauses, conveying a gruff, skeptical, and mildly amused attitude. The speaker questions an absurd claim with deadpan skepticism, punctuated by a dry cough, a suppressed snicker, and a low chuckle. The recording is dry and clean with negligible background noise, capturing the textured, theatrical gravel of the voice with high fidelity.
SCRIPT — timed speech, local cues and free-form burst descriptions
(very full, very authoritative, very disdain, enjoyable; 0.67 seconds duration) Shoo,
[0.29 seconds pause]
(very full, very authoritative, very disdain, enjoyable; 1.21 seconds duration) get off that fence post.
[1.35 seconds pause]
(full, chesty, teasing, enjoyable; 0.38 seconds duration) Wait,
("A quick, audible sharp inhalation occurs in the pause between 'Wait' and 'Did you just', signaling sudden realization and surprise before questioning the other character.", 0.42 seconds)
[0.40 seconds pause]
(full, chesty, teasing, enjoyable; 2.89 seconds duration) did you just offer to weed my South Acreage?
[0.25 seconds pause]
("A dry, raspy, two-pulse throat cough occurs in the pause following 'Agric', sounding like a deliberate character cough or clearing of the throat before resuming speech.", 1.48 seconds)
[0.42 seconds pause]
(very full, very denasal, ribbing, enjoyable; 2.37 seconds duration) Crows don't pull thistle, you scoundrel.
[1.71 seconds pause]
(very clean-bg, very full, very bantering, enjoyable; 1.25 seconds duration) Shiny beetles
[0.50 seconds pause]
(very clean-bg, very full, very bantering, enjoyable; 1.34 seconds duration) and ("A brief, quiet chuckle escapes between 'Shiny beetles' and 'and dried corn', expressing amused skepticism at the absurd bartering offer.", 0.60 seconds) dried corn.
[0.43 seconds pause]
(very tonal, very accelerating, very astonishment; 0.75 seconds duration) That's your price.
[0.72 seconds pause]
(very denasal, very full, hope, enjoyable; 2.48 seconds duration) ("A low, breathy, half-amused exhalation following 'Nice try, featherhead', functioning as a dismissive, weary sigh of mild exasperation.", 0.80 seconds) Nice try, featherhead. ("A quick, audible preparatory intake of breath through the mouth and nose occurs right before the speaker delivers 'I might be bone tired', naturally resetting respiratory support for the next phrase.", 0.54 seconds)
[0.21 seconds pause]
(very denasal, very chesty, burnout, enjoyable; 2.98 seconds duration) I might be bone tired, but I'm not bargaining with a bird.
[0.69 seconds pause]
(overlapping non-speech vocal events; 1.06 seconds duration) ("A brief, low-pitched, breathy chuckle occurs in the pause before 'Though', conveying amused skepticism and smoothly leading into the final rhetorical question.", 0.80 seconds) ("A brief, low-pitched nasal snort of derisive amusement and skepticism right before uttering 'Though', blending naturally into the cynical tone of the final rhetorical question.", 0.50 seconds)
[0.63 seconds pause]
(chesty, denasal, very provoking lightly, enjoyable; 0.51 seconds duration) Though
[1.01 seconds pause]
(chesty, denasal, very provoking lightly, enjoyable; 2.21 seconds duration) you did say you'd keep the rabbits out, didn't you?
[0.28 seconds pause]
Automated review note The script contains significant burst timing mismatches and overlaps. For instance, an exhalation is tagged overlapping the start of 'Nice try' (18.828s), and around 25-26s two overlapping bursts are claimed where there is primarily one breathy chuckle. Several emotion tags (notably 'hope' on segment-009) mischaracterize the cynical/skeptical delivery.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
0.42 s duration · HVA_EVENT_0000
A quick, audible sharp inhalation occurs in the pause between 'Wait' and 'Did you just', signaling sudden realization and surprise before questioning the other character.
A dry, raspy, two-pulse throat cough occurs in the pause following 'Agric', sounding like a deliberate character cough or clearing of the throat before resuming speech.
Top-three type softmax: Cough: 88.1%; Chuckle: 3.2%; Snicker: 1.6%
0.60 s duration · HVA_EVENT_0002
A brief, quiet chuckle escapes between 'Shiny beetles' and 'and dried corn', expressing amused skepticism at the absurd bartering offer.
Top-three type softmax: Chuckle: 43.5%; Snort: 14.5%; Snorting Giggle: 7.4%
0.80 s duration · HVA_EVENT_0003
A low, breathy, half-amused exhalation following 'Nice try, featherhead', functioning as a dismissive, weary sigh of mild exasperation.
A quick, audible preparatory intake of breath through the mouth and nose occurs right before the speaker delivers 'I might be bone tired', naturally resetting respiratory support for the next phrase.
A brief, low-pitched, breathy chuckle occurs in the pause before 'Though', conveying amused skepticism and smoothly leading into the final rhetorical question.
A brief, low-pitched nasal snort of derisive amusement and skepticism right before uttering 'Though', blending naturally into the cynical tone of the final rhetorical question.
Top-three type softmax: Exasperated Sigh: 29.2%; Breath: 9.4%; Snort: 7.5%
Local emotion and delivery review — 12 speech segments
Absolute interval
Recognized speech and cue
Gemini ratings
Reviewer explanation
segment-001 Source sentence 1
Shoo, very full, very authoritative, very disdain, enjoyable
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
56 words preserved; 7 selected events serialized; printed elapsed budget 30.890 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English6 selected intervals
Whispered in the Dark
en-12110 · 26.57 s · Voice: Iapetus · eric_morris_inspired
3.40 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A single young-adult to mature-adult masculine-coded speaker delivers an intensely dramatic, confessional monologue in General American English. The vocal delivery is predominantly hushed, breathy, and close-mic'd, oscillating between a strained stage whisper and a low, raspy chest-throat resonance. The tempo is slow and heavily fragmented, marked by frequent pauses, audible gasps, and weary sighs that convey deep guilt, shame, and emotional exhaustion. Articulation remains clear despite the heavy breathiness and suppressed volume. The acoustic environment is completely clean and dry with negligible background noise, highlighting the intimate, soliloquy-like nature of the performance.
SCRIPT — timed speech, local cues and free-form burst descriptions
(very fading, very whispered, very apathy, enjoyable; 2.42 seconds duration) I keep telling myself it doesn't matter anymore.
[0.54 seconds pause]
(very clean-bg, denasal, tranquility, enjoyable; 2.64 seconds duration) That keeping quiet protects us. ("A sharp, audible intake of air through the mouth and nose occurs during a pause between phrases, reflecting emotional tension and physical preparation before continuing the confession.", 0.76 seconds)
(very breathy, very noisy, very fondness, enjoyable; 0.60 seconds duration) But ("A sharp, audible intake of air through the mouth and nose occurs right before the speaker stumbles on 'But when you look at me like that', signaling emotional hesitation and physical tension.", 0.68 seconds)
(very breathy, very noisy, very fondness, enjoyable; 2.46 seconds duration) when you look at me like that, ("A sharp, audible intake of air through the mouth and throat occurs during a pause after 'like that', signaling emotional hesitation and physical tension before confessing vulnerability.", 0.98 seconds)
[0.30 seconds pause]
(very breathy, very noisy, very fondness, enjoyable; 1.36 seconds duration) I just want you to see
[0.24 seconds pause]
(very breathy, very noisy, very fondness, enjoyable; 0.96 seconds duration) all of it.
[0.48 seconds pause]
("A slow, labored, and audible inhalation and exhalation cycle occurs in the silence following 'all of it', reflecting heavy emotional burden and physical exhaustion.", 0.92 seconds)
[0.68 seconds pause]
(very breathy, very noisy, very mortification, enjoyable; 2.66 seconds duration) God, I feel so ugly saying it out loud.
("A heavy, weary exhalation and breath release following 'saying it out loud', conveying deep self-reproach and emotional exhaustion.", 1.10 seconds)
[0.32 seconds pause]
(very denasal, very negative, very fear, enjoyable; 2.30 seconds duration) I ran away before anyone could leave me
[0.26 seconds pause]
(very denasal, very negative, very fear, enjoyable; 0.32 seconds duration) first.
("A prolonged, breathy sigh of defeat and exhaustion following 'leave me first', marking a heavy pause before the final confession.", 1.68 seconds)
[0.72 seconds pause]
(very whispered, very fading, very dejection, enjoyable; 1.50 seconds duration) I ruined that family.
[0.42 seconds pause]
(very fading, very fading, very indifference; 1.28 seconds duration) And I did it on purpose.
[0.41 seconds pause]
Automated review note Acoustic and performance descriptions are largely faithful, though some local emotion tags diverge slightly from the prevailing somber affect, and burst timings show minor overlaps with speech.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
0.76 s duration · HVA_EVENT_0000
A sharp, audible intake of air through the mouth and nose occurs during a pause between phrases, reflecting emotional tension and physical preparation before continuing the confession.
A sharp, audible intake of air through the mouth and nose occurs right before the speaker stumbles on 'But when you look at me like that', signaling emotional hesitation and physical tension.
A sharp, audible intake of air through the mouth and throat occurs during a pause after 'like that', signaling emotional hesitation and physical tension before confessing vulnerability.
A slow, labored, and audible inhalation and exhalation cycle occurs in the silence following 'all of it', reflecting heavy emotional burden and physical exhaustion.
Local emotion and delivery review — 11 speech segments
Absolute interval
Recognized speech and cue
Gemini ratings
Reviewer explanation
segment-001 Source sentence 1
I keep telling myself it doesn't matter anymore. very fading, very whispered, very apathy, enjoyable
Emotion 3/5 Delivery 4/5
Fairly accurate tags; delivery aligns with whispery quality.
segment-002 Source sentence 2
That keeping quiet protects us. very clean-bg, denasal, tranquility, enjoyable
Emotion 2/5 Delivery 3/5
Labeling emotion as contentment misses the underlying somber tone.
segment-003 Source sentence 3
But very breathy, very noisy, very fondness, enjoyable
Emotion 3/5 Delivery 3/5
Very short word; tagged traits are overstated.
segment-004 Source sentence 3
when you look at me like that, very breathy, very noisy, very fondness, enjoyable
Emotion 3/5 Delivery 3/5
Fondness is partly supported by tone, but breathiness dominates.
segment-005 Source sentence 3
I just want you to see very breathy, very noisy, very fondness, enjoyable
Emotion 3/5 Delivery 3/5
Emotion label is acceptable; delivery fits breathy cue.
segment-006 Source sentence 3
all of it. very breathy, very noisy, very fondness, enjoyable
Emotion 3/5 Delivery 3/5
Quiet resolution fits general tone well.
segment-007 Source sentence 4
God, I feel so ugly saying it out loud. very breathy, very noisy, very mortification, enjoyable
Emotion 4/5 Delivery 3/5
Mortification appropriately reflects the vocal anguish.
segment-008 Source sentence 5
I ran away before anyone could leave me very denasal, very negative, very fear, enjoyable
Emotion 3/5 Delivery 3/5
Fear tag is reasonable, though distress is the primary tone.
segment-009 Source sentence 5
first. very denasal, very negative, very fear, enjoyable
Emotion 3/5 Delivery 3/5
Short isolated word matches preceding segment delivery.
segment-010 Source sentence 6
I ruined that family. very whispered, very fading, very dejection, enjoyable
Emotion 4/5 Delivery 4/5
Dejection fits the mournful tone nicely.
segment-011 Source sentence 7
And I did it on purpose. very fading, very fading, very indifference
Emotion 3/5 Delivery 3/5
Arousal and volume drop well described.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
4/5
Whole-recording emotion fit
4/5
Local speech emotion fit
3/5
Whole-recording talking-style fit
4/5
Local speech delivery fit
3/5
Burst-description fit
3/5
Burst-timing fit
3/5
Event precision rating
3/5
Event coverage rating
3/5
Transcript fidelity (diagnostic)
4/5
Speech-timing fit (diagnostic)
3/5
Readability (diagnostic)
3/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
58 words preserved; 6 selected events serialized; printed elapsed budget 26.570 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English9 selected intervals
Chopping Wood While Telling a Story
en-13004 · 24.41 s · Voice: Iapetus · voicenet_extension
3.15 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A young adult masculine-coded speaker delivers an engaging, humorous anecdote in General American English. The vocal delivery is rich, warm, and resonant, centered in a comfortable baritone register with balanced chest and oral resonance. The pacing is natural and conversational, punctuated by expressive pitch variations and deliberate pauses that build comedic tension. The speaker exhibits high amusement and playful storytelling flair, punctuated by a suppressed chuckle and a brief, sharp gasp of mock surprise before describing the climax of the story. The recording is studio-clean with virtually no background noise or distortion, capturing a highly genuine, spontaneous, and entertaining performance.
SCRIPT — timed speech, local cues and free-form burst descriptions
(denasal, full, surprise, enjoyable; 0.66 seconds duration) ("A quick, audible preparatory sharp inhale occurs right before the speaker launches into the opening word 'So', setting up the narrative with natural respiratory readiness.", 0.48 seconds) So
[0.46 seconds pause]
(denasal, full, surprise, enjoyable; 3.30 seconds duration) Marcus swore the goat was totally tame,
[0.38 seconds pause]
(denasal, full, surprise, enjoyable; 0.15 seconds duration) right?
(very clean-bg, denasal, very sharpness, enjoyable; 2.00 seconds duration) ("A brief, quiet chuckle occurs right after the tag question 'right?', featuring a soft breathy release that naturally blends into the subsequent narrative clause.", 0.44 seconds) He walks up with the feed bucket,
(very clean-bg, denasal, very sharpness, enjoyable; 0.84 seconds duration) ("A brief, quiet chuckle occurs right after 'feed bucket' and before 'makes this ridiculous whistling sound', expressing suppressed amusement at the absurdity of the situation.", 0.44 seconds) makes this
[0.21 seconds pause]
(very clean-bg, denasal, very sharpness, enjoyable; 1.59 seconds duration) ridiculous whistling sound,
(very clean-bg, denasal, very sharpness, enjoyable; 1.78 seconds duration) ("A brief, quiet chuckle with a couple of soft breathy pulses occurs right after describing the ridiculous whistling sound, naturally releasing amusement before continuing the story.", 0.84 seconds) completely ("A brief, quiet chuckle with a soft breathy release occurs right before 'completely confident', naturally underscoring the speaker's amusement at the absurdity of the situation.", 0.54 seconds) confident.
[0.63 seconds pause]
(surging, emphatic, surprise, enjoyable; 2.96 seconds duration) Next thing, the goat just charges him, ("A brief, quiet chuckle with a slight breathy release occurs right after 'dead in the shins', underscoring the absurdity of the goat's sudden attack.", 0.52 seconds)
(surging, emphatic, surprise, enjoyable; 1.67 seconds duration) ("A brief, breathy chuckle occurs right after 'charges him', blending naturally into the pause before 'dead in the shins' with audible smile-voice.", 0.88 seconds) dead in the shins. ("A brief, quiet chuckle occurs right after 'dead in the shins', punctuating the punchline of the goat's sudden attack with natural amusement.", 0.58 seconds)
(clean-bg, full, slightly curiosity, enjoyable; 2.64 seconds duration) ("A brief, quiet chuckle with a couple of soft breathy pulses occurs right after 'dead in the shins', naturally releasing the built-up comedic tension before continuing the story.", 0.84 seconds) Marcus drops the entire bucket,
[0.33 seconds pause]
(clean-bg, full, slightly curiosity, enjoyable; 3.08 seconds duration) leaps backward, and ends up waist deep in the duck pond.
[0.52 seconds pause]
(very hi-fi, very darkening, very letdown; 0.98 seconds duration) Didn't even look back.
[0.23 seconds pause]
Automated review note The speaker delivers a humorous story with amused vocal fry, smiling voice, and several breathy chuckles. The annotation hallucinates multiple redundant, overlapping chuckle events (particularly around 16.0-17.5s, where three separate chuckles are claimed), and inappropriately assigns 'sourness' and 'disappointment' tags to playful comedic lines.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
0.48 s duration · HVA_EVENT_0000
A quick, audible preparatory sharp inhale occurs right before the speaker launches into the opening word 'So', setting up the narrative with natural respiratory readiness.
A brief, quiet chuckle occurs right after the tag question 'right?', featuring a soft breathy release that naturally blends into the subsequent narrative clause.
A brief, quiet chuckle occurs right after 'feed bucket' and before 'makes this ridiculous whistling sound', expressing suppressed amusement at the absurdity of the situation.
Top-three type softmax: Chuckle: 56.5%; Snort: 7.0%; Breathy Giggle: 5.3%
0.84 s duration · HVA_EVENT_0003
A brief, quiet chuckle with a couple of soft breathy pulses occurs right after describing the ridiculous whistling sound, naturally releasing amusement before continuing the story.
A brief, quiet chuckle with a soft breathy release occurs right before 'completely confident', naturally underscoring the speaker's amusement at the absurdity of the situation.
A brief, breathy chuckle occurs right after 'charges him', blending naturally into the pause before 'dead in the shins' with audible smile-voice.
Top-three type softmax: Chuckle: 22.0%; Snort: 8.6%; Exasperated Sigh: 6.8%
0.58 s duration · HVA_EVENT_0006
A brief, quiet chuckle occurs right after 'dead in the shins', punctuating the punchline of the goat's sudden attack with natural amusement.
Top-three type softmax: Chuckle: 36.6%; Breathy Giggle: 14.4%; Snort: 9.1%
0.52 s duration · HVA_EVENT_0007
A brief, quiet chuckle with a slight breathy release occurs right after 'dead in the shins', underscoring the absurdity of the goat's sudden attack.
Top-three type softmax: Chuckle: 19.6%; Snort: 17.3%; Exasperated Sigh: 10.8%
0.84 s duration · HVA_EVENT_0008
A brief, quiet chuckle with a couple of soft breathy pulses occurs right after 'dead in the shins', naturally releasing the built-up comedic tension before continuing the story.
Top-three type softmax: Cough: 68.1%; Chuckle: 10.4%; Throat Clearing: 7.6%
Local emotion and delivery review — 12 speech segments
Absolute interval
Recognized speech and cue
Gemini ratings
Reviewer explanation
segment-001 Source sentence 1
So denasal, full, surprise, enjoyable
Emotion 3/5 Delivery 3/5
Segment starts with an audible breath before speech at around 0.5s; emotion is mildly conversational rather than full astonishment.
segment-002 Source sentence 1
Marcus swore the goat was totally tame, denasal, full, surprise, enjoyable
Emotion 3/5 Delivery 4/5
Engaging storytelling tone; 'surprise' is slightly overstated for this premise setup.
segment-003 Source sentence 1
right? denasal, full, surprise, enjoyable
Emotion 4/5 Delivery 4/5
A rhetorical tag question delivered with an amused upward lilt.
segment-004 Source sentence 2
He walks up with the feed bucket, very clean-bg, denasal, very sharpness, enjoyable
Emotion 2/5 Delivery 3/5
Tagging 'very sharpness / sourness' does not match the amused, humorous delivery.
segment-005 Source sentence 2
makes this very clean-bg, denasal, very sharpness, enjoyable
Emotion 2/5 Delivery 3/5
Short phrase transitioning smoothly; not sour or sharp.
segment-006 Source sentence 2
ridiculous whistling sound, very clean-bg, denasal, very sharpness, enjoyable
Emotion 3/5 Delivery 4/5
Amused description of the whistling sound; delivery is smiling and rhythmic.
segment-007 Source sentence 2
completely confident. very clean-bg, denasal, very sharpness, enjoyable
Emotion 2/5 Delivery 3/5
Preceded by an audible breath/chuckle pause, but emotion is playful irony, not sourness.
segment-008 Source sentence 3
Next thing, the goat just charges him, surging, emphatic, surprise, enjoyable
Emotion 4/5 Delivery 4/5
Good pacing shift and comedic punchiness describing the charge.
segment-009 Source sentence 3
dead in the shins. surging, emphatic, surprise, enjoyable
Emotion 4/5 Delivery 4/5
Punchline delivery delivered with clipped timing and amusement.
segment-010 Source sentence 4
Marcus drops the entire bucket, clean-bg, full, slightly curiosity, enjoyable
Emotion 3/5 Delivery 3/5
Narrative pacing remains amused and light; 'curiosity' is an odd emotion tag here.
segment-011 Source sentence 4
leaps backward, and ends up waist deep in the duck pond. clean-bg, full, slightly curiosity, enjoyable
Emotion 4/5 Delivery 4/5
Climax of the anecdote delivered with vivid narrative inflection.
segment-012 Source sentence 5
Didn't even look back. very hi-fi, very darkening, very letdown
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
54 words preserved; 9 selected events serialized; printed elapsed budget 24.410 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English4 selected intervals
A Lie of Love
en-13064 · 26.30 s · Voice: Algenib · laions_got_talent
3.40 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A single young-adult to mature-adult masculine-coded speaker delivers an intimate, deeply comforting message in General American English. The vocal delivery is exceptionally warm, low-pitched, and grounded in chest resonance, characterized by a soft, breathy half-voice that borders on an ASMR or close-mic bedroom intimacy. The tempo is slow and measured with generous, natural pauses between short, reassuring clauses. Articulation is clear yet gentle, avoiding harsh transients or sharp attacks. The emotional tone is saturated with affection, tenderness, and soothing reassurance, aimed directly at calming an anxious listener. Two soft, breathy chuckles punctuate the speech, blending seamlessly into the quiet, close-mic acoustic space. The recording is pristine with a dead-quiet background and no audible room reflections or technical artifacts.
