Procedural voice captions with detected vocal bursts
⚠️ Older demo — rendered at locator threshold 0.7 with the earlier Variant-A/B framing. The current
default is locator@0.75 + duration & no-burst gates + terse tags, with the LLM path rewording the draft
while keeping burst positions. See the up-to-date grid →
character-captions.
This page is the previous pipeline.
It was produced with locator model.pt (v1) and the multi-label burst classifier at the v1
post-processing defaults (threshold 0.75, merge_gap 0.30 s, min_duration 0.30 s). The current defaults are
model_v2.pt + laion/vocal-burst-detector-v2 at 0.50 / 0.10 / 0.10 —
see the current output here. This page is kept for comparison.
What this page adds. The base procedural captioner turns
model predictions into a voice description. Here we additionally detect the non-verbal vocal bursts
(laughs, gasps, sighs, screams, sobs, grunts, slaps…) that the audio really contains and write them into
the captions. Every number below is produced by models listening to the clip — nothing is hand-written.
Scoring stack (all run on the audio): VoiceNet 57-dim predictors + genuineness + vocal-burst-blend
(