SCRIPT — timed speech, local cues and free-form burst descriptions
(very denasal, very throaty, very infatuation, enjoyable; 2.48 seconds duration) Of ("A soft, warm, breathy exhalation occurs right at the start of the recording before speech begins, conveying deep affectionate relief and setting an intimate, soothing tone.", 0.88 seconds) course I missed you.
("A soft, warm, breathy chuckle occurs in the pause following 'I missed you', featuring gentle low-pitched vocal pulses that convey affectionate amusement and comfort before transitioning smoothly into the next phrase.", 0.94 seconds)
[0.50 seconds pause]
(very fading, very scattered, very having a crush; 0.75 seconds duration) Look at you.
[0.45 seconds pause]
(very full, very chesty, brooding, enjoyable; 2.70 seconds duration) You always worry about the smallest things.
[0.28 seconds pause]
("A soft, warm, breathy exhalation occurs in the pause after 'the smallest things', releasing gentle tension and transitioning smoothly into the comforting invitation 'Come here'.", 1.04 seconds)
(very non-narrative, very scattered, enjoyable; 0.86 seconds duration) Come here.
[0.24 seconds pause]
(very denasal, very full, very feeling horny, enjoyable; 2.40 seconds duration) Just let me hold you for a second.
[0.40 seconds pause]
(very whispered, very asmr, very fondness, enjoyable; 1.59 seconds duration) We're completely fine.
[0.57 seconds pause]
(very flat, very fading, slightly distrust; 0.26 seconds duration) Really.
[1.18 seconds pause]
(very full, very denasal, very hope, enjoyable; 0.88 seconds duration) I promise
[0.24 seconds pause]
(very full, very denasal, very hope, enjoyable; 3.68 seconds duration) everything is going to be just the way we talked about. ("A soft, quiet breathy chuckle with gentle vocal fold vibration, expressing affectionate reassurance and mild amusement.", 0.74 seconds)
[0.56 seconds pause]
(very darkening, denasal, reverence, enjoyable; 2.05 seconds duration) You don't ever need to ask me that again,
[0.27 seconds pause]
(very darkening, denasal, reverence, enjoyable; 0.18 seconds duration) okay?
[0.62 seconds pause]
(very scattered, relaxed, slightly satisfaction; 0.79 seconds duration) I'm right here.
[0.39 seconds pause]
Automated review note Acoustic and style descriptions are solid, but local emotion tags contain several inaccurate and overly dramatic descriptors.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
0.88 s duration · HVA_EVENT_0000
A soft, warm, breathy exhalation occurs right at the start of the recording before speech begins, conveying deep affectionate relief and setting an intimate, soothing tone.
A soft, warm, breathy chuckle occurs in the pause following 'I missed you', featuring gentle low-pitched vocal pulses that convey affectionate amusement and comfort before transitioning smoothly into the next phrase.
Top-three type softmax: Chuckle: 43.6%; Snicker: 12.8%; Snort: 10.7%
1.04 s duration · HVA_EVENT_0002
A soft, warm, breathy exhalation occurs in the pause after 'the smallest things', releasing gentle tension and transitioning smoothly into the comforting invitation 'Come here'.
A soft, quiet breathy chuckle with gentle vocal fold vibration, expressing affectionate reassurance and mild amusement.
Top-three type softmax: Chuckle: 18.2%; Snort: 14.8%; Exasperated Sigh: 7.4%
Local emotion and delivery review — 12 speech segments
Absolute interval
Recognized speech and cue
Gemini ratings
Reviewer explanation
segment-001 Source sentence 1
Of course I missed you. very denasal, very throaty, very infatuation, enjoyable
Emotion 3/5 Delivery 3/5
Starts over initial breath; overly dramatic emotion tag.
segment-002 Source sentence 2
Look at you. very fading, very scattered, very having a crush
Emotion 3/5 Delivery 3/5
Intimate tone fits, descriptors slightly exaggerated.
segment-003 Source sentence 3
You always worry about the smallest things. very full, very chesty, brooding, enjoyable
Emotion 3/5 Delivery 3/5
Brooding tag is an odd match for reassurance.
segment-004 Source sentence 4
Come here. very non-narrative, very scattered, enjoyable
Emotion 3/5 Delivery 3/5
Delivery fits soft, intimate vocal tone.
segment-005 Source sentence 5
Just let me hold you for a second. very denasal, very full, very feeling horny, enjoyable
Emotion 2/5 Delivery 3/5
Emotion tag is overly sexualized for the calm delivery.
segment-006 Source sentence 6
We're completely fine. very whispered, very asmr, very fondness, enjoyable
Emotion 4/5 Delivery 4/5
Accurately notes the whispered ASMR quality.
segment-007 Source sentence 7
Really. very flat, very fading, slightly distrust
Emotion 2/5 Delivery 3/5
Distrust tag contradicts the reassuring intent.
segment-008 Source sentence 8
I promise very full, very denasal, very hope, enjoyable
Emotion 3/5 Delivery 3/5
Tone is soft and steady.
segment-009 Source sentence 8
everything is going to be just the way we talked about. very full, very denasal, very hope, enjoyable
Emotion 3/5 Delivery 3/5
Speech continues steadily; minor overlap with chuckle.
segment-010 Source sentence 9
You don't ever need to ask me that again, very darkening, denasal, reverence, enjoyable
Emotion 3/5 Delivery 3/5
Delivery remains warm rather than dark.
segment-011 Source sentence 9
okay? very darkening, denasal, reverence, enjoyable
Emotion 3/5 Delivery 3/5
Short tag matches the gentle tag question.
segment-012 Source sentence 10
I'm right here. very scattered, relaxed, slightly satisfaction
Emotion 4/5 Delivery 3/5
Soft, relaxing conclusion matches well.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
4/5
Whole-recording emotion fit
4/5
Local speech emotion fit
3/5
Whole-recording talking-style fit
4/5
Local speech delivery fit
3/5
Burst-description fit
3/5
Burst-timing fit
3/5
Event precision rating
3/5
Event coverage rating
3/5
Transcript fidelity (diagnostic)
4/5
Speech-timing fit (diagnostic)
3/5
Readability (diagnostic)
4/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
55 words preserved; 4 selected events serialized; printed elapsed budget 26.300 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English10 selected intervals
The Late Shift
en-14243 · 28.06 s · Voice: Pulcherrima · situation_inspired
3.55 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A young adult masculine-coded speaker delivers an informal, exasperated monologue in General American English, complaining about a broken glass and a cold candy bar. The vocal delivery is heavily fragmented, slow, and punctuated by frequent pauses, groans, and sighs of physical discomfort and annoyance. The timbre is warm and grounded in chest resonance, with noticeable breathiness and mild vocal fry reflecting fatigue and cold. The performance feels highly genuine and spontaneous, with natural respiratory resets and organic transitions between speech and vocalized frustration.
SCRIPT — timed speech, local cues and free-form burst descriptions
(very breathy, very noisy, very lethargy; 0.44 seconds duration) Hey
("A sharp, breathy exhale of physical pain and exasperation immediately following the initial greeting 'Hey', blending directly into the strained delivery of the next word.", 0.64 seconds)
[0.52 seconds pause]
(very breathy, very noisy, very lethargy; 0.40 seconds duration) Kid,
[0.56 seconds pause]
(very breathy, very noisy, very lethargy; 0.82 seconds duration) don't move.
[0.38 seconds pause]
(very noisy, very breathy, very burnout, less-enjoyable; 1.10 seconds duration) Ugh, ("A low-pitched, strained groan of physical discomfort and exhaustion following the command 'Don't move', blending directly into a breathy release.", 1.10 seconds)
[0.24 seconds pause]
("A low-pitched, strained groan of physical discomfort and exhaustion, emerging directly from the preceding groan and leading into the muttered complaint about the back.", 0.86 seconds)
[0.54 seconds pause]
("A strained, low-pitched vocalized groan of physical discomfort directly preceding the phrase 'my back'.", 0.70 seconds)
(very noisy, very breathy, very burnout, less-enjoyable; 0.26 seconds duration) my
[0.22 seconds pause]
(very noisy, very breathy, very burnout, less-enjoyable; 0.32 seconds duration) back.
[0.80 seconds pause]
(very scattered, very choppy, very pessimism; 0.40 seconds duration) W what
[0.24 seconds pause]
(very scattered, very choppy, very pessimism; 0.53 seconds duration) time is it?
[0.43 seconds pause]
(very explicit, very nasal, slightly uncertainty, less-enjoyable; 0.37 seconds duration) Three?
[0.23 seconds pause]
(overlapping non-speech vocal events; 1.12 seconds duration) ("A sharp, breathy exhale of exasperation and physical discomfort following the question 'Three?', blending directly into the next sentence.", 0.78 seconds) ("A short, breathy groan of physical discomfort and exasperation right before delivering the line about breaking the glass.", 0.66 seconds)
[0.36 seconds pause]
(very effortless, very smooth, very teasing, enjoyable; 2.72 seconds duration) You really broke the glass for a candy bar.
[0.24 seconds pause]
(very darkening, very non-narrative, ribbing; 1.76 seconds duration) You think I'm gonna let you run?
[0.32 seconds pause]
(very non-narrative, very flat, very numbness; 0.23 seconds duration) No.
[0.49 seconds pause]
(very accelerating, very clean-content, detachment; 1.90 seconds duration) Sit down. ("A low-pitched, breathy groan of physical discomfort and exhaustion following the command 'sit down'.", 0.74 seconds) ("A low-pitched, breathy groan of physical discomfort and exhaustion following the command 'sit down', blending directly into the next complaint.", 0.76 seconds)
[0.42 seconds pause]
(very non-instructional, vulnerable, very anxiety, enjoyable; 3.01 seconds duration) I've been freezing on this linoleum all night and now
[0.67 seconds pause]
(very non-instructional, vulnerable, very anxiety, enjoyable; 1.44 seconds duration) Now I get to call your mother.
(very fading, very relaxed, slightly brooding; 2.07 seconds duration) ("A quiet, breathy sigh of resignation and mild exasperation right before delivering the final sarcastic remark.", 0.74 seconds) Wonder ("A quiet, breathy sigh of resignation and mild exasperation right before delivering the final sarcastic remark.", 0.64 seconds) how she'll like that.
[0.31 seconds pause]
Automated review note The audio features a tired, aching speaker complaining in the middle of the night. Several proposed nonverbal events are heavily duplicated/split into overlapping intervals where only a single continuous sigh or groan occurs (e.g. 11.0-12.12s, 18.92-20.14s, 25.74-26.86s).
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
0.64 s duration · HVA_EVENT_0000
A sharp, breathy exhale of physical pain and exasperation immediately following the initial greeting 'Hey', blending directly into the strained delivery of the next word.
A low-pitched, strained groan of physical discomfort and exhaustion following the command 'Don't move', blending directly into a breathy release.
Top-three type softmax: Frustrated Groan: 31.4%; Exhausted Groan: 22.1%; Other / rare burst: 9.5%
0.86 s duration · HVA_EVENT_0002
A low-pitched, strained groan of physical discomfort and exhaustion, emerging directly from the preceding groan and leading into the muttered complaint about the back.
Sit down. very accelerating, very clean-content, detachment
Emotion 3/5 Delivery 3/5
Short command followed by a groaning sigh.
segment-014 Source sentence 9
I've been freezing on this linoleum all night and now very non-instructional, vulnerable, very anxiety, enjoyable
Emotion 3/5 Delivery 3/5
Weary complaint about cold floor.
segment-015 Source sentence 9
Now I get to call your mother. very non-instructional, vulnerable, very anxiety, enjoyable
Emotion 3/5 Delivery 3/5
Resigned statement.
segment-016 Source sentence 10
Wonder how she'll like that. very fading, very relaxed, slightly brooding
Emotion 4/5 Delivery 3/5
Sarcastic, weary remark preceded by a sigh.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
4/5
Whole-recording emotion fit
4/5
Local speech emotion fit
3/5
Whole-recording talking-style fit
4/5
Local speech delivery fit
3/5
Burst-description fit
4/5
Burst-timing fit
3/5
Event precision rating
3/5
Event coverage rating
4/5
Transcript fidelity (diagnostic)
4/5
Speech-timing fit (diagnostic)
3/5
Readability (diagnostic)
4/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
54 words preserved; 10 selected events serialized; printed elapsed budget 28.060 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English10 selected intervals
Stage Light Crush
en-14983 · 28.37 s · Voice: Orus · situation_inspired
3.35 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A single young-adult to mature-adult masculine-coded speaker delivers an intimate, deeply contemplative monologue in General American English. The vocal delivery is exceptionally slow, quiet, and breathy, hovering between a soft stage whisper and a half-voiced ASMR-like register. The timbre is warm, dark, and rounded, with prominent chest and pharyngeal resonance, low harmonicity due to heavy air leakage, and very low muscular tension. Phrasing is fragmented into short, deliberate breath groups separated by long, reflective pauses and audible, gentle inhalations. The emotional tone is saturated with awe, contemplation, and tender affection, conveying a sense of quiet revelation and emotional vulnerability. The recording is pristine, dry, and close-mic'd with no discernible background noise, allowing the subtle respiratory details and delicate vocal textures to remain completely clear and organically integrated.
SCRIPT — timed speech, local cues and free-form burst descriptions
[0.24 seconds pause]
(very whispered, very noisy, very desire, enjoyable; 2.84 seconds duration) And if the stars should turn to ash to night
(very whispered, very noisy, very desire, enjoyable; 3.91 seconds duration) ("A soft, deliberate intake of breath occurs in the pause between the initial and repeated opening lines, functioning as a quiet emotional reset before continuing.", 0.68 seconds) and if ("A soft, deliberate intake of breath occurs in the pause between the two repetitions of the opening line, functioning as a quiet emotional reset before continuing.", 0.54 seconds) the stars should turn to ash to night, ("A slow, audible inhalation and soft exhalation occurs in the pause between the repeated opening line and 'We would still have the morning', grounding the speaker's contemplative pacing.", 0.74 seconds)
(very whispered, very noisy, very desire, enjoyable; 2.21 seconds duration) ("A slow, audible inhalation and soft exhalation occurs in the pause between 'tonight' and 'we would still have', grounding the speaker's contemplative pacing.", 0.78 seconds) we would still have the morning
[0.34 seconds pause]
(very whispered, very noisy, very desire, enjoyable; 3.46 seconds duration) ("A soft, deliberate intake of air occurs in the pause before 'or at least the promise of one', maintaining the quiet, intimate pacing.", 0.64 seconds) or at least the promise of one. ("A slow, audible inhalation and soft exhalation during the pause after 'the promise of one', grounding the speaker's contemplative pacing.", 0.96 seconds)
[0.92 seconds pause]
(very full, very negative, very insensitivity, enjoyable; 3.96 seconds duration) You said once that distance doesn't change what's real. ("A soft, deliberate inhalation taken in the pause after 'what's real', grounding the speaker before admitting past disbelief.", 0.60 seconds) ("A soft, deliberate inhalation taken in the pause after 'what's real', grounding the speaker before admitting past disbelief.", 0.74 seconds)
(very asmr, very whispered, very solace, enjoyable; 1.60 seconds duration) I didn't believe it then,
[0.36 seconds pause]
(very asmr, very whispered, very solace, enjoyable; 0.32 seconds duration) but
[0.24 seconds pause]
(very asmr, very whispered, very solace, enjoyable; 0.69 seconds duration) looking out
[0.27 seconds pause]
(very asmr, very whispered, very solace, enjoyable; 0.46 seconds duration) right now,
[0.90 seconds pause]
(very asmr, very whispered, very solace, enjoyable; 1.82 seconds duration) seeing you there ("A soft, quiet exhale of emotional release and contemplation occurs right after 'seeing you there', bridging the pause before the final reflective realization.", 0.74 seconds)
(very asmr, very whispered, very solace, enjoyable; 3.34 seconds duration) ("A slow, deep audible inhalation and soft exhalation occurs in the pause before 'I think maybe', grounding the speaker's emotional realization and blending seamlessly into the quiet, breathy delivery.", 1.08 seconds) I think maybe I finally understand what you meant.
[0.49 seconds pause]
Automated review note Several burst events are duplicated with heavily overlapping timestamps (e.g. 0000/0001, 0002/0003, 0006/0007, 0008/0009), creating redundant and misaligned event spans. Segment 005 has unfitting acoustic/emotion tags ('very full', 'insensitivity').
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
0.68 s duration · HVA_EVENT_0000
A soft, deliberate intake of breath occurs in the pause between the initial and repeated opening lines, functioning as a quiet emotional reset before continuing.
A soft, deliberate intake of breath occurs in the pause between the two repetitions of the opening line, functioning as a quiet emotional reset before continuing.
A slow, audible inhalation and soft exhalation occurs in the pause between the repeated opening line and 'We would still have the morning', grounding the speaker's contemplative pacing.
A slow, audible inhalation and soft exhalation occurs in the pause between 'tonight' and 'we would still have', grounding the speaker's contemplative pacing.
A soft, quiet exhale of emotional release and contemplation occurs right after 'seeing you there', bridging the pause before the final reflective realization.
A slow, deep audible inhalation and soft exhalation occurs in the pause before 'I think maybe', grounding the speaker's emotional realization and blending seamlessly into the quiet, breathy delivery.
but very asmr, very whispered, very solace, enjoyable
Emotion 3/5 Delivery 4/5
Very short isolated conjunction, tags are inherited.
segment-008 Source sentence 3
looking out very asmr, very whispered, very solace, enjoyable
Emotion 3/5 Delivery 4/5
Intimate whispered phrase matches delivery.
segment-009 Source sentence 3
right now, very asmr, very whispered, very solace, enjoyable
Emotion 3/5 Delivery 4/5
Short isolated segment accurately marked as whispery.
segment-010 Source sentence 3
seeing you there very asmr, very whispered, very solace, enjoyable
Emotion 3/5 Delivery 4/5
Soft whispered delivery aligns well with ASMR descriptors.
segment-011 Source sentence 3
I think maybe I finally understand what you meant. very asmr, very whispered, very solace, enjoyable
Emotion 4/5 Delivery 4/5
Tender, quiet resolution matches the delivery well.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
4/5
Whole-recording emotion fit
4/5
Local speech emotion fit
3/5
Whole-recording talking-style fit
4/5
Local speech delivery fit
3/5
Burst-description fit
3/5
Burst-timing fit
2/5
Event precision rating
3/5
Event coverage rating
4/5
Transcript fidelity (diagnostic)
4/5
Speech-timing fit (diagnostic)
3/5
Readability (diagnostic)
4/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
64 words preserved; 10 selected events serialized; printed elapsed budget 28.370 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English6 selected intervals
Boredom Settling
en-15782 · 28.90 s · Voice: Vindemiatrix · haiku_batch
3.35 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A single young adult masculine-coded speaker delivers a casual, conversational anecdote in General American English with a relaxed, slightly weary tone. The vocal timbre is grounded in a baritone register with balanced chest and oral resonance, moderate warmth, and a clean, natural harmonic core. The delivery is unhurried and segmented into short, reflective phrases punctuated by noticeable pauses, hesitations, and audible sighs. The speaker recounts a mundane meeting with mild amusement and growing fatigue, culminating in a dismissive, resigned attitude. The recording is technically clean with minimal background interference, capturing natural respiratory resets and subtle vocal fry that enhance the lived-in authenticity of the performance.
SCRIPT — timed speech, local cues and free-form burst descriptions
[0.32 seconds pause]
(very narrator, teacherly, slightly pensiveness, enjoyable; 0.24 seconds duration) So
[0.24 seconds pause]
(very narrator, teacherly, slightly pensiveness, enjoyable; 2.47 seconds duration) first he went over the quarterly revisions,
[0.25 seconds pause]
(very narrator, teacherly, slightly pensiveness, enjoyable; 0.06 seconds duration) which
[0.42 seconds pause]
(very narrator, teacherly, slightly pensiveness, enjoyable; 1.20 seconds duration) sounded important,
[0.72 seconds pause]
(very narrator, teacherly, slightly pensiveness, enjoyable; 2.08 seconds duration) but then came the spreadsheets,
[0.24 seconds pause]
(very narrator, teacherly, slightly pensiveness, enjoyable; 2.07 seconds duration) and then the appendices.
[0.39 seconds pause]
(overlapping non-speech vocal events; 1.78 seconds duration) ("A heavy, vocalized exasperated sigh occurs in the pause following 'the appendices', releasing built-up frustration and mental fatigue before continuing the narrative.", 1.34 seconds) ("A heavy, vocalized exasperated sigh occurs in the pause following 'the appendices', releasing built-up frustration and mental fatigue before continuing the narrative.", 0.84 seconds)
(very darkening, slightly non-whispered, slightly curiosity, enjoyable; 3.20 seconds duration) And then someone asked about the slide formatting.
[0.64 seconds pause]
(very meek, very thin-mask, very burnout, enjoyable; 0.50 seconds duration) I just
[0.22 seconds pause]
("A quick, audible sharp inhalation occurs in the pause between 'I just' and 'I don't know', signaling cognitive hesitation and mental fatigue as the speaker searches for words.", 0.66 seconds)
[0.86 seconds pause]
(very meek, very thin-mask, very burnout, enjoyable; 4.12 seconds duration) I don't know, honestly, my brain kinda checked out halfway through.
(overlapping non-speech vocal events; 1.30 seconds duration) ("A low-pitched, closed-mouth vocalized groan and hum of mental fatigue and resignation occurs in the pause before 'Whatever', blending smoothly into the speaker's weary, deflated delivery.", 0.70 seconds) ("A low-pitched, closed-mouth vocalized groan and hum of mental exhaustion and resignation occurs in the pause before 'Whatever', blending smoothly into the speaker's weary, deflated delivery.", 0.78 seconds)
[0.66 seconds pause]
(very non-narrative, very tonal, very boredom; 0.59 seconds duration) Whatever.
[0.45 seconds pause]
(very non-narrative, very cold, very relief; 0.75 seconds duration) It's fine.
[0.27 seconds pause]
(very fading, very decelerating, very indifference, enjoyable; 1.80 seconds duration) ("A quick, quiet nasal-oral sharp inhale occurs in the brief pause before 'I really didn't care anymore', functioning as a weary preparatory breath that naturally leads into the final resigned phrase.", 0.58 seconds) I really didn't care anymore.
[0.40 seconds pause]
Automated review note The non-speech vocal bursts contain several redundant duplicated overlapping intervals (pairs HVA_EVENT_0000/0001 and HVA_EVENT_0003/0004 cover the exact same single sigh and groan events twice). Segment 012 also includes an inline breath detection that overlaps with spoken speech. The speech transcription and overall narrative description are otherwise solid.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
1.34 s duration · HVA_EVENT_0000
A heavy, vocalized exasperated sigh occurs in the pause following 'the appendices', releasing built-up frustration and mental fatigue before continuing the narrative.
A heavy, vocalized exasperated sigh occurs in the pause following 'the appendices', releasing built-up frustration and mental fatigue before continuing the narrative.
A quick, audible sharp inhalation occurs in the pause between 'I just' and 'I don't know', signaling cognitive hesitation and mental fatigue as the speaker searches for words.
A low-pitched, closed-mouth vocalized groan and hum of mental fatigue and resignation occurs in the pause before 'Whatever', blending smoothly into the speaker's weary, deflated delivery.
A low-pitched, closed-mouth vocalized groan and hum of mental exhaustion and resignation occurs in the pause before 'Whatever', blending smoothly into the speaker's weary, deflated delivery.
A quick, quiet nasal-oral sharp inhale occurs in the brief pause before 'I really didn't care anymore', functioning as a weary preparatory breath that naturally leads into the final resigned phrase.
Local emotion and delivery review — 12 speech segments
Absolute interval
Recognized speech and cue
Gemini ratings
Reviewer explanation
segment-001 Source sentence 1
So very narrator, teacherly, slightly pensiveness, enjoyable
Emotion 3/5 Delivery 3/5
Short introductory particle; labels are heavily carried over from whole sentence context.
segment-002 Source sentence 1
first he went over the quarterly revisions, very narrator, teacherly, slightly pensiveness, enjoyable
Emotion 4/5 Delivery 4/5
Casual narrative tone accurately captures the recount of the meeting.
segment-003 Source sentence 1
which very narrator, teacherly, slightly pensiveness, enjoyable
Emotion 3/5 Delivery 3/5
Brief connecting word tagged with identical full-clause descriptors.
segment-004 Source sentence 1
sounded important, very narrator, teacherly, slightly pensiveness, enjoyable
Emotion 3/5 Delivery 3/5
Speaker delivers this mildly, contemplation tag is slightly overstated.
segment-005 Source sentence 1
but then came the spreadsheets, very narrator, teacherly, slightly pensiveness, enjoyable
Emotion 3/5 Delivery 3/5
Tone shifts toward mild exasperation rather than didactic/teacherly style.
segment-006 Source sentence 1
and then the appendices. very narrator, teacherly, slightly pensiveness, enjoyable
Emotion 3/5 Delivery 3/5
Speaker sounds bored/weary, continuing 'teacherly' label is a mismatch.
segment-007 Source sentence 2
And then someone asked about the slide formatting. very darkening, slightly non-whispered, slightly curiosity, enjoyable
Emotion 3/5 Delivery 3/5
Speaker sounds annoyed and fatigued rather than curious.
segment-008 Source sentence 3
I just very meek, very thin-mask, very burnout, enjoyable
Emotion 4/5 Delivery 4/5
Hesitant false start well characterized by fatigue/weariness.
segment-009 Source sentence 3
I don't know, honestly, my brain kinda checked out halfway through. very meek, very thin-mask, very burnout, enjoyable
Emotion 4/5 Delivery 4/5
Burnout and fatigued delivery fit the resigned conversational style.
segment-010 Source sentence 4
Whatever. very non-narrative, very tonal, very boredom
Emotion 4/5 Delivery 4/5
Dismissive and bored delivery accurately labeled.
segment-011 Source sentence 5
It's fine. very non-narrative, very cold, very relief
Emotion 3/5 Delivery 3/5
Tagged as 'relief' but tone is flat and resigned/apathetic rather than truly relieved.
segment-012 Source sentence 6
I really didn't care anymore. very fading, very decelerating, very indifference, enjoyable
Emotion 4/5 Delivery 4/5
Decelerating pace and indifference accurately capture the trailing off ending.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
4/5
Whole-recording emotion fit
4/5
Local speech emotion fit
3/5
Whole-recording talking-style fit
4/5
Local speech delivery fit
3/5
Burst-description fit
3/5
Burst-timing fit
2/5
Event precision rating
3/5
Event coverage rating
4/5
Transcript fidelity (diagnostic)
5/5
Speech-timing fit (diagnostic)
4/5
Readability (diagnostic)
4/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
49 words preserved; 6 selected events serialized; printed elapsed budget 28.900 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English5 selected intervals
Taking a Bite of a Juicy Peach
en-15998 · 22.70 s · Voice: Vindemiatrix · laions_got_talent
3.40 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A young adult masculine-coded speaker delivers an enthusiastic, sensory-rich food review in General American English. The vocal delivery is warm, bright, and highly expressive, characterized by a relaxed conversational tempo, clear articulation, and frequent breathy pauses that convey genuine physical sensation. The speaker's timbre is resonant with balanced chest and oral placement, maintaining a friendly, inviting, and slightly playful tone throughout. Prominent paralinguistic events include an initial surprised gasp, a contented sigh of relief, and an appreciative pleasure moan, all of which blend seamlessly into the spoken phrases. The recording is studio-clean with minimal background noise and high fidelity, capturing the speaker's dynamic pitch excursions and authentic enthusiasm.
SCRIPT — timed speech, local cues and free-form burst descriptions
(very scattered, very mumbled, very startlement, less-enjoyable; 0.78 seconds duration) Oh, ("A soft, breathy vocalized sigh of pleasure and sensory delight opens the recording, transitioning directly into the exclamation 'Oh wow' with natural vocal warmth.", 0.74 seconds)
(very scattered, very mumbled, very startlement, less-enjoyable; 0.76 seconds duration) wow.
[0.30 seconds pause]
(very energetic, very wide-pitch, very astonishment; 0.76 seconds duration) Look at that.
[0.20 seconds pause]
(very dramatic, very darkening, very irritability, enjoyable; 3.02 seconds duration) It's literally dripping down my wrist already.
(very positive, very brightening, very infatuation, enjoyable; 2.56 seconds duration) ("A soft, breathy vocalized sigh of sensory pleasure and relief occurs in the pause between 'already' and 'It's so sweet', releasing tension and leading smoothly into the next appreciative phrase.", 0.86 seconds) ("A soft, breathy vocalized sigh of pleasure and sensory satisfaction occurs in the pause before 'It's so sweet', releasing tension and naturally leading into the next descriptive phrase.", 0.64 seconds) It's so sweet,
[0.26 seconds pause]
(very positive, very brightening, very infatuation, enjoyable; 1.50 seconds duration) like pure sunshine.
[0.66 seconds pause]
(thin-throat, bright, startlement, enjoyable; 2.75 seconds duration) But it's got that tiny bit of tartness right near the pit.
(very breathy, very asmr, very adoration, enjoyable; 0.83 seconds duration) Oh ("A soft, expressive vocalized sigh of pleasure and sensory satisfaction ('Oh') that bridges the description of tartness and the melting texture.", 0.68 seconds)
(very breathy, very asmr, very adoration, enjoyable; 3.36 seconds duration) it just completely melts in your mouth. ("A soft, warm, contented sigh of pure culinary satisfaction and pleasure, leading directly into the earnest recommendation 'Seriously, you have to try one of these.'", 0.62 seconds)
(very darkening, dialogic, very short-temperedness; 2.03 seconds duration) Seriously, you have to try one of these.
[0.23 seconds pause]
(very clean-bg, hi-fi, very beatitude, enjoyable; 1.89 seconds duration) It's the absolute best part of summer.
[0.81 seconds pause]
Automated review note Global traits are well captured, though negative emotional tags like irritability appear inappropriately on enthusiastic segments.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
0.74 s duration · HVA_EVENT_0000
A soft, breathy vocalized sigh of pleasure and sensory delight opens the recording, transitioning directly into the exclamation 'Oh wow' with natural vocal warmth.
A soft, breathy vocalized sigh of sensory pleasure and relief occurs in the pause between 'already' and 'It's so sweet', releasing tension and leading smoothly into the next appreciative phrase.
A soft, breathy vocalized sigh of pleasure and sensory satisfaction occurs in the pause before 'It's so sweet', releasing tension and naturally leading into the next descriptive phrase.
A soft, warm, contented sigh of pure culinary satisfaction and pleasure, leading directly into the earnest recommendation 'Seriously, you have to try one of these.'
Local emotion and delivery review — 11 speech segments
Absolute interval
Recognized speech and cue
Gemini ratings
Reviewer explanation
segment-001 Source sentence 1
Oh, very scattered, very mumbled, very startlement, less-enjoyable
Emotion 3/5 Delivery 2/5
Tagged as mumbled and disorganized, but speech is audible and open.
segment-002 Source sentence 1
wow. very scattered, very mumbled, very startlement, less-enjoyable
Emotion 3/5 Delivery 3/5
Mildly astonished tone, not particularly scattered.
segment-003 Source sentence 2
Look at that. very energetic, very wide-pitch, very astonishment
Emotion 4/5 Delivery 4/5
High arousal and pitch movement are present.
segment-004 Source sentence 3
It's literally dripping down my wrist already. very dramatic, very darkening, very irritability, enjoyable
Emotion 2/5 Delivery 3/5
Irritability tag does not match the appreciative delivery.
segment-005 Source sentence 4
It's so sweet, very positive, very brightening, very infatuation, enjoyable
Emotion 4/5 Delivery 4/5
Positive valence fits the sweet description.
segment-006 Source sentence 4
like pure sunshine. very positive, very brightening, very infatuation, enjoyable
Emotion 4/5 Delivery 4/5
Warm, infatuated tone matches well.
segment-007 Source sentence 5
But it's got that tiny bit of tartness right near the pit. thin-throat, bright, startlement, enjoyable
Emotion 3/5 Delivery 3/5
Conversational observation rather than clear startlement.
segment-008 Source sentence 6
Oh very breathy, very asmr, very adoration, enjoyable
Emotion 4/5 Delivery 3/5
Sigh-like exclamation fits breathy adoration.
segment-009 Source sentence 6
it just completely melts in your mouth. very breathy, very asmr, very adoration, enjoyable
Emotion 4/5 Delivery 4/5
Soft, appreciative delivery aligns with description.
segment-010 Source sentence 7
Seriously, you have to try one of these. very darkening, dialogic, very short-temperedness
Emotion 2/5 Delivery 3/5
Short-tempered tag misidentifies enthusiastic insistence.
segment-011 Source sentence 8
It's the absolute best part of summer. very clean-bg, hi-fi, very beatitude, enjoyable
Emotion 4/5 Delivery 4/5
Contented, positive tone matches well.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
4/5
Whole-recording emotion fit
4/5
Local speech emotion fit
3/5
Whole-recording talking-style fit
4/5
Local speech delivery fit
3/5
Burst-description fit
3/5
Burst-timing fit
3/5
Event precision rating
3/5
Event coverage rating
3/5
Transcript fidelity (diagnostic)
4/5
Speech-timing fit (diagnostic)
3/5
Readability (diagnostic)
4/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
53 words preserved; 5 selected events serialized; printed elapsed budget 22.700 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English13 selected intervals
Confronting the Parent Who Chose Silence
en-16375 · 37.13 s · Voice: Sadachbia · existing_challenge_inspired
3.35 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A single young-adult to mature-adult masculine-coded speaker delivers an intensely dramatic, emotionally vulnerable monologue in General American English. The performance is dominated by profound sadness, distress, and helplessness, characterized by a slow tempo, fragmented phrasing, and heavy, labored breathing. The vocal timbre is warm yet strained, with prominent chest and pharyngeal resonance, noticeable breathiness, and a fragile, cracking quality that borders on a stage whisper. The speaker repeatedly breaks into audible gasps, shuddering sobs, and weary sighs between short, halting clauses, creating a deeply exposed and sorrowful atmosphere. The recording is dry and clean with negligible background noise, capturing intimate vocal micro-details and breath dynamics with high fidelity.
SCRIPT — timed speech, local cues and free-form burst descriptions
[0.04 seconds pause]
("A heavy, weary exhalation opens the recording before any speech begins, establishing a state of deep emotional exhaustion and suppressed distress.", 0.82 seconds)
[1.62 seconds pause]
(very denasal, very full, very mortification, enjoyable; 3.72 seconds duration) I came here today because I needed to say this calmly. ("A sharp, audible intake of air through the mouth and throat occurs during a pause between phrases, reflecting emotional strain and physical preparation before speaking.", 0.70 seconds) ("A sharp, audible intake of air through the mouth and throat occurs during a pause between phrases, reflecting emotional strain and physical preparation before speaking.", 0.84 seconds)
(very full, very chesty, horror, enjoyable; 3.01 seconds duration) But ("A sharp, audible intake of air through the mouth and throat occurs during a pause between phrases, reflecting emotional tension and physical preparation before speaking.", 0.90 seconds) hearing you say you were scared too.
[0.31 seconds pause]
(overlapping non-speech vocal events; 0.96 seconds duration) ("A quick, tense inhalation taken through the mouth during a pause, reflecting emotional strain and preparation before delivering the next line.", 0.64 seconds) ("A quick, tense inhalation through the mouth and nose occurs in the pause before 'You were forty', reflecting emotional bracing and physical constriction.", 0.62 seconds)
[0.96 seconds pause]
(very surging, very tense, very fury; 1.18 seconds duration) You were forty.
(very fading, very meek, slightly discomfiture; 2.06 seconds duration) ("A quick, tense inhalation taken between 'You were forty' and 'I was eight years old', highlighting emotional distress and physical breathlessness.", 0.68 seconds) I was eight years old.
[0.68 seconds pause]
(very tense, very darkening, very enragement, enjoyable; 2.39 seconds duration) ("A sharp, audible intake of air following 'I was eight years old', marking a sudden surge of emotional distress and physical tension before the accusatory question.", 0.86 seconds) Why didn't you stop him?
[0.62 seconds pause]
("A heavy, despondent exhalation following the accusatory question 'Why didn't you stop him?', releasing built-up tension with a weary, defeated quality.", 1.52 seconds)
[1.23 seconds pause]
(very denasal, very whispered, very bitterness, enjoyable; 2.58 seconds duration) Just look at me and say you should have done something.
[0.70 seconds pause]
(very non-narrative, choppy, slightly provoking lightly, less-enjoyable; 0.64 seconds duration) Please.
[0.63 seconds pause]
("A heavy, despondent sigh of resignation and exhaustion following the quiet plea 'Please', marking the transition into bitter defeat.", 1.22 seconds)
[1.59 seconds pause]
(very non-narrative, very flat, slightly detachment, less-enjoyable; 0.36 seconds duration) Right.
[1.24 seconds pause]
(very denasal, very darkening, very bitterness, enjoyable; 1.60 seconds duration) You're never gonna say it, are you?
[1.13 seconds pause]
(overlapping non-speech vocal events; 1.68 seconds duration) ("A heavy, despondent exhalation following the resigned question 'are you?', marking the collapse of emotional energy.", 1.08 seconds) ("A final deflated, weary exhalation that leads directly into the quiet, defeated closing line 'There's nothing left here.'", 1.08 seconds)
[0.36 seconds pause]
(very fading, very denasal, very detachment; 2.03 seconds duration) ("A final deflated, breathy sigh of resignation and exhaustion right before the concluding line 'There's nothing left here.'", 1.02 seconds) There's nothing left here.
[0.25 seconds pause]
Automated review note The global description captures the sorrowful, dramatic acting performance well. However, multiple redundant and duplicate overlapping burst events are hallucinated (e.g., repeating breaths around 6s, 9.8s, and 33-35s), and segment-007 severely misinterprets 'Please' as teasing.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
0.82 s duration · HVA_EVENT_0000
A heavy, weary exhalation opens the recording before any speech begins, establishing a state of deep emotional exhaustion and suppressed distress.
A sharp, audible intake of air through the mouth and throat occurs during a pause between phrases, reflecting emotional strain and physical preparation before speaking.
A sharp, audible intake of air through the mouth and throat occurs during a pause between phrases, reflecting emotional strain and physical preparation before speaking.
A sharp, audible intake of air through the mouth and throat occurs during a pause between phrases, reflecting emotional tension and physical preparation before speaking.
A quick, tense inhalation through the mouth and nose occurs in the pause before 'You were forty', reflecting emotional bracing and physical constriction.
Top-three type softmax: Sniff: 28.9%; Sharp Inhale: 12.8%; Snort: 7.3%
0.68 s duration · HVA_EVENT_0006
A quick, tense inhalation taken between 'You were forty' and 'I was eight years old', highlighting emotional distress and physical breathlessness.
A sharp, audible intake of air following 'I was eight years old', marking a sudden surge of emotional distress and physical tension before the accusatory question.
A heavy, despondent exhalation following the accusatory question 'Why didn't you stop him?', releasing built-up tension with a weary, defeated quality.
Local emotion and delivery review — 10 speech segments
Absolute interval
Recognized speech and cue
Gemini ratings
Reviewer explanation
segment-001 Source sentence 1
I came here today because I needed to say this calmly. very denasal, very full, very mortification, enjoyable
Emotion 3/5 Delivery 4/5
Speaker sounds restrained and somber rather than displaying overt mortification.
segment-002 Source sentence 2
But hearing you say you were scared too. very full, very chesty, horror, enjoyable
Emotion 3/5 Delivery 4/5
Delivery is emotionally pained and hurt rather than direct horror.
segment-003 Source sentence 3
You were forty. very surging, very tense, very fury
Emotion 4/5 Delivery 4/5
High tension and sharp emotional intensity are well noted.
segment-004 Source sentence 4
I was eight years old. very fading, very meek, slightly discomfiture
Emotion 3/5 Delivery 3/5
Fading energy fits, but the emotion is grief and vulnerability rather than meek discomfiture.
segment-005 Source sentence 5
Why didn't you stop him? very tense, very darkening, very enragement, enjoyable
Emotion 4/5 Delivery 4/5
Accusatory tone and emotional anguish captured reasonably well with tension and anger tags.
segment-006 Source sentence 6
Just look at me and say you should have done something. very denasal, very whispered, very bitterness, enjoyable
Emotion 4/5 Delivery 4/5
Husky, strained bitterness fits the delivery accurately.
segment-007 Source sentence 7
Please. very non-narrative, choppy, slightly provoking lightly, less-enjoyable
Emotion 1/5 Delivery 3/5
Tagging 'Please' with teasing or provoking is completely inaccurate; it is a desperate plea.
segment-008 Source sentence 8
Right. very non-narrative, very flat, slightly detachment, less-enjoyable
Emotion 4/5 Delivery 4/5
Quiet resignation and flat delivery match 'Right.'
segment-009 Source sentence 9
You're never gonna say it, are you? very denasal, very darkening, very bitterness, enjoyable
Emotion 4/5 Delivery 4/5
Bitterness and defeat fit this resigned question.
segment-010 Source sentence 10
There's nothing left here. very fading, very denasal, very detachment
Emotion 4/5 Delivery 4/5
Detachment, exhaustion, and fading energy suit the final line well.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
4/5
Whole-recording emotion fit
4/5
Local speech emotion fit
3/5
Whole-recording talking-style fit
4/5
Local speech delivery fit
3/5
Burst-description fit
3/5
Burst-timing fit
2/5
Event precision rating
3/5
Event coverage rating
4/5
Transcript fidelity (diagnostic)
5/5
Speech-timing fit (diagnostic)
4/5
Readability (diagnostic)
4/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
56 words preserved; 13 selected events serialized; printed elapsed budget 37.130 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English8 selected intervals
Curdled Pride on the Tongue
en-16625 · 22.54 s · Voice: Zubenelgenubi · extreme_physical
3.45 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A mature adult masculine-coded voice delivers a dramatic, character-driven monologue in General American English. The timbre is deep, chest-resonant, and gravelly with noticeable vocal fry and roughness, maintaining a low baritone register. The delivery is slow, deliberate, and heavily chunked, marked by prominent breathiness, weary sighs, and cynical chuckles. The speaker conveys a mix of bitter amusement, contempt, and physical exhaustion, reflecting on a failed attempt to break milk. The recording is dry and clean with minimal background interference, capturing intimate vocal textures and natural respiratory pauses.
SCRIPT — timed speech, local cues and free-form burst descriptions
(very full, very denasal, slightly self-consciousness, enjoyable; 2.94 seconds duration) ("A low, gravelly guttural grunt and breath intake precedes the opening line, setting a menacing, physically grounded tone before speech begins.", 0.44 seconds) He thinks he's the first one to break milk this way.
[0.74 seconds pause]
(very full, very denasal, very self-confidently, enjoyable; 2.53 seconds duration) I've buried men who brag far less.
[0.99 seconds pause]
(very naturalistic, relaxed, relaxation; 1.01 seconds duration) Smells right, though.
[0.59 seconds pause]
(very tense, very breathy, very abhorrence, less-enjoyable; 1.58 seconds duration) ("A low, breathy, half-voiced groan of physical discomfort and cold shivering occurs in the pause before the speaker exclaims 'Oh god', blending naturally with the labored breathing.", 1.00 seconds) ("A low, strained, breathy groan of physical discomfort and cold shivering, uttered by the second speaker just before saying 'Oh god', blending directly into the shivering vocalization.", 0.86 seconds) Oh
[0.26 seconds pause]
(very tense, very breathy, very abhorrence, less-enjoyable; 0.09 seconds duration) God.
[0.21 seconds pause]
(overlapping non-speech vocal events; 0.86 seconds duration) ("A low, strained guttural grunt of physical discomfort and cold shivering, uttered in isolation between phrases.", 0.58 seconds) ("A brief, low-pitched guttural grunt of displeasure and physical strain, uttered in isolation between phrases.", 0.54 seconds)
[1.00 seconds pause]
(very denasal, very darkening, very loathing, enjoyable; 2.01 seconds duration) Still not as sharp as my own first batch,
[0.23 seconds pause]
(very denasal, very darkening, very loathing, enjoyable; 0.65 seconds duration) you bastard.
[0.33 seconds pause]
(very breathy, very fading, very lethargy, enjoyable; 2.22 seconds duration) ("A heavy, shivering exhalation and breathy sigh of cold discomfort occurs in the pause before 'My teeth are buzzing', blending naturally with the speaker's physical shivering.", 1.34 seconds) My teeth are buzzing.
(overlapping non-speech vocal events; 1.40 seconds duration) ("A quick, audible gasp-like inhalation through the mouth and teeth, reflecting physical coldness and bracing against the biting cold.", 0.76 seconds) ("A low, breathy, half-suppressed chuckle and amused exhalation following the comment about buzzing teeth, blending naturally into the final taunting remark.", 0.82 seconds)
[0.36 seconds pause]
(very full, very denasal, slightly mocking playfully, enjoyable; 2.32 seconds duration) But he got the curds to bite, didn't he?
[0.22 seconds pause]
Automated review note Accurate overall acoustic description, though burst events contain multiple redundant duplicate intervals.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
0.44 s duration · HVA_EVENT_0000
A low, gravelly guttural grunt and breath intake precedes the opening line, setting a menacing, physically grounded tone before speech begins.
A low, breathy, half-voiced groan of physical discomfort and cold shivering occurs in the pause before the speaker exclaims 'Oh god', blending naturally with the labored breathing.
A low, strained, breathy groan of physical discomfort and cold shivering, uttered by the second speaker just before saying 'Oh god', blending directly into the shivering vocalization.
A heavy, shivering exhalation and breathy sigh of cold discomfort occurs in the pause before 'My teeth are buzzing', blending naturally with the speaker's physical shivering.
A low, breathy, half-suppressed chuckle and amused exhalation following the comment about buzzing teeth, blending naturally into the final taunting remark.
Local emotion and delivery review — 9 speech segments
Absolute interval
Recognized speech and cue
Gemini ratings
Reviewer explanation
segment-001 Source sentence 1
He thinks he's the first one to break milk this way. very full, very denasal, slightly self-consciousness, enjoyable
Emotion 3/5 Delivery 3/5
Gravelly delivery is accurately captured, but pride cue is slightly off.
segment-002 Source sentence 2
I've buried men who brag far less. very full, very denasal, very self-confidently, enjoyable
Emotion 3/5 Delivery 3/5
Matches the dry, grim vocal performance fairly well.
segment-003 Source sentence 3
Smells right, though. very naturalistic, relaxed, relaxation
Emotion 3/5 Delivery 3/5
Relaxation tag is passable though performance stays gritty.
segment-004 Source sentence 4
Oh very tense, very breathy, very abhorrence, less-enjoyable
Emotion 3/5 Delivery 3/5
Disgust and tension are reasonable matches for the strained groan.
segment-005 Source sentence 4
God. very tense, very breathy, very abhorrence, less-enjoyable
Emotion 3/5 Delivery 3/5
Segment is very short and matches the preceding strained tone.
segment-006 Source sentence 5
Still not as sharp as my own first batch, very denasal, very darkening, very loathing, enjoyable
Emotion 4/5 Delivery 3/5
Contempt and dark tone match the spoken batch comment.
segment-007 Source sentence 5
you bastard. very denasal, very darkening, very loathing, enjoyable
Emotion 4/5 Delivery 3/5
Fits the bitter, biting spoken insult well.
segment-008 Source sentence 6
My teeth are buzzing. very breathy, very fading, very lethargy, enjoyable
Emotion 3/5 Delivery 3/5
Breathy, shivered delivery is captured reasonably well.
segment-009 Source sentence 7
But he got the curds to bite, didn't he? very full, very denasal, slightly mocking playfully, enjoyable
Emotion 3/5 Delivery 3/5
Playful mocking is slightly overstated, but teasing fit is acceptable.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
4/5
Whole-recording emotion fit
4/5
Local speech emotion fit
3/5
Whole-recording talking-style fit
4/5
Local speech delivery fit
3/5
Burst-description fit
3/5
Burst-timing fit
3/5
Event precision rating
3/5
Event coverage rating
4/5
Transcript fidelity (diagnostic)
4/5
Speech-timing fit (diagnostic)
3/5
Readability (diagnostic)
4/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
47 words preserved; 8 selected events serialized; printed elapsed budget 22.540 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English12 selected intervals
Wheel of Ruin
en-17449 · 20.69 s · Voice: Fenrir · extreme_physical
3.30 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A young adult masculine-coded speaker delivers an intensely dramatic, breathless monologue in General American English, depicting a high-stakes driving scenario. The performance begins with an intimate, breathy, and sensual tone characterized by warm chest resonance, low pitch, and close-mic proximity, before rapidly escalating into acute panic, fear, and helplessness. The tempo is initially slow and fragmented with heavy, audible respiration, but accelerates into rapid, choppy, and strained phrasing as the speaker realizes they are losing control. Articulation remains clear despite the heavy breathiness and vocal constriction. The recording is pristine with a dead-quiet background, capturing every subtle gasp, sharp inhale, and trembling vocal tremor with high fidelity.
SCRIPT — timed speech, local cues and free-form burst descriptions
(very non-instructional, very vulnerable, very bliss, enjoyable; 0.52 seconds duration) Oh ("A soft, breathy vocalized sigh of pleasure and relaxation opens the recording right before the speaker utters 'God', blending smoothly into the initial phrase with an open, relaxed throat.", 0.66 seconds)
(very non-instructional, very vulnerable, very bliss, enjoyable; 2.15 seconds duration) god, I love the pull of this road.
[0.33 seconds pause]
(very whispered, very organic, very feeling horny; 1.92 seconds duration) ("A quick, audible gasp-like inhalation occurs in the pause between 'pull of this road' and 'Just taking it in deeper', signaling a sudden shift in focus and physical tension.", 0.52 seconds) Just taking it in deeper.
(very whispered, very noisy, very infatuation, enjoyable; 1.50 seconds duration) ("A quick, audible gasp-like inhalation occurs in the pause between 'deeper' and 'The way', marking a sudden shift in focus and rising anxiety.", 0.60 seconds) The way everything
[0.54 seconds pause]
(very whispered, very noisy, very infatuation, enjoyable; 0.21 seconds duration) just
[0.27 seconds pause]
(very whispered, very noisy, very infatuation, enjoyable; 1.32 seconds duration) holds you. ("A sharp, constricted gasp of sudden realization and alarm right before the speaker exclaims 'No', marking an abrupt shift from relaxed appreciation to panic.", 0.74 seconds)
(very non-narrative, very scattered, less-enjoyable; 1.42 seconds duration) Wait. ("A sharp, audible intake of air following 'holds you' that abruptly interrupts the calm tone, signaling sudden realization and alarm.", 1.06 seconds)
(very fading, very dissociated, insensitivity, less-enjoyable; 0.36 seconds duration) No.
[0.34 seconds pause]
(very scattered, very breathy, slightly indifference, enjoyable; 0.76 seconds duration) The wheel.
(very non-narrative, very fading, burnout; 0.91 seconds duration) ("A quick, tense inhalation interrupting the phrase 'The wheel', signaling sudden panic and physical strain.", 0.54 seconds) It ("A quick, tense inhalation interrupting the phrase 'It won't turn', signaling sudden panic and physical strain.", 0.56 seconds) won't turn.
(very cold, metallic, slightly startlement; 0.93 seconds duration) It's locked up. ("A sharp, constricted gasp of air taken in preparation for the desperate, shouted command 'Turn!'.", 0.66 seconds)
(very breathy, very fading, very desperation; 1.35 seconds duration) ("A sharp, constricted fearful gasp of air taken in between 'It's locked up' and 'I can't hold this straight', reflecting sudden panic and physical strain.", 0.80 seconds) I can't hold it straight.
(very breathy, very noisy, very nervousness, enjoyable; 2.37 seconds duration) ("A quick, tense inhalation taken in the pause before pleading 'Please don't slip', showing rising panic and physical strain.", 0.54 seconds) Oh, please don't slip. ("A sharp, constricted gasp of air taken in panic right before shouting 'Turn!', showing high physiological arousal and sudden respiratory demand.", 0.64 seconds)
[0.56 seconds pause]
(very non-narrative, very decelerating, less-enjoyable; 0.56 seconds duration) Turn!
[0.40 seconds pause]
(very bright, very metallic, very startlement; 0.80 seconds duration) Damn ("A sharp, constricted gasp of panic and exasperation right before the final shouted command 'turn it'.", 0.60 seconds) it,
[0.24 seconds pause]
(very bright, very metallic, very startlement; 0.47 seconds duration) turn!
[0.46 seconds pause]
Automated review note Several burst events in the second half directly overlap spoken words (e.g., 18.72 to 19.32 covering 'Damn it') rather than true pauses, and some segments in the panic phase are misclassified with numbness/indifference/metallic tags.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
0.66 s duration · HVA_EVENT_0000
A soft, breathy vocalized sigh of pleasure and relaxation opens the recording right before the speaker utters 'God', blending smoothly into the initial phrase with an open, relaxed throat.
Top-three type softmax: Exasperated Sigh: 31.9%; Frustrated Groan: 12.4%; Other / rare burst: 7.6%
0.52 s duration · HVA_EVENT_0001
A quick, audible gasp-like inhalation occurs in the pause between 'pull of this road' and 'Just taking it in deeper', signaling a sudden shift in focus and physical tension.
A sharp, constricted gasp of sudden realization and alarm right before the speaker exclaims 'No', marking an abrupt shift from relaxed appreciation to panic.
A sharp, constricted fearful gasp of air taken in between 'It's locked up' and 'I can't hold this straight', reflecting sudden panic and physical strain.
Damn it, very bright, very metallic, very startlement
Emotion 3/5 Delivery 2/5
Bright/metallic tags do not fit the angry/frantic shout.
segment-016 Source sentence 12
turn! very bright, very metallic, very startlement
Emotion 3/5 Delivery 2/5
Desperate final shout; metallic tag is poorly fitting.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
4/5
Whole-recording emotion fit
4/5
Local speech emotion fit
3/5
Whole-recording talking-style fit
4/5
Local speech delivery fit
3/5
Burst-description fit
3/5
Burst-timing fit
2/5
Event precision rating
3/5
Event coverage rating
3/5
Transcript fidelity (diagnostic)
4/5
Speech-timing fit (diagnostic)
3/5
Readability (diagnostic)
4/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
43 words preserved; 12 selected events serialized; printed elapsed budget 20.690 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English8 selected intervals
Correction in Front of Crowd
en-17663 · 20.45 s · Voice: Schedar · haiku_batch
2.90 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A single young adult masculine-coded speaker delivers a casual, unscripted conversational monologue in General American English. The vocal delivery is grounded in a comfortable baritone register with balanced chest and oral resonance, moderate warmth, and a clean, natural harmonic profile. The pacing is standard conversational speed with natural chunking, punctuated by brief pauses, self-corrections, and audible breaths. The speaker exhibits a relaxed, cooperative stance and an open, reflective emotional state marked by mild contemplation, slight self-conscious amusement, and a touch of embarrassment over an awkward conversational tangent. The recording is technically clean with minimal background interference, capturing natural respiratory resets and organic speech disfluencies that contribute to a highly genuine, lived-in performance.
SCRIPT — timed speech, local cues and free-form burst descriptions
(very non-narrative, very jerky, very shock; 0.82 seconds duration) Oh. ("A voiced, low-pitched 'oh' sound expressing sudden realization and mild surprise, opening the clip with a relaxed, open throat and transitioning smoothly into a brief pause.", 0.70 seconds) ("A brief, low-pitched chuckle occurs right after the initial realization marker 'Oh', expressing mild self-deprecating amusement before continuing with 'huh'.", 0.50 seconds)
(very non-narrative, very clean-content, less-enjoyable; 0.94 seconds duration) Huh. ("A brief, low-pitched chuckle occurs right after the initial realization 'Oh', featuring a soft, breathy vocal release that conveys mild self-deprecating amusement before continuing into the next phrase.", 0.60 seconds) ("A brief, low-pitched chuckle occurs right after the initial 'Oh' and before 'I actually didn't know that', expressing mild self-deprecating amusement and conversational realization.", 0.40 seconds)
[0.40 seconds pause]
(very fading, very clean-content, very amazement, enjoyable; 0.32 seconds duration) I
[0.24 seconds pause]
(very fading, very clean-content, very amazement, enjoyable; 1.68 seconds duration) actually didn't know that. ("A soft, audible inhalation occurs during a pause between phrases, serving as a natural respiratory reset before the speaker continues with 'I mean, I was pretty sure'.", 0.64 seconds)
(very fading, clean-content, very skepticism, enjoyable; 2.96 seconds duration) I mean I was pretty sure the data pointed the other way,
[0.28 seconds pause]
(very fading, clean-content, very skepticism, enjoyable; 0.74 seconds duration) but ("A soft, audible exhalation of breath occurs during a pause after 'but', signaling mild resignation and cognitive processing before continuing with 'well'.", 0.74 seconds)
(very fading, clean-content, very skepticism, enjoyable; 3.38 seconds duration) well, I guess the general principle was close enough.
[0.88 seconds pause]
(very denasal, very accelerating, very apathy; 1.12 seconds duration) Or maybe not.
[0.24 seconds pause]
(very clean-content, fading, slightly pondering, enjoyable; 0.31 seconds duration) Look,
[0.25 seconds pause]
(very clean-content, fading, slightly pondering, enjoyable; 0.04 seconds duration) just
[0.28 seconds pause]
(very clean-content, fading, slightly pondering, enjoyable; 0.88 seconds duration) forget I said
[0.24 seconds pause]
(very clean-content, fading, slightly pondering, enjoyable; 0.80 seconds duration) that. ("A soft, audible exhalation of breath occurs during a pause after 'forget I said that', functioning as a self-deprecating release of tension before yielding the floor to the other speaker.", 0.80 seconds)
(very full, very chesty, enjoyable; 3.65 seconds duration) ("A soft, audible exhalation of relief and mild self-deprecation occurs right before yielding the floor to the other speaker, blending seamlessly into the subsequent conversational turn.", 0.64 seconds) Go ahead, Dave, finish what you were gonna say before I jumped in.
Automated review note The script contains excessive hallucinated and overlapping burst events, including classifying speech words ('Oh') as bursts and multiplying chuckle events across 0.6s to 1.76s. Sentences are also overly fragmented into unnatural single-word chunks.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
0.70 s duration · HVA_EVENT_0000
A voiced, low-pitched 'oh' sound expressing sudden realization and mild surprise, opening the clip with a relaxed, open throat and transitioning smoothly into a brief pause.
A brief, low-pitched chuckle occurs right after the initial realization marker 'Oh', expressing mild self-deprecating amusement before continuing with 'huh'.
A brief, low-pitched chuckle occurs right after the initial realization 'Oh', featuring a soft, breathy vocal release that conveys mild self-deprecating amusement before continuing into the next phrase.
A brief, low-pitched chuckle occurs right after the initial 'Oh' and before 'I actually didn't know that', expressing mild self-deprecating amusement and conversational realization.
A soft, audible inhalation occurs during a pause between phrases, serving as a natural respiratory reset before the speaker continues with 'I mean, I was pretty sure'.
A soft, audible exhalation of breath occurs during a pause after 'but', signaling mild resignation and cognitive processing before continuing with 'well'.
A soft, audible exhalation of breath occurs during a pause after 'forget I said that', functioning as a self-deprecating release of tension before yielding the floor to the other speaker.
A soft, audible exhalation of relief and mild self-deprecation occurs right before yielding the floor to the other speaker, blending seamlessly into the subsequent conversational turn.
Local emotion and delivery review — 13 speech segments
Absolute interval
Recognized speech and cue
Gemini ratings
Reviewer explanation
segment-001 Source sentence 1
Oh. very non-narrative, very jerky, very shock
Emotion 4/5 Delivery 2/5
Spoken word 'Oh'. Tagged as very jerky, but it is a single isolated syllable.
segment-002 Source sentence 2
Huh. very non-narrative, very clean-content, less-enjoyable
Emotion 3/5 Delivery 3/5
Spoken word 'Huh.' followed by a quiet laugh/breath.
segment-003 Source sentence 3
I very fading, very clean-content, very amazement, enjoyable
Emotion 2/5 Delivery 2/5
Splits single clause artificially on 'I'. Rating of extreme amazement is an overstatement.
segment-004 Source sentence 3
actually didn't know that. very fading, very clean-content, very amazement, enjoyable
Emotion 3/5 Delivery 3/5
Delivery is conversational admission, not extreme amazement.
segment-005 Source sentence 4
I mean I was pretty sure the data pointed the other way, very fading, clean-content, very skepticism, enjoyable
Emotion 4/5 Delivery 3/5
Expresses qualification and doubt reasonably well.
segment-006 Source sentence 4
but very fading, clean-content, very skepticism, enjoyable
Emotion 3/5 Delivery 2/5
Isolated conjunction 'but' with trailing pause/breath.
segment-007 Source sentence 4
well, I guess the general principle was close enough. very fading, clean-content, very skepticism, enjoyable
Emotion 4/5 Delivery 3/5
Casual concession fits the doubt/skepticism tag somewhat.
segment-008 Source sentence 5
Or maybe not. very denasal, very accelerating, very apathy
Emotion 2/5 Delivery 2/5
Tagged as very apathy, but delivery is ironic/rueful realization.
segment-009 Source sentence 6
Look, very clean-content, fading, slightly pondering, enjoyable
Emotion 3/5 Delivery 3/5
Single word segment 'Look' from an unnatural phrase chop.
segment-010 Source sentence 6
just very clean-content, fading, slightly pondering, enjoyable
Emotion 3/5 Delivery 2/5
Extremely brief chop on 'just' (0.04s) is broken timing.
segment-011 Source sentence 6
forget I said very clean-content, fading, slightly pondering, enjoyable
Emotion 3/5 Delivery 3/5
Part of 'forget I said that'.
segment-012 Source sentence 6
that. very clean-content, fading, slightly pondering, enjoyable
Emotion 3/5 Delivery 2/5
Final word 'that.' followed by breath, unnaturally fragmented.
segment-013 Source sentence 7
Go ahead, Dave, finish what you were gonna say before I jumped in. very full, very chesty, enjoyable
Emotion 4/5 Delivery 3/5
Clear, grounded conversational turn yielding to Dave.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
4/5
Whole-recording emotion fit
4/5
Local speech emotion fit
3/5
Whole-recording talking-style fit
4/5
Local speech delivery fit
3/5
Burst-description fit
1/5
Burst-timing fit
1/5
Event precision rating
1/5
Event coverage rating
3/5
Transcript fidelity (diagnostic)
4/5
Speech-timing fit (diagnostic)
2/5
Readability (diagnostic)
3/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
51 words preserved; 8 selected events serialized; printed elapsed budget 20.450 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English5 selected intervals
The Reconciliation of Ghosts
en-17999 · 27.38 s · Voice: Kore · eric_morris_inspired
3.35 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A feminine-coded, young-adult speaker delivers an exasperated, confrontational monologue in General American English. The vocal delivery is marked by prominent throat and oral resonance, moderate brightness, and noticeable vocal tension that conveys irritation, disappointment, and bitterness. The tempo is deliberate and conversational, broken into short, punchy phrases with clear, crisp articulation. The speaker maintains an assertive, direct interpersonal focus, addressing an interlocutor with sharp emphasis on key words like 'cruel', 'ready', and 'hate'. Two distinct exasperated sighs punctuate the speech, naturally integrated into the pauses between accusations. The recording is studio-clean with minimal background noise and high intelligibility.
SCRIPT — timed speech, local cues and free-form burst descriptions
[0.32 seconds pause]
(very decelerating, very bright, very acrimony, enjoyable; 4.64 seconds duration) Five years of absolute silence and now you just text me?
[0.32 seconds pause]
(very clean-bg, very bright, very resentment, enjoyable; 2.18 seconds duration) I had every single cruel word
[0.38 seconds pause]
(very clean-bg, very bright, very resentment, enjoyable; 0.96 seconds duration) ready. ("A heavy, weary exhalation occurs in the pause following 'ready', releasing built-up tension and signaling reluctant emotional deflation before shifting to a softer, more vulnerable tone.", 0.80 seconds)
(very explicit, clean-bg, slightly pondering, enjoyable; 2.59 seconds duration) ("A heavy, weary exhalation occurs in the pause between 'ready' and 'But then I see', releasing built-up tension and signaling reluctant emotional surrender before shifting to a softer, more vulnerable tone.", 1.04 seconds) But then I see these
[0.21 seconds pause]
(very explicit, clean-bg, slightly pondering, enjoyable; 2.28 seconds duration) faded pencil marks behind your door.
[0.44 seconds pause]
(very darkening, very decelerating, very teasing, enjoyable; 2.23 seconds duration) You cheated on that height chart, remember?
[0.25 seconds pause]
(very hi-fi, very even-rhythm, slightly serenity, enjoyable; 3.60 seconds duration) You stood on your tiptoes every June. ("A soft, breathy exasperated sigh occurs in the pause following 'every June', releasing built-up tension and signaling reluctant emotional surrender before admitting lingering feelings.", 0.68 seconds) ("A heavy, weary exhalation with slight vocalization occurs in the pause after 'every June', conveying deep emotional exhaustion and lingering resentment before the speaker admits 'God, I still hate what you did'.", 1.18 seconds)
(very fading, very whispered, very resentment, enjoyable; 1.56 seconds duration) God, I still hate what you
(very fading, very whispered, very resentment, enjoyable; 0.84 seconds duration) did. ("A soft, audible inhalation taken in the pause before 'But I hate', grounding the speaker's emotional composure and preparing the final vulnerable admission.", 0.64 seconds)
[0.56 seconds pause]
(very clean-bg, very non-instructional, very butterflies in the stomach, enjoyable; 4.02 seconds duration) But I hate how badly I want to hear you explain it.
Automated review note The pipeline double-counted single sighs by producing overlapping duplicate burst events (events 0000/0001 and 0002/0003 cover the same actual sighs around 8.9-9.9s and 19.2-20.3s). Segment timestamps also incorrectly subsume burst pauses directly into speech segments.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
0.80 s duration · HVA_EVENT_0000
A heavy, weary exhalation occurs in the pause following 'ready', releasing built-up tension and signaling reluctant emotional deflation before shifting to a softer, more vulnerable tone.
A heavy, weary exhalation occurs in the pause between 'ready' and 'But then I see', releasing built-up tension and signaling reluctant emotional surrender before shifting to a softer, more vulnerable tone.
A soft, breathy exasperated sigh occurs in the pause following 'every June', releasing built-up tension and signaling reluctant emotional surrender before admitting lingering feelings.
A heavy, weary exhalation with slight vocalization occurs in the pause after 'every June', conveying deep emotional exhaustion and lingering resentment before the speaker admits 'God, I still hate what you did'.
A soft, audible inhalation taken in the pause before 'But I hate', grounding the speaker's emotional composure and preparing the final vulnerable admission.
Local emotion and delivery review — 10 speech segments
Absolute interval
Recognized speech and cue
Gemini ratings
Reviewer explanation
segment-001 Source sentence 1
Five years of absolute silence and now you just text me? very decelerating, very bright, very acrimony, enjoyable
Emotion 4/5 Delivery 4/5
Captures the bitter, exasperated opening well.
segment-002 Source sentence 2
I had every single cruel word very clean-bg, very bright, very resentment, enjoyable
Emotion 4/5 Delivery 4/5
Resentment fits the clipped, tense delivery.
segment-003 Source sentence 2
ready. very clean-bg, very bright, very resentment, enjoyable
Emotion 3/5 Delivery 3/5
Word 'ready' is spoken firmly; note that the sigh actually starts right after it around 8.8s rather than on the word itself.
segment-004 Source sentence 3
But then I see these very explicit, clean-bg, slightly pondering, enjoyable
Emotion 4/5 Delivery 3/5
Tone softens to reflective/pondering; segment timing overlaps with the preceding audible sigh.
segment-005 Source sentence 3
faded pencil marks behind your door. very explicit, clean-bg, slightly pondering, enjoyable
Emotion 4/5 Delivery 4/5
Pensive, slightly subdued delivery matches well.
segment-006 Source sentence 4
You cheated on that height chart, remember? very darkening, very decelerating, very teasing, enjoyable
Emotion 4/5 Delivery 3/5
Gentle, nostalgic teasing is audible.
segment-007 Source sentence 5
You stood on your tiptoes every June. very hi-fi, very even-rhythm, slightly serenity, enjoyable
Emotion 3/5 Delivery 3/5
Affectionate memory, followed by a noticeable sigh in the pause.
segment-008 Source sentence 6
God, I still hate what you very fading, very whispered, very resentment, enjoyable
Emotion 4/5 Delivery 4/5
Whispered, breathy resentment is accurate.
segment-009 Source sentence 6
did. very fading, very whispered, very resentment, enjoyable
Emotion 4/5 Delivery 3/5
Whispered completion of the phrase.
segment-010 Source sentence 7
But I hate how badly I want to hear you explain it. very clean-bg, very non-instructional, very butterflies in the stomach, enjoyable
Emotion 3/5 Delivery 4/5
Vulnerable and conflicted rather than romantic 'butterflies in the stomach'.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
4/5
Whole-recording emotion fit
4/5
Local speech emotion fit
3/5
Whole-recording talking-style fit
4/5
Local speech delivery fit
3/5
Burst-description fit
3/5
Burst-timing fit
2/5
Event precision rating
3/5
Event coverage rating
4/5
Transcript fidelity (diagnostic)
5/5
Speech-timing fit (diagnostic)
3/5
Readability (diagnostic)
4/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
62 words preserved; 5 selected events serialized; printed elapsed budget 27.380 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English0 selected intervals
Conclude Strongly
en-18040 · 22.66 s · Voice: Charon · haiku_batch
3.50 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A single mature adult masculine-coded speaker delivers a formal, authoritative statement in General American English with a deep baritone register. The vocal timbre is grounded with prominent chest resonance, solid warmth, and a clean, controlled harmonic core with minimal roughness. The delivery is measured, deliberate, and highly structured, maintaining a steady, low-to-moderate arousal level and a serious, slightly stern emotional tone. Articulation is crisp and precise throughout, with clear emphasis placed on key legal and moral terms such as 'lie', 'none', 'truth', and 'accountability'. The pacing is slow and segmented into distinct, deliberate breath groups with virtually no disfluencies or hesitations, projecting an unyielding, commanding stance. The recording is technically clean and dry with an acoustically quiet background, though the highly regular, rehearsed cadence gives it a slightly synthetic or read quality. No non-speech vocal bursts are present.
SCRIPT — timed speech, local cues and free-form burst descriptions
[0.32 seconds pause]
(very full, very serious, existential void, enjoyable; 4.34 seconds duration) The timeline doesn't lie, and the physical evidence cannot be wished away.
[0.70 seconds pause]
(very full, very serious, very boredom, enjoyable; 2.21 seconds duration) The defense asked you to look for doubt,
[0.43 seconds pause]
(very full, very serious, very boredom, enjoyable; 1.39 seconds duration) but the facts leave none,
[0.37 seconds pause]
(very full, very serious, very boredom, enjoyable; 0.96 seconds duration) none at all.
[0.32 seconds pause]
(very full, very steady, slightly gladness, enjoyable; 1.84 seconds duration) A family is waiting for the truth
[0.32 seconds pause]
(very full, very steady, slightly gladness, enjoyable; 2.00 seconds duration) to matter in this room.
[0.40 seconds pause]
(very full, very even-rhythm, slightly insensitivity, enjoyable; 2.65 seconds duration) Justice demands accountability, members of the Jerry.
[0.71 seconds pause]
(very denasal, very full, slightly numbness, enjoyable; 1.59 seconds duration) Do not look away from what happened.
[0.57 seconds pause]
(very full, very accelerating, very insensitivity, enjoyable; 1.18 seconds duration) Hold him responsible.
[0.36 seconds pause]
Automated review note Accurate overall performance description, but local emotion tags like boredom and gladness mischaracterize the serious delivery, and jury was misheard as Jerry.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
No interval selected.
Local emotion and delivery review — 9 speech segments
Absolute interval
Recognized speech and cue
Gemini ratings
Reviewer explanation
segment-001 Source sentence 1
The timeline doesn't lie, and the physical evidence cannot be wished away. very full, very serious, existential void, enjoyable
Emotion 2/5 Delivery 3/5
Tone is serious rather than an existential void.
segment-002 Source sentence 2
The defense asked you to look for doubt, very full, very serious, very boredom, enjoyable
Emotion 2/5 Delivery 3/5
Serious emphasis rather than audible boredom.
segment-003 Source sentence 2
but the facts leave none, very full, very serious, very boredom, enjoyable
Emotion 2/5 Delivery 3/5
Firm delivery rather than boredom.
segment-004 Source sentence 2
none at all. very full, very serious, very boredom, enjoyable
Emotion 2/5 Delivery 3/5
Firm emphasis rather than boredom.
segment-005 Source sentence 3
A family is waiting for the truth very full, very steady, slightly gladness, enjoyable
Emotion 2/5 Delivery 3/5
Solemn delivery rather than gladness.
segment-006 Source sentence 3
to matter in this room. very full, very steady, slightly gladness, enjoyable
Emotion 2/5 Delivery 3/5
Solemn tone rather than gladness.
segment-007 Source sentence 4
Justice demands accountability, members of the Jerry. very full, very even-rhythm, slightly insensitivity, enjoyable
Emotion 2/5 Delivery 3/5
Spoken text says jury instead of Jerry.
segment-008 Source sentence 5
Do not look away from what happened. very denasal, very full, slightly numbness, enjoyable
Emotion 2/5 Delivery 3/5
Firm command rather than numbness.
segment-009 Source sentence 6
Hold him responsible. very full, very accelerating, very insensitivity, enjoyable
Emotion 2/5 Delivery 3/5
Pace is steady rather than accelerating.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
4/5
Whole-recording emotion fit
4/5
Local speech emotion fit
2/5
Whole-recording talking-style fit
4/5
Local speech delivery fit
3/5
Burst-description fit
4/5
Burst-timing fit
4/5
Event precision rating
4/5
Event coverage rating
4/5
Transcript fidelity (diagnostic)
3/5
Speech-timing fit (diagnostic)
4/5
Readability (diagnostic)
4/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
57 words preserved; 0 selected events serialized; printed elapsed budget 22.660 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English10 selected intervals
The Thin Ribbon of Blue
en-18796 · 24.62 s · Voice: Fenrir · extreme_physical
3.65 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A young adult masculine-coded speaker delivers an increasingly panicked, breathless monologue in General American English, depicting a character trapped in a failing environment. The performance begins with a relatively controlled, curious observation before rapidly deteriorating into acute distress, fear, and helplessness. The vocal timbre is moderately bright with prominent throat and oral resonance, marked by high breathiness, audible gasps, and strained glottal attacks. Phrasing is fragmented and choppy, interrupted by frequent gasps, panting, and a frustrated groan as the speaker realizes the dire situation. The recording is dry and clean with negligible background noise, capturing the actor's escalating physiological panic with high dramatic believability.
SCRIPT — timed speech, local cues and free-form burst descriptions
[0.24 seconds pause]
(very wide-pitch, tonal, awe, enjoyable; 1.86 seconds duration) You can see why the gods left.
[0.38 seconds pause]
(very clean-bg, very negative, foreboding, enjoyable; 2.10 seconds duration) There is so much air down there
[0.22 seconds pause]
(very clean-bg, very negative, foreboding, enjoyable; 2.49 seconds duration) and they waste it fighting over puddles.
[0.47 seconds pause]
(very fading, very fading, very insensitivity, enjoyable; 1.52 seconds duration) They don't even look up.
[0.26 seconds pause]
("A sharp, constricted gasp of air taken in sudden alarm right before the speaker interrupts their own thought with 'Wait'.", 0.68 seconds)
[0.42 seconds pause]
(very scattered, very fading, stoicism, less-enjoyable; 0.42 seconds duration) Wait.
[0.62 seconds pause]
(clean-bg, naturalistic, very perplexity; 0.91 seconds duration) What was that drop?
[1.33 seconds pause]
("A sudden, constricted gasp of air taken in shock and panic right before the speaker stammers 'I can't', showing an abrupt shift from detached observation to acute distress.", 0.62 seconds)
[0.34 seconds pause]
(very meek, very whispered, very burnout, enjoyable; 0.30 seconds duration) Can't
(very meek, very whispered, very burnout, enjoyable; 1.03 seconds duration) ("A quick, shallow panting breath taken in isolation between fragmented words, reflecting acute respiratory strain.", 0.52 seconds) breathe.
[0.31 seconds pause]
(very breathy, very whispered, very powerlessness, enjoyable; 1.50 seconds duration) Mask ("A labored, raspy inhalation and exhalation cycle occurring in isolation between 'can't' and 'breathe', demonstrating severe respiratory strain.", 1.06 seconds) won't ("A quick, constricted gasp of air taken in a pause, reflecting mounting panic and physical strain.", 0.56 seconds)
[0.38 seconds pause]
(very breathy, very whispered, very powerlessness, enjoyable; 1.06 seconds duration) No air up here.
(very breathy, very vulnerable, very helplessness, enjoyable; 1.46 seconds duration) ("A quick, shallow panting breath between 'No air up here' and 'Pilot', showing escalating panic and physical strain.", 0.68 seconds) Pilot ("A quick, shallow panting breath between 'Pilot' and 'Pressure', showing rapid respiratory distress.", 0.54 seconds)
[0.30 seconds pause]
(very breathy, very vulnerable, very helplessness, enjoyable; 1.38 seconds duration) ("A strained, audible intake of air following 'Pilot', showing acute respiratory struggle and panic.", 0.60 seconds) Pressure, I can't
(overlapping non-speech vocal events; 0.98 seconds duration) ("A sharp, constricted gasp of air taken in desperation right before the final whispered word 'Helpless'.", 0.58 seconds) ("A faint, breathy vocalized groan of helplessness and physical strain just before the final whispered word 'Helpless'.", 0.68 seconds)
(very breathy, very vulnerable, very helplessness, enjoyable; 0.74 seconds duration) helpless.
[0.30 seconds pause]
Automated review note The script breaks up words and overlapping breath events heavily in the latter half. Several burst annotations overlap with spoken words or are redundant duplicates (such as multiple overlapping breaths around 16.5-18.0s and 22.6-23.5s).
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
0.68 s duration · HVA_EVENT_0000
A sharp, constricted gasp of air taken in sudden alarm right before the speaker interrupts their own thought with 'Wait'.
A sudden, constricted gasp of air taken in shock and panic right before the speaker stammers 'I can't', showing an abrupt shift from detached observation to acute distress.
Local emotion and delivery review — 13 speech segments
Absolute interval
Recognized speech and cue
Gemini ratings
Reviewer explanation
segment-001 Source sentence 1
You can see why the gods left. very wide-pitch, tonal, awe, enjoyable
Emotion 4/5 Delivery 4/5
Well matched; the opening delivery is clear, thoughtful, and reflective.
segment-002 Source sentence 2
There is so much air down there very clean-bg, very negative, foreboding, enjoyable
Emotion 3/5 Delivery 4/5
Delivery fits well, though 'foreboding' / 'very negative' is slightly overstated for this reflective observation.
segment-003 Source sentence 2
and they waste it fighting over puddles. very clean-bg, very negative, foreboding, enjoyable
Emotion 4/5 Delivery 4/5
Accurately captures the disdainful, cynical tone and clean acoustic profile.
segment-004 Source sentence 3
They don't even look up. very fading, very fading, very insensitivity, enjoyable
Emotion 3/5 Delivery 4/5
'Insensitivity' is a slight mischaracterization of the tone, which is dismissive and quiet.
segment-005 Source sentence 4
Wait. very scattered, very fading, stoicism, less-enjoyable
Emotion 3/5 Delivery 3/5
'Stoicism' does not fit; the speaker sounds startled and attentive after hearing something.
segment-006 Source sentence 5
What was that drop? clean-bg, naturalistic, very perplexity
Emotion 4/5 Delivery 4/5
Perplexity and naturalistic questioning delivery are accurate.
segment-007 Source sentence 6
Can't very meek, very whispered, very burnout, enjoyable
Emotion 3/5 Delivery 3/5
Segment boundaries are heavily fragmented around the gasp and struggle; burnout tag does not capture acute panic.
segment-008 Source sentence 6
breathe. very meek, very whispered, very burnout, enjoyable
Emotion 4/5 Delivery 4/5
Whispered, breathy struggle to breathe is represented adequately.
segment-009 Source sentence 7
Mask won't very breathy, very whispered, very powerlessness, enjoyable
Emotion 4/5 Delivery 4/5
Helpless, breathless whisper-talk matches well.
segment-010 Source sentence 7
No air up here. very breathy, very whispered, very powerlessness, enjoyable
Emotion 4/5 Delivery 4/5
Fits the breathless, terrified realization well.
segment-011 Source sentence 8
Pilot very breathy, very vulnerable, very helplessness, enjoyable
Emotion 4/5 Delivery 4/5
Captures the acute distress and vulnerability.
segment-012 Source sentence 8
Pressure, I can't very breathy, very vulnerable, very helplessness, enjoyable
Emotion 4/5 Delivery 4/5
Strained, choked delivery accurately reflects the tagging.
segment-013 Source sentence 8
helpless. very breathy, very vulnerable, very helplessness, enjoyable
Emotion 4/5 Delivery 4/5
Final breathless resignation is captured appropriately.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
4/5
Whole-recording emotion fit
4/5
Local speech emotion fit
3/5
Whole-recording talking-style fit
4/5
Local speech delivery fit
4/5
Burst-description fit
4/5
Burst-timing fit
3/5
Event precision rating
3/5
Event coverage rating
4/5
Transcript fidelity (diagnostic)
4/5
Speech-timing fit (diagnostic)
3/5
Readability (diagnostic)
4/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
44 words preserved; 10 selected events serialized; printed elapsed budget 24.620 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English7 selected intervals
Final Phone Call from the ICU
en-18876 · 28.58 s · Voice: Sadaltager · existing_challenge_inspired
3.15 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A young adult masculine-coded speaker delivers an intimate, deeply vulnerable confession in General American English. The vocal delivery is exceptionally soft, quiet, and breathy, hovering between a half-whisper and an ASMR-like close-mic proximity. The tempo is slow and measured with frequent, deliberate pauses and audible, heavy breathing that conveys profound emotional exposure, relief, and affection. The timbre is warm and grounded in chest resonance, though slightly raspy and breath-depleted due to the hushed register. The performance feels highly genuine and lived-in, with natural respiratory resets and sighs seamlessly integrated into the spoken phrases. The recording is pristine, dry, and close, with virtually no background noise.
SCRIPT — timed speech, local cues and free-form burst descriptions
(very noisy, very whispered, very weariness; 1.12 seconds duration) Hey ("A slow, audible, close-mic inhalation and soft exhalation precedes the opening word, establishing an intimate, weary, and vulnerable tone.", 1.08 seconds)
[0.72 seconds pause]
(very noisy, very whispered, very weariness; 0.81 seconds duration) it's me.
[0.71 seconds pause]
(very fading, very fading, emotional blunting; 1.52 seconds duration) I know it's late.
[0.80 seconds pause]
(very naturalistic, denasal, craving; 2.40 seconds duration) Just wanted to hear your voice for a second.
(very breathy, very vulnerable, very shyness; 3.76 seconds duration) ("A heavy, weary exhalation and breath release occurs in the pause before 'God, I should have said this', conveying deep emotional weight and regret.", 0.96 seconds) ("A heavy, weary exhalation that functions as an exasperated sigh, leading directly into the phrase 'God, I should have said this 20 years ago'.", 0.88 seconds) God, I should have said this twenty years ago, but
[0.96 seconds pause]
(very breathy, very vulnerable, very shyness; 1.84 seconds duration) I was so proud of you.
[0.32 seconds pause]
(very thin-mask, very fading, infatuation; 0.49 seconds duration) Always.
[0.95 seconds pause]
(very noisy, very denasal, very guilt, enjoyable; 2.44 seconds duration) I'm sorry I made everything so damn hard.
(very organic, very fading, very weariness; 2.84 seconds duration) ("A soft, prolonged exhalation of relief and emotional release following 'so damn hard', releasing built-up tension before offering reassurance.", 0.86 seconds) ("A soft, prolonged exhalation of relief and emotional release following the admission of making things hard, leading into a gentle reassurance.", 1.06 seconds) Don't worry about me tonight.
[0.48 seconds pause]
(very fading, very organic, stoicism; 0.55 seconds duration) I'm okay.
[1.29 seconds pause]
(very dissociated, very fading, emotional blunting; 0.46 seconds duration) Really.
[0.42 seconds pause]
(very whispered, very breathy, very comfort, enjoyable; 0.30 seconds duration) Just
(overlapping non-speech vocal events; 1.56 seconds duration) ("A soft, warm exhalation of relief and affection occurs in the pause before 'I'm so glad you're in the world', carrying a gentle release of tension and deep emotional tenderness.", 0.98 seconds) ("A soft, warm exhalation of relief and affection occurs in the pause before 'I'm so glad you're in the world', carrying a gentle release of tension and deep emotional warmth.", 0.96 seconds)
(very whispered, very breathy, very comfort, enjoyable; 1.44 seconds duration) so glad you're in the world.
[0.40 seconds pause]
Automated review note The burst event list contains redundant duplicate events that heavily overlap each other (e.g., pairs of near-identical sighs at ~8.5-9.5s, ~19-20.5s, and ~25.5-26.7s), resulting in degraded burst precision and timing scores despite the sound events actually being real sigh/breath releases.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
1.08 s duration · HVA_EVENT_0000
A slow, audible, close-mic inhalation and soft exhalation precedes the opening word, establishing an intimate, weary, and vulnerable tone.
Top-three type softmax: Exasperated Sigh: 19.8%; Exhausted Groan: 15.3%; Other / rare burst: 9.9%
0.96 s duration · HVA_EVENT_0001
A heavy, weary exhalation and breath release occurs in the pause before 'God, I should have said this', conveying deep emotional weight and regret.
A soft, warm exhalation of relief and affection occurs in the pause before 'I'm so glad you're in the world', carrying a gentle release of tension and deep emotional tenderness.
A soft, warm exhalation of relief and affection occurs in the pause before 'I'm so glad you're in the world', carrying a gentle release of tension and deep emotional warmth.
Local emotion and delivery review — 13 speech segments
Absolute interval
Recognized speech and cue
Gemini ratings
Reviewer explanation
segment-001 Source sentence 1
Hey very noisy, very whispered, very weariness
Emotion 3/5 Delivery 4/5
Hey is delivered softly and intimately, though labelling it as extremely weariness is somewhat exaggerated.
segment-002 Source sentence 1
it's me. very noisy, very whispered, very weariness
Emotion 3/5 Delivery 4/5
Soft and quiet delivery matches well, mild emotional fatigue.
segment-003 Source sentence 2
I know it's late. very fading, very fading, emotional blunting
Emotion 2/5 Delivery 3/5
Labelled as emotional blunting, but the voice sounds tender, vulnerable, and regretful rather than emotionally blunt.
segment-004 Source sentence 3
Just wanted to hear your voice for a second. very naturalistic, denasal, craving
Emotion 4/5 Delivery 4/5
Craving/longing fits the context and delivery well.
segment-005 Source sentence 4
God, I should have said this twenty years ago, but very breathy, very vulnerable, very shyness
Emotion 3/5 Delivery 4/5
Tags shyness/embarrassment, whereas the audible affect is more sorrowful regret and vulnerability.
segment-006 Source sentence 4
I was so proud of you. very breathy, very vulnerable, very shyness
Emotion 3/5 Delivery 4/5
Tagging shyness is slightly off; it is sincere and emotionally warm pride.
segment-007 Source sentence 5
Always. very thin-mask, very fading, infatuation
Emotion 3/5 Delivery 3/5
Brief whisper-spoken tag 'infatuation' is a bit overly specific, but fits the tender tone.
segment-008 Source sentence 6
I'm sorry I made everything so damn hard. very noisy, very denasal, very guilt, enjoyable
Emotion 4/5 Delivery 4/5
Guilt tag accurately reflects the delivery of the apology.
segment-009 Source sentence 7
Don't worry about me tonight. very organic, very fading, very weariness
Emotion 3/5 Delivery 3/5
Tender and reassuring tone, weariness is moderately present.
segment-010 Source sentence 8
I'm okay. very fading, very organic, stoicism
Emotion 3/5 Delivery 3/5
Stoicism tag is reasonable for 'I'm okay'.
segment-011 Source sentence 9
Really. very dissociated, very fading, emotional blunting
Emotion 2/5 Delivery 3/5
Emotional blunting and dissociation do not quite match the intimate, sincere tone.
segment-012 Source sentence 10
Just very whispered, very breathy, very comfort, enjoyable
Emotion 3/5 Delivery 4/5
Short whispery connector word.
segment-013 Source sentence 10
so glad you're in the world. very whispered, very breathy, very comfort, enjoyable
Emotion 4/5 Delivery 4/5
Comfort, warmth, and relief fit the tender conclusion.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
4/5
Whole-recording emotion fit
4/5
Local speech emotion fit
3/5
Whole-recording talking-style fit
4/5
Local speech delivery fit
3/5
Burst-description fit
2/5
Burst-timing fit
2/5
Event precision rating
2/5
Event coverage rating
3/5
Transcript fidelity (diagnostic)
5/5
Speech-timing fit (diagnostic)
4/5
Readability (diagnostic)
3/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
56 words preserved; 7 selected events serialized; printed elapsed budget 28.580 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
English7 selected intervals
Post-Apocalyptic Radio Broadcast: Is Anyone Out There?
en-19162 · 34.10 s · Voice: Umbriel · voicenet_extension
3.30 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A single young-adult to mature-adult masculine-coded speaker delivers a dramatic, radio-transmission-style monologue in General American English. The performance begins in a low, weary, and contemplative baritone register with noticeable vocal fry, breathiness, and deliberate pacing. The speaker reads a call sign and repeats a frequency before expressing profound fatigue and disorientation. As the monologue progresses, the delivery shifts from quiet, low-arousal exhaustion to sudden, startled confusion and mild panic. This shift is marked by a sharp, audible gasp of surprise around the 20-second mark, followed by a quickened tempo, elevated pitch, and increased vocal tension as the speaker questions their surroundings and demands to be tapped on the microphone. The acoustic environment is clean and dry with negligible background noise, and the vocal bursts integrate naturally into the dramatic performance.
SCRIPT — timed speech, local cues and free-form burst descriptions
[0.32 seconds pause]
(very accelerating, very hi-fi, very emotional blunting; 1.37 seconds duration) This is station seven.
[1.27 seconds pause]
(very full, very serious, slightly absorption, enjoyable; 0.66 seconds duration) Anyone,
[0.38 seconds pause]
(very full, very serious, slightly absorption, enjoyable; 1.47 seconds duration) if anyone can hear this,
[0.29 seconds pause]
(very full, very serious, slightly absorption, enjoyable; 2.36 seconds duration) respond on frequency one four two.
[0.22 seconds pause]
(very steady, very full, slightly absorption, enjoyable; 2.20 seconds duration) ("A slow, audible inhalation and exhalation cycle occurs during a long pause between radio calls, reflecting physical fatigue and mental weariness.", 1.22 seconds) ("A slow, audible inhalation and exhalation cycle occurs during a pause between repeated calls for response, reflecting physical fatigue and mental strain.", 0.92 seconds) Repeating
[0.42 seconds pause]
(very steady, very full, slightly absorption, enjoyable; 1.92 seconds duration) frequency one four two.
[1.76 seconds pause]
(very fading, very organic, very contemplation, enjoyable; 0.99 seconds duration) It's Tuesday,
[0.37 seconds pause]
(very fading, very organic, very contemplation, enjoyable; 0.64 seconds duration) I think.
[0.56 seconds pause]
(very accelerating, very effortless, very stoicism, enjoyable; 0.87 seconds duration) Overcast.
[0.81 seconds pause]
(very denasal, very full, very apathy, enjoyable; 1.87 seconds duration) Not that the clouds care much anymore.
[0.23 seconds pause]
("A soft, weary exhalation and breath release occurs in the pause after 'anymore', conveying fatigue and mild resignation before resuming speech.", 0.74 seconds)
[0.68 seconds pause]
(very full, very denasal, mocking playfully, enjoyable; 1.98 seconds duration) You know, I almost missed traffic, James.
[0.50 seconds pause]
(very fading, very naturalistic, very nostalgia, enjoyable; 1.69 seconds duration) Just to sit near someone honking.
[0.61 seconds pause]
(very non-narrative, very accelerating; 0.96 seconds duration) ("A quick, audible gasp-like sharp intake of breath occurs right before the speaker calls out 'Hello?', signaling sudden alertness and mild startle.", 0.70 seconds) Hello?
[0.28 seconds pause]
(very decelerating, very non-narrative, very surprise, less-enjoyable; 0.62 seconds duration) ("A sudden, sharp intake of breath occurs right after the speaker says 'Hello?', signaling an abrupt startle or sudden realization of an unexpected sound in the background.", 0.52 seconds) Huh?
(very non-narrative, very scattered, very disorientation; 0.46 seconds duration) ("A sudden, sharp intake of breath occurs right after the speaker says 'Hello?', signaling an abrupt startle or sudden realization of an unexpected sound before asking 'Was that?'.", 0.50 seconds) Was that?
[0.28 seconds pause]
(very scattered, non-narrative, very skepticism, enjoyable; 1.66 seconds duration) ("A quick, tense inhalation taken between 'Was that?' and 'Is someone there?', reflecting sudden alertness and mild panic.", 0.76 seconds) Is someone there?
[0.40 seconds pause]
(naturalistic, steady, slightly powerlessness, enjoyable; 0.47 seconds duration) Please
[0.25 seconds pause]
(naturalistic, steady, slightly powerlessness, enjoyable; 1.14 seconds duration) just tap the mic.
[0.40 seconds pause]
Automated review note The script hallucinates multiple duplicate breath events in overlapping time windows (around 8.3-10.2s and 28.4-29.4s). Transcript error at 00:23 ('traffic jams' transcribed as 'traffic, James'). Overall acoustic description is solid.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
1.22 s duration · HVA_EVENT_0000
A slow, audible inhalation and exhalation cycle occurs during a long pause between radio calls, reflecting physical fatigue and mental weariness.
A slow, audible inhalation and exhalation cycle occurs during a pause between repeated calls for response, reflecting physical fatigue and mental strain.
A sudden, sharp intake of breath occurs right after the speaker says 'Hello?', signaling an abrupt startle or sudden realization of an unexpected sound in the background.
A sudden, sharp intake of breath occurs right after the speaker says 'Hello?', signaling an abrupt startle or sudden realization of an unexpected sound before asking 'Was that?'.
just tap the mic. naturalistic, steady, slightly powerlessness, enjoyable
Emotion 4/5 Delivery 4/5
Quiet pleading desperation.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
4/5
Whole-recording emotion fit
4/5
Local speech emotion fit
3/5
Whole-recording talking-style fit
4/5
Local speech delivery fit
3/5
Burst-description fit
3/5
Burst-timing fit
2/5
Event precision rating
2/5
Event coverage rating
4/5
Transcript fidelity (diagnostic)
4/5
Speech-timing fit (diagnostic)
3/5
Readability (diagnostic)
4/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
58 words preserved; 7 selected events serialized; printed elapsed budget 34.100 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
German7 selected intervals
The Inheritance of Wounds
de-00093 · 26.78 s · Voice: Erinome · eric_morris_inspired
3.45 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A young adult feminine-coded speaker delivers an emotionally charged, dramatic confrontation in Standard German. The performance begins with sharp, accusatory directness and high vocal tension, marked by clear articulation and forward oral-mask resonance. As the monologue progresses, the speaker's composure unravels into profound distress, shame, and helplessness, shifting from an assertive, bitter tone to a hushed, breathy, and tearful whisper. The acoustic environment is completely clean and dry with professional studio clarity, allowing subtle respiratory catches, trembling vocal fry, and quiet sobs to integrate seamlessly into the spoken delivery.
SCRIPT — timed speech, local cues and free-form burst descriptions
(very even-rhythm, very hi-fi, very desire, enjoyable; 2.94 seconds duration) Jahrelang dachte ich, du hättest uns einfach im Stich gelassen.
[0.34 seconds pause]
(very even-rhythm, hi-fi, slightly anticipation, enjoyable; 3.36 seconds duration) Ich habe mir jedes einzelne Wort zurechtgelegt, um dir weh zu tun.
[0.32 seconds pause]
(very meek, very fading, very bewilderment; 0.50 seconds duration) Aber ("A sharp, audible intake of breath occurs in the pause before 'Aber', signaling emotional distress and preparing the speaker for an accusatory question.", 0.48 seconds)
[0.78 seconds pause]
(very meek, very fading, very bewilderment; 2.40 seconds duration) warum zitterst du so? ("A sharp, audible intake of breath occurs in the pause between 'warum zitterst du so?' and 'Das', marked by sudden throat constriction and emotional distress, blending naturally into the trembling onset of the following word.", 0.86 seconds)
(very whispered, very breathy, very amazement; 0.40 seconds duration) Das
[1.12 seconds pause]
(very whispered, very breathy, very amazement; 0.55 seconds duration) war er.
[1.05 seconds pause]
(very smooth, darkening, very acerbity, enjoyable; 2.49 seconds duration) Er hat dir das angetan, damit er mich nicht anfasst.
[1.11 seconds pause]
(very breathy, very meek, very weariness, enjoyable; 0.39 seconds duration) Gott,
(very breathy, very meek, very weariness, enjoyable; 2.05 seconds duration) ("A soft, tremulous breathy whimper and catch in the throat right before pleading 'Hey, sieh mich an', conveying deep vulnerability and emotional distress.", 0.54 seconds) hey, sieh mich an. ("A soft, tremulous breathy whimper and catch in the throat occurs right after 'sieh mich an', conveying deep vulnerability and emotional pleading.", 0.58 seconds)
(very breathy, very thin, slightly exhaustion; 1.64 seconds duration) Bitte hör auf zu weiten. ("A soft, tremulous breathy release and quiet sob-like exhale following 'aufzuhören', conveying deep sorrow and emotional exhaustion.", 0.58 seconds)
(very meek, very scattered, very respite; 0.83 seconds duration) Es ist vorbei.
[0.29 seconds pause]
(very whispered, very asmr, very reverence; 1.00 seconds duration) Ich bin hier, Mama.
("A soft, tremulous sobbing breath and quiet catch in the throat following 'Ich bin hier, Mama', conveying deep sorrow and emotional release.", 0.88 seconds)
[0.04 seconds pause]
("A soft, tremulous sobbing breath and catch in the throat precedes the final apology 'Es tut mir so leid', blending seamlessly into the tearful, whispered delivery.", 0.60 seconds)
(very whispered, very meek, very shame, less-enjoyable; 1.31 seconds duration) Es tut mir so leid.
[0.39 seconds pause]
Automated review note The monologue shifts from cold confrontation to broken, whispered distress. In early segments, automated emotional tags ('desire', 'anticipation') completely miss the accusatory anger. Several burst events over-segment continuous speech or contain normal syllable releases, but captured gasps and sobs are genuinely present. Transcript contains a typo ('weiten' instead of 'weinen').
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
0.48 s duration · HVA_EVENT_0000
A sharp, audible intake of breath occurs in the pause before 'Aber', signaling emotional distress and preparing the speaker for an accusatory question.
A sharp, audible intake of breath occurs in the pause between 'warum zitterst du so?' and 'Das', marked by sudden throat constriction and emotional distress, blending naturally into the trembling onset of the following word.
A soft, tremulous breathy whimper and catch in the throat right before pleading 'Hey, sieh mich an', conveying deep vulnerability and emotional distress.
A soft, tremulous sobbing breath and catch in the throat precedes the final apology 'Es tut mir so leid', blending seamlessly into the tearful, whispered delivery.
Local emotion and delivery review — 13 speech segments
Absolute interval
Recognized speech and cue
Gemini ratings
Reviewer explanation
segment-001 Source sentence 1
Jahrelang dachte ich, du hättest uns einfach im Stich gelassen. very even-rhythm, very hi-fi, very desire, enjoyable
Emotion 1/5 Delivery 3/5
The speaker sounds stern and reproachful, not expressing longing or desire.
segment-002 Source sentence 2
Ich habe mir jedes einzelne Wort zurechtgelegt, um dir weh zu tun. very even-rhythm, hi-fi, slightly anticipation, enjoyable
Emotion 2/5 Delivery 3/5
Tone is cold, hurt, and confrontational, not hopeful anticipation.
segment-003 Source sentence 3
Aber very meek, very fading, very bewilderment
Emotion 4/5 Delivery 4/5
Delivery falters with confusion and sudden hesitation as labeled.
segment-004 Source sentence 3
warum zitterst du so? very meek, very fading, very bewilderment
Emotion 4/5 Delivery 4/5
Bewildered and concerned question matching the delivery tags well.
segment-005 Source sentence 4
Das very whispered, very breathy, very amazement
Emotion 4/5 Delivery 4/5
Quiet shocked whisper captures the sudden realization.
segment-006 Source sentence 4
war er. very whispered, very breathy, very amazement
Emotion 4/5 Delivery 4/5
Breathy, stunned delivery fits the realization well.
segment-007 Source sentence 5
Er hat dir das angetan, damit er mich nicht anfasst. very smooth, darkening, very acerbity, enjoyable
Emotion 3/5 Delivery 3/5
Delivery is soft, distressed, and horrified rather than smooth acerbity.
segment-008 Source sentence 6
Gott, very breathy, very meek, very weariness, enjoyable
Emotion 3/5 Delivery 3/5
A gasping whisper of shock rather than simple weariness.
segment-009 Source sentence 6
hey, sieh mich an. very breathy, very meek, very weariness, enjoyable
Emotion 4/5 Delivery 4/5
Desperate, pleading whisper fits the tags well.
segment-010 Source sentence 7
Bitte hör auf zu weiten. very breathy, very thin, slightly exhaustion
Emotion 4/5 Delivery 3/5
Transcript mishears 'weinen' as 'weiten'; tone is tearful and pleading.
segment-011 Source sentence 8
Es ist vorbei. very meek, very scattered, very respite
Emotion 4/5 Delivery 3/5
Reassuring whisper, quiet and gentle.
segment-012 Source sentence 9
Ich bin hier, Mama. very whispered, very asmr, very reverence
Emotion 4/5 Delivery 4/5
Affectionate comforting whisper fits the tags reasonably.
segment-013 Source sentence 10
Es tut mir so leid. very whispered, very meek, very shame, less-enjoyable
Emotion 4/5 Delivery 4/5
Sorrowful, apologetic whisper matches shame and meek delivery.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
4/5
Whole-recording emotion fit
4/5
Local speech emotion fit
3/5
Whole-recording talking-style fit
4/5
Local speech delivery fit
3/5
Burst-description fit
3/5
Burst-timing fit
3/5
Event precision rating
3/5
Event coverage rating
4/5
Transcript fidelity (diagnostic)
4/5
Speech-timing fit (diagnostic)
3/5
Readability (diagnostic)
4/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
62 words preserved; 7 selected events serialized; printed elapsed budget 26.780 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
German13 selected intervals
The Fever Dream of Forgiveness
de-00725 · 33.34 s · Voice: Achernar · eric_morris_inspired
3.75 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A young adult feminine-coded speaker delivers an intensely dramatic, breathless monologue in Standard German, depicting severe physical distress, suffocation, and helplessness. The vocal production is heavily unvoiced and whispered, characterized by low harmonicity, prominent throat resonance, and audible respiratory strain. The tempo is very slow and fragmented into short, labored breath groups separated by frequent gasps, sharp inhales, and exhausted groans. The emotional tone is dominated by acute fear, physical pain, and despair, with a slight downward shift toward numb, resigned exhaustion toward the end. The recording is dry and clean with minimal background interference, capturing intimate vocal-tract friction and breath dynamics.
SCRIPT — timed speech, local cues and free-form burst descriptions
(very whispered, very breathy, very fatigue, enjoyable; 1.68 seconds duration) ("A sharp, labored intake of breath precedes the first spoken words, establishing immediate physical distress and breathlessness.", 0.92 seconds) Es fühlt sich an
[0.24 seconds pause]
(very whispered, very breathy, very fatigue, enjoyable; 0.80 seconds duration) wie ("A sharp, constricted inhalation occurs in the pause between 'Es fühlt sich an wie' and 'nasse Beton', marked by audible throat friction and breathiness that reflects physical distress.", 0.86 seconds)
(very whispered, very breathy, very fatigue, enjoyable; 2.23 seconds duration) ("A sharp, constricted inhalation occurs in the pause between 'Es fühlt sich an wie' and 'nasse Beton', marked by audible throat friction and breathiness that reflects physical distress.", 0.98 seconds) nasser Beton auf meinen Rippen.
(very whispered, very vulnerable, very alarm, enjoyable; 4.86 seconds duration) ("A labored, audible inhalation and exhalation cycle occurring in the pause after 'auf meinen Rippen', demonstrating physical strain and breathlessness.", 0.70 seconds) Ich versuche ruhig zu atmen, wirklich, aber der Druck bricht einfach durch.
(very whispered, very breathy, very physical pain, enjoyable; 2.79 seconds duration) ("A strained, breathy vocalized groan of physical agony and breathlessness following 'durch', blending directly into the next phrase.", 0.76 seconds) ("A strained, breathy vocalized groan of physical agony and breathlessness, emerging directly from the word 'durch'.", 0.70 seconds) Mein ganzer Nacken brennt,
[0.52 seconds pause]
(very whispered, very breathy, very physical pain, enjoyable; 1.58 seconds duration) als würde jemand zudrücken.
[0.50 seconds pause]
(very whispered, very negative, very submission, enjoyable; 1.35 seconds duration) Ich kann mich kaum halten.
[0.33 seconds pause]
(overlapping non-speech vocal events; 1.68 seconds duration) ("A labored, shuddering breath cycle with audible friction in the throat, reflecting ongoing physical struggle.", 1.06 seconds) ("A quick, shallow breath intake and release in the pause before 'Es zieht nach unten', maintaining the labored respiratory state.", 0.68 seconds)
(very hi-fi, very even-rhythm, numbness; 1.30 seconds duration) Es zieht nach unten.
("A slow, deep audible inhalation and exhalation cycle, marking a momentary pause in the physical struggle.", 0.88 seconds)
(very whispered, very breathy, very solace, enjoyable; 1.86 seconds duration) Und ("A slow, labored breath cycle with audible friction in the throat, marking a pause before the word 'Und'.", 0.88 seconds) plötzlich ("A quick, shallow breath intake and release between 'Und plötzlich' and 'lässt es nach', marking a sudden physiological shift.", 0.74 seconds)
[0.30 seconds pause]
("A slow, deep audible inhalation and release in the pause before 'Und plötzlich', marking a shift toward quiet contemplation.", 0.94 seconds)
(very whispered, very breathy, very solace, enjoyable; 1.70 seconds duration) lässt es nach. ("A soft, strained intake of air following 'nach', leading into the hushed, breathy delivery of 'Es ist nicht weg'.", 0.84 seconds)
[0.38 seconds pause]
(very whispered, very meek, very mournfulness, enjoyable; 0.81 seconds duration) Es ist nicht weg,
[0.31 seconds pause]
(very whispered, very meek, very mournfulness, enjoyable; 0.07 seconds duration) aber
[0.25 seconds pause]
(very whispered, very meek, very mournfulness, enjoyable; 0.87 seconds duration) es wird still.
[1.21 seconds pause]
(very whispered, very asmr, very awestruck, enjoyable; 0.88 seconds duration) Wie kühles
[0.24 seconds pause]
(very whispered, very asmr, very awestruck, enjoyable; 0.24 seconds duration) Wasser
[0.72 seconds pause]
(very whispered, very asmr, very awestruck, enjoyable; 1.43 seconds duration) nach einem langen Brand.
[0.39 seconds pause]
Automated review note The global description captures the whispered, breathless, dramatic physical distress monologue accurately. However, several burst events overlap heavily with each other or directly over spoken speech segments rather than pauses.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
0.92 s duration · HVA_EVENT_0000
A sharp, labored intake of breath precedes the first spoken words, establishing immediate physical distress and breathlessness.
A sharp, constricted inhalation occurs in the pause between 'Es fühlt sich an wie' and 'nasse Beton', marked by audible throat friction and breathiness that reflects physical distress.
A sharp, constricted inhalation occurs in the pause between 'Es fühlt sich an wie' and 'nasse Beton', marked by audible throat friction and breathiness that reflects physical distress.
Local emotion and delivery review — 16 speech segments
Absolute interval
Recognized speech and cue
Gemini ratings
Reviewer explanation
segment-001 Source sentence 1
Es fühlt sich an very whispered, very breathy, very fatigue, enjoyable
Emotion 4/5 Delivery 4/5
Whispered, breathy delivery fits well.
segment-002 Source sentence 1
wie very whispered, very breathy, very fatigue, enjoyable
Emotion 4/5 Delivery 4/5
Short whispered word with audible inhalation.
segment-003 Source sentence 1
nasser Beton auf meinen Rippen. very whispered, very breathy, very fatigue, enjoyable
Emotion 4/5 Delivery 4/5
Labored, strained whisper fits the exhaustion tags.
segment-004 Source sentence 2
Ich versuche ruhig zu atmen, wirklich, aber der Druck bricht einfach durch. very whispered, very vulnerable, very alarm, enjoyable
Emotion 4/5 Delivery 4/5
Displays distress and vulnerability as labeled.
segment-005 Source sentence 3
Mein ganzer Nacken brennt, very whispered, very breathy, very physical pain, enjoyable
Emotion 4/5 Delivery 4/5
Whispered, pain-tinged delivery matches well.
segment-006 Source sentence 3
als würde jemand zudrücken. very whispered, very breathy, very physical pain, enjoyable
Emotion 4/5 Delivery 4/5
Delivery fits the physical pain context.
segment-007 Source sentence 4
Ich kann mich kaum halten. very whispered, very negative, very submission, enjoyable
Emotion 4/5 Delivery 4/5
Despairing and helpless delivery matches tags.
segment-008 Source sentence 5
Es zieht nach unten. very hi-fi, very even-rhythm, numbness
Emotion 3/5 Delivery 3/5
Delivery remains strained and breathless rather than evenly paced.
segment-009 Source sentence 6
Und plötzlich very whispered, very breathy, very solace, enjoyable
Emotion 4/5 Delivery 4/5
Breathy, transitioning tone matches reasonably well.
segment-010 Source sentence 6
lässt es nach. very whispered, very breathy, very solace, enjoyable
Emotion 4/5 Delivery 4/5
Sense of mild release matches solace/relief tags.
segment-011 Source sentence 7
Es ist nicht weg, very whispered, very meek, very mournfulness, enjoyable
Emotion 4/5 Delivery 4/5
Quiet, gentle sad whisper.
segment-012 Source sentence 7
aber very whispered, very meek, very mournfulness, enjoyable
Emotion 4/5 Delivery 4/5
Very short transitional segment, delivery fits.
segment-013 Source sentence 7
es wird still. very whispered, very meek, very mournfulness, enjoyable
Emotion 4/5 Delivery 4/5
Very quiet, still whispered delivery.
segment-014 Source sentence 8
Wie kühles very whispered, very asmr, very awestruck, enjoyable
Emotion 4/5 Delivery 4/5
Soft, hushed ASMR-like delivery fits.
segment-015 Source sentence 8
Wasser very whispered, very asmr, very awestruck, enjoyable
Emotion 4/5 Delivery 4/5
Short whisper continuation fits.
segment-016 Source sentence 8
nach einem langen Brand. very whispered, very asmr, very awestruck, enjoyable
Emotion 4/5 Delivery 4/5
Quiet, contemplative, intimate finish.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
4/5
Whole-recording emotion fit
4/5
Local speech emotion fit
4/5
Whole-recording talking-style fit
4/5
Local speech delivery fit
4/5
Burst-description fit
3/5
Burst-timing fit
3/5
Event precision rating
3/5
Event coverage rating
4/5
Transcript fidelity (diagnostic)
5/5
Speech-timing fit (diagnostic)
3/5
Readability (diagnostic)
4/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
59 words preserved; 13 selected events serialized; printed elapsed budget 33.340 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
German10 selected intervals
The Wedding Song at the Funeral
de-01381 · 32.45 s · Voice: Zubenelgenubi · eric_morris_inspired
2.90 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A young adult masculine-coded speaker delivers an intimate, deeply contemplative monologue in Standard German. The vocal delivery is exceptionally soft, warm, and close-mic'd, characterized by a low baritone pitch center, rich chest resonance, and pervasive breathiness bordering on an ASMR or bedroom whisper style. The tempo is slow and unhurried, segmented into short, reflective breath groups with gentle, soft attacks and delicate articulation. The emotional tone is saturated with profound contentment, peaceful awe, and tender affection, evoking a sense of serene joy and emotional safety. Multiple natural paralinguistic events—including contented sighs, soft breathy chuckles, and a gentle intake of air—are seamlessly woven into the pauses between phrases, enhancing the lived-in, vulnerable authenticity of the performance. The recording is pristine, dry, and free of background noise, allowing the subtle micro-textures of the speaker's breath and vocal fry to remain clearly audible.
SCRIPT — timed speech, local cues and free-form burst descriptions
(very organic, very warm, very warmth, enjoyable; 1.92 seconds duration) ("A soft, wistful sighing exhalation opens the recording before speech begins, setting a melancholic, reflective tone.", 0.88 seconds) Schau sie dir an,
[0.32 seconds pause]
(very organic, very warm, very warmth, enjoyable; 1.43 seconds duration) wie glücklich sie tanzen.
[0.65 seconds pause]
(very fading, naturalistic, very yearning, enjoyable; 1.71 seconds duration) Ich will einfach nur für sie lächeln.
[0.45 seconds pause]
(very fading, very meek, very liberation; 1.04 seconds duration) Wirklich. ("A soft, wistful sighing exhalation occurs in the pause after 'wirklich', carrying a gentle drop in pitch and breathy release that reflects tender longing.", 0.88 seconds)
(very whispered, very noisy, very sadness, enjoyable; 1.69 seconds duration) ("A soft, wistful sighing exhalation occurs in the pause before 'Aber dieses Lied', conveying bittersweet longing and emotional vulnerability.", 0.80 seconds) Aber dieses Lied,
[0.79 seconds pause]
(very whispered, very noisy, very sadness, enjoyable; 2.71 seconds duration) es dröhnt so laut und gerade jetzt erfahre ich
[0.33 seconds pause]
(very whispered, very noisy, very sadness, enjoyable; 0.32 seconds duration) das.
[0.58 seconds pause]
(very denasal, very naturalistic, very longing, enjoyable; 3.20 seconds duration) ("A soft, quiet sobbing breath catch and tremulous exhalation occurs in the pause after 'erleidet das', conveying deep emotional pain and vulnerability.", 0.86 seconds) Mein Herz zerreißt förmlich bei jedem Takt.
[0.86 seconds pause]
(very meek, very organic, very awestruck, enjoyable; 1.51 seconds duration) Und trotzdem, ("A soft, deliberate inhalation and gentle release between 'Und trotzdem' and 'es ist so wunderschön', grounding the speaker's emotional reflection.", 0.84 seconds)
(very meek, very organic, very awestruck, enjoyable; 2.35 seconds duration) es ist so wunderschön hier draußen. ("A soft, quiet breathy sigh of wistful longing and contemplation in the pause before 'Man kann das Leben doch nicht einfach anhalten'.", 0.76 seconds)
(very fading, very clean-content, very helplessness, enjoyable; 2.69 seconds duration) ("A quiet, warm chuckle with a gentle breathy release, expressing fond amusement and affectionate resignation.", 0.92 seconds) Man kann das Leben doch nicht einfach anhalten,
(very fading, very clean-content, very helplessness, enjoyable; 1.50 seconds duration) oder? ("A quiet, warm sigh of contentment and release following the rhetorical question 'nicht einfach anhalten, oder?'.", 0.86 seconds) ("A soft, quiet nasal sniffle and breath catch in the pause before the final declaration, conveying lingering emotional vulnerability.", 0.76 seconds)
("A quiet, gentle nasal sniff and breath catch in the pause before the final declaration, conveying emotional vulnerability and suppressed tears.", 0.76 seconds)
[0.71 seconds pause]
(very organic, very whispered, very envy; 0.78 seconds duration) Ich weine
[1.14 seconds pause]
(very organic, very whispered, very envy; 0.92 seconds duration) und ich tanze.
[0.76 seconds pause]
(very accelerating, very steady, insensitivity, enjoyable; 0.65 seconds duration) Für beide.
[0.68 seconds pause]
Automated review note The speaker delivers a melancholic, intimate monologue in German. The annotation over-identifies nonverbal burst events, hallucinating non-existent chuckles and sniffling catches where only sighs and pauses exist, and mislabels grief as 'envy' and 'insensitivity'.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
0.88 s duration · HVA_EVENT_0000
A soft, wistful sighing exhalation opens the recording before speech begins, setting a melancholic, reflective tone.
A soft, wistful sighing exhalation occurs in the pause after 'wirklich', carrying a gentle drop in pitch and breathy release that reflects tender longing.
Local emotion and delivery review — 15 speech segments
Absolute interval
Recognized speech and cue
Gemini ratings
Reviewer explanation
segment-001 Source sentence 1
Schau sie dir an, very organic, very warm, very warmth, enjoyable
Emotion 3/5 Delivery 4/5
Tone is soft and melancholic rather than purely affectionate.
segment-002 Source sentence 1
wie glücklich sie tanzen. very organic, very warm, very warmth, enjoyable
Emotion 3/5 Delivery 4/5
Breathy, wistful delivery fits well.
segment-003 Source sentence 2
Ich will einfach nur für sie lächeln. very fading, naturalistic, very yearning, enjoyable
Emotion 4/5 Delivery 4/5
Yearning and sadness come through accurately.
segment-004 Source sentence 3
Wirklich. very fading, very meek, very liberation
Emotion 2/5 Delivery 3/5
Tagged with liberation/relief, but tone remains sad and heavy.
segment-005 Source sentence 4
Aber dieses Lied, very whispered, very noisy, very sadness, enjoyable
Emotion 4/5 Delivery 4/5
Sadness tag fits the vulnerable delivery.
segment-006 Source sentence 4
es dröhnt so laut und gerade jetzt erfahre ich very whispered, very noisy, very sadness, enjoyable
Emotion 4/5 Delivery 4/5
Breathy, subdued whisper-like tone correctly captured.
segment-007 Source sentence 4
das. very whispered, very noisy, very sadness, enjoyable
Emotion 3/5 Delivery 3/5
Brief completion word continuing the previous segment's sad tone.
segment-008 Source sentence 5
Mein Herz zerreißt förmlich bei jedem Takt. very denasal, very naturalistic, very longing, enjoyable
Emotion 4/5 Delivery 3/5
Conveys heartbreak and sorrow rather than mere longing.
segment-009 Source sentence 6
Und trotzdem, very meek, very organic, very awestruck, enjoyable
Emotion 3/5 Delivery 3/5
Awe is overstated; the delivery is contemplative.
segment-010 Source sentence 6
es ist so wunderschön hier draußen. very meek, very organic, very awestruck, enjoyable
Emotion 3/5 Delivery 4/5
Gentle appreciation and wistfulness present, but awe is exaggerated.
segment-011 Source sentence 7
Man kann das Leben doch nicht einfach anhalten, very fading, very clean-content, very helplessness, enjoyable
Emotion 4/5 Delivery 4/5
Helplessness fits the quiet resigned tone.
segment-012 Source sentence 7
oder? very fading, very clean-content, very helplessness, enjoyable
Emotion 3/5 Delivery 3/5
Short rhetorical tag matching previous segment.
segment-013 Source sentence 8
Ich weine very organic, very whispered, very envy
Emotion 2/5 Delivery 4/5
Tagged with envy, but the emotion is straightforward grief and sorrow.
segment-014 Source sentence 8
und ich tanze. very organic, very whispered, very envy
Emotion 2/5 Delivery 4/5
Also tagged with envy; speech sounds mournful and reflective.
segment-015 Source sentence 9
Für beide. very accelerating, very steady, insensitivity, enjoyable
Emotion 2/5 Delivery 3/5
Tagged as insensitivity/numbness, but delivery remains tender and emotional.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
3/5
Whole-recording emotion fit
3/5
Local speech emotion fit
3/5
Whole-recording talking-style fit
4/5
Local speech delivery fit
3/5
Burst-description fit
2/5
Burst-timing fit
2/5
Event precision rating
2/5
Event coverage rating
3/5
Transcript fidelity (diagnostic)
5/5
Speech-timing fit (diagnostic)
4/5
Readability (diagnostic)
4/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
60 words preserved; 10 selected events serialized; printed elapsed budget 32.450 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
German12 selected intervals
The Last Light in Your Eyes
de-01938 · 30.10 s · Voice: Rasalgethi · eric_morris_inspired
3.45 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A young adult masculine-coded speaker delivers an intimate, deeply emotional monologue in Standard German. The performance is marked by a very slow tempo, fragmented phrasing, and heavy breathiness, bordering on a stage whisper or ASMR-like close-mic proximity. The vocal timbre is warm, soft, and slightly raspy, with low harmonicity and prominent chest and pharyngeal resonance. The emotional tone is saturated with profound affection, tenderness, and quiet relief, tinged with lingering vulnerability and fatigue. Multiple audible breaths, a soft sigh of relief, and a gentle breathy chuckle are seamlessly integrated into the pauses between phrases, enhancing the lived-in, highly genuine quality of the delivery. The recording is pristine, dry, and close-mic'd with no background noise.
SCRIPT — timed speech, local cues and free-form burst descriptions
(very clean-bg, non-narrative, slightly burnout; 1.08 seconds duration) Bleib ("A soft, tremulous sharp intake of breath precedes the opening words, establishing a fragile, intimate, and emotionally strained state.", 0.54 seconds) noch kurz.
(very darkening, hi-fi, irritability; 3.43 seconds duration) ("A sharp, audible intake of air following 'Bleib noch kurz', marked by slight pharyngeal constriction and breathiness, signaling physical strain and emotional distress.", 0.60 seconds) ("A fragile, breathy vocalized whimper and catch in the throat occurs during a pause after 'Bleib noch kurz', conveying deep emotional vulnerability and sorrow.", 0.70 seconds) Ich kann mir deine Hände gar nicht schnell genug merken.
[0.37 seconds pause]
(very hi-fi, accelerating, slightly curiosity, enjoyable; 1.66 seconds duration) Diese feinen Linien, das Licht.
(very breathy, very whispered, very desperation; 0.90 seconds duration) ("A soft, tremulous breath catch and quiet sob-like whimper occurs in the pause before 'Warte nicht vergessen', conveying deep sorrow and emotional fragility.", 0.74 seconds) Warte
[0.24 seconds pause]
(very breathy, very whispered, very desperation; 1.14 seconds duration) nicht weg hin. ("A soft, tremulous breathy sigh and quiet sob-like release following 'Warte nicht vergessen', conveying deep sorrow and helplessness.", 0.74 seconds)
[0.70 seconds pause]
(very scattered, breathy, slightly boredom; 1.36 seconds duration) Aber schau dich an.
[0.24 seconds pause]
(very whispered, very darkening, slightly indifference; 1.16 seconds duration) Du bist so ruhig.
(overlapping non-speech vocal events; 0.94 seconds duration) ("A soft, audible intake of air following 'Du bist so ruhig', marking a quiet pause filled with sorrowful contemplation.", 0.64 seconds) ("A soft, audible intake of air following 'Du bist so ruhig', functioning as a quiet, intimate breath reset.", 0.58 seconds)
[0.22 seconds pause]
(very whispered, very darkening, very helplessness, enjoyable; 1.44 seconds duration) Kein Kampf mehr.
[0.40 seconds pause]
(very fading, very whispered, very liberation; 1.86 seconds duration) Nur noch dieser sanfte Atemzug.
[0.38 seconds pause]
("A soft, quiet exhale of relief and sorrowful release following 'sanfte Atemzug', blending naturally into the hushed tone of the subsequent sentence.", 0.82 seconds)
(very clean-content, hi-fi, slightly surprise, enjoyable; 1.64 seconds duration) Deine Finger sind noch ganz warm.
[1.14 seconds pause]
(very darkening, very negative, very agony; 1.02 seconds duration) Es tut so weh.
("A soft, quiet sobbing breath catch and tremulous release following 'Es tut so weh', conveying deep sorrow and helplessness.", 0.78 seconds)
[0.36 seconds pause]
(very accelerating, very fading, slightly insensitivity; 0.56 seconds duration) Trotzdem.
[0.58 seconds pause]
(overlapping non-speech vocal events; 1.18 seconds duration) ("A soft, quiet exhale of relief and sorrowful acceptance bridges the pause before 'Ich darf hier bei dir sein', blending seamlessly into the hushed, intimate tone.", 0.76 seconds) ("A soft, deliberate intake of air right before 'Ich darf hier bei dir sein', grounding the final tender declaration.", 0.78 seconds)
(very hi-fi, clean-bg, courage, enjoyable; 1.84 seconds duration) Ich darf hier bei dir sein.
[0.32 seconds pause]
(very cold, very flat-resonance, detachment, less-enjoyable; 0.42 seconds duration) Danke.
(very positive, very fading, very thankfulness, enjoyable; 1.61 seconds duration) ("A soft, quiet intake of air between 'Danke' and 'Danke für alles', maintaining the hushed, intimate tone.", 0.54 seconds) Danke für alles.
[0.31 seconds pause]
Automated review note The global description captures the intimate, whispered, ASMR-like German performance nicely. However, several segment-level emotion tags (e.g., irritability, indifference, coldness/detachment) are distinctly inaccurate given the persistently tender, mournful, and affectionate context. In addition, duplicate overlapping breath bursts were artificially created in multiple pauses.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
0.54 s duration · HVA_EVENT_0000
A soft, tremulous sharp intake of breath precedes the opening words, establishing a fragile, intimate, and emotionally strained state.
A sharp, audible intake of air following 'Bleib noch kurz', marked by slight pharyngeal constriction and breathiness, signaling physical strain and emotional distress.
A fragile, breathy vocalized whimper and catch in the throat occurs during a pause after 'Bleib noch kurz', conveying deep emotional vulnerability and sorrow.
A soft, tremulous breath catch and quiet sob-like whimper occurs in the pause before 'Warte nicht vergessen', conveying deep sorrow and emotional fragility.
A soft, quiet exhale of relief and sorrowful acceptance bridges the pause before 'Ich darf hier bei dir sein', blending seamlessly into the hushed, intimate tone.
Nur noch dieser sanfte Atemzug. very fading, very whispered, very liberation
Emotion 4/5 Delivery 4/5
Gentle, whispered fading delivery matches.
segment-010 Source sentence 9
Deine Finger sind noch ganz warm. very clean-content, hi-fi, slightly surprise, enjoyable
Emotion 3/5 Delivery 4/5
Quiet awe/tenderness rather than standard surprise.
segment-011 Source sentence 10
Es tut so weh. very darkening, very negative, very agony
Emotion 4/5 Delivery 4/5
Pain and distress are audible here.
segment-012 Source sentence 11
Trotzdem. very accelerating, very fading, slightly insensitivity
Emotion 2/5 Delivery 3/5
A quiet sigh/word, definitely not insensitive.
segment-013 Source sentence 12
Ich darf hier bei dir sein. very hi-fi, clean-bg, courage, enjoyable
Emotion 4/5 Delivery 4/5
Intimate, tender emotional grounding.
segment-014 Source sentence 13
Danke. very cold, very flat-resonance, detachment, less-enjoyable
Emotion 2/5 Delivery 2/5
Tagged as cold and detached, but the voice remains warm, broken, and intimate.
segment-015 Source sentence 14
Danke für alles. very positive, very fading, very thankfulness, enjoyable
Emotion 4/5 Delivery 4/5
Soft, grateful, whispered sign-off.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
4/5
Whole-recording emotion fit
4/5
Local speech emotion fit
3/5
Whole-recording talking-style fit
4/5
Local speech delivery fit
3/5
Burst-description fit
3/5
Burst-timing fit
3/5
Event precision rating
3/5
Event coverage rating
4/5
Transcript fidelity (diagnostic)
4/5
Speech-timing fit (diagnostic)
4/5
Readability (diagnostic)
4/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
59 words preserved; 12 selected events serialized; printed elapsed budget 30.100 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.
German12 selected intervals
The Taste of Copper and Victory
de-02578 · 36.17 s · Voice: Gacrux · eric_morris_inspired
2.85 / 5Automated composite
Original audio bytes, not regenerated. Select a colored interval or a timestamp to play from that point.
A young adult feminine-coded speaker delivers an intensely intimate, sensual monologue in Standard German, describing the burning sensation of a hot beverage. The performance is delivered in a close-mic, ASMR-like whisper-talk style with heavy breathiness, low harmonicity, and prominent oral and pharyngeal resonance. The tempo is slow and fragmented into short, breathless phrases separated by frequent, audible inhalations and soft sighs. The emotional tone is dominated by sensual pleasure, longing, and physical warmth, with a playful, teasing undertone. The recording is pristine with a dead-quiet background, allowing every subtle breath and lip movement to be clearly heard.
SCRIPT — timed speech, local cues and free-form burst descriptions
(very accelerating, very hi-fi, enjoyable; 0.98 seconds duration) Da liegst
(very accelerating, very hi-fi, enjoyable; 0.74 seconds duration) du. ("A sharp, audible intake of air through the mouth and throat occurs right after 'Da liegst du', conveying physical strain and breathlessness.", 0.56 seconds)
(very cold, very thin-chest; 1.20 seconds duration) ("A sharp, audible intake of air through the mouth and throat occurs in the pause between 'Da liegst du' and 'Siehst du das?', conveying physical strain and heightened tension.", 0.56 seconds) Siehst du das?
[0.30 seconds pause]
("A sharp, raspy intake of breath through constricted vocal cords, occurring in the pause between 'Siehst du das?' and 'Ich hab's geschafft', conveying physical strain and breathlessness.", 0.74 seconds)
(very clean-bg, very choppy, very triumph; 1.48 seconds duration) Ich hab's geschafft.
(very breathy, whispered, burnout; 1.44 seconds duration) ("A sharp, raspy intake of breath through constricted vocal cords occurs in the pause after 'Ich hab's geschafft', conveying physical strain and breathlessness.", 0.76 seconds) Ganz genau so.
(overlapping non-speech vocal events; 1.24 seconds duration) ("A sharp, raspy intake of breath through constricted vocal cords, signaling physical strain and breathlessness.", 0.88 seconds) ("A sharp, raspy intake of air following 'Ganz genau so', conveying physical strain and breathlessness.", 0.72 seconds)
[0.52 seconds pause]
(very breathy, very tense, very peevishness; 0.06 seconds duration) Du
[0.26 seconds pause]
(very breathy, very tense, very peevishness; 2.28 seconds duration) bist am Boden, verdammt. ("A sharp, raspy intake of breath following 'verdammt', capturing physical strain and throat constriction.", 0.92 seconds)
[0.60 seconds pause]
(very whispered, very fading, very fatigue; 1.25 seconds duration) Mein Mund brennt.
[0.59 seconds pause]
(very whispered, very asmr, very carnal desire, enjoyable; 2.82 seconds duration) Dieses warme, metallische Zeug auf der Zunge.
[0.62 seconds pause]
(very tense, very cold, slightly emotional blunting; 0.52 seconds duration) Kupfer!
[0.92 seconds pause]
(very hi-fi, very clean-bg, very bliss; 1.57 seconds duration) So schmeckt also das Gewinn.
("A quick, breathy intake of air following 'das Gewinn', maintaining the close-mic, strained respiratory state.", 0.68 seconds)
(very whispered, very asmr, slightly romantic desire; 1.90 seconds duration) Es klebt überall an den Zähnen.
[0.73 seconds pause]
(very dissociated, very flat, boredom; 0.51 seconds duration) Schmeckt.
[1.41 seconds pause]
(very asmr, very whispered, devotion; 2.25 seconds duration) Schmeckt genau wie damals. ("A slow, raspy inhalation and exhalation cycle in the pause before 'Gott, ich bin genauso geworden', emphasizing the speaker's physical exhaustion and bodily transformation.", 0.84 seconds)
[0.63 seconds pause]
(very breathy, very whispered, burnout; 2.32 seconds duration) Gott, ich bin genauso geworden wie du.
[0.96 seconds pause]
(very whispered, very naturalistic, very detachment; 1.43 seconds duration) Da ist überhaupt kein Sieg,
[0.25 seconds pause]
(very whispered, very naturalistic, very detachment; 1.33 seconds duration) nur. ("A slow, raspy inhalation and exhalation cycle in the pause before 'Dieser bittere Geschmack', emphasizing the speaker's physical exhaustion and lingering sensory memory.", 0.80 seconds) ("A low, strained, breathy groan of physical suffering and exhaustion right before the final whispered phrase.", 0.76 seconds)
(very whispered, very noisy, slightly emotional blunting; 1.64 seconds duration) Dieser bittere Geschmack. ("A faint, breathy trailing exhalation at the very end of the clip, releasing residual tension.", 0.38 seconds)
Automated review note The speech is delivered in a sustained close-up ASMR whisper. The prompt invents a story about a hot beverage and erotic desire ('carnal desire', 'romantic desire', 'bliss'), when the tone and text clearly point to an injured/bitter reflection (blood in the mouth tasting like copper/defeat). Multiple burst event annotations duplicate overlapping intervals.
Selected vocal events: exact starts/ends and top-three types
These are detector-selected intervals. MOSS prose describes the full-audio interval; softmax percentages indicate type ranking and are not calibrated correctness probabilities.
0.56 s duration · HVA_EVENT_0000
A sharp, audible intake of air through the mouth and throat occurs right after 'Da liegst du', conveying physical strain and breathlessness.
A sharp, audible intake of air through the mouth and throat occurs in the pause between 'Da liegst du' and 'Siehst du das?', conveying physical strain and heightened tension.
A sharp, raspy intake of breath through constricted vocal cords, occurring in the pause between 'Siehst du das?' and 'Ich hab's geschafft', conveying physical strain and breathlessness.
A sharp, raspy intake of breath through constricted vocal cords occurs in the pause after 'Ich hab's geschafft', conveying physical strain and breathlessness.
A slow, raspy inhalation and exhalation cycle in the pause before 'Gott, ich bin genauso geworden', emphasizing the speaker's physical exhaustion and bodily transformation.
A slow, raspy inhalation and exhalation cycle in the pause before 'Dieser bittere Geschmack', emphasizing the speaker's physical exhaustion and lingering sensory memory.
Da ist überhaupt kein Sieg, very whispered, very naturalistic, very detachment
Emotion 3/5 Delivery 4/5
Detached whisper fits the utterance.
segment-017 Source sentence 14
nur. very whispered, very naturalistic, very detachment
Emotion 3/5 Delivery 3/5
Short trailing whisper.
segment-018 Source sentence 15
Dieser bittere Geschmack. very whispered, very noisy, slightly emotional blunting
Emotion 3/5 Delivery 3/5
Final fading whisper.
All recording-level review ratings
Criterion
Rating
Whole-recording caption fidelity
3/5
Whole-recording emotion fit
2/5
Local speech emotion fit
2/5
Whole-recording talking-style fit
4/5
Local speech delivery fit
3/5
Burst-description fit
3/5
Burst-timing fit
3/5
Event precision rating
3/5
Event coverage rating
3/5
Transcript fidelity (diagnostic)
4/5
Speech-timing fit (diagnostic)
3/5
Readability (diagnostic)
4/5
All 192 raw scalar predictions, reference z-scores and auxiliary targets
Raw values retain each target's own units. Reference z-scores use the speech normalizer, not the training-target normalizer. Negative values are preserved. Full embeddings, frames, local raw scores and all proposals are in the downloadable JSON.
60 words preserved; 12 selected events serialized; printed elapsed budget 36.170 s. Independent format proof: passed. This does not establish acoustic correctness.
Voice and source metadata are preserved. Full raw predictions and embeddings are in canonical annotations; reviewer response and token usage are in all reviews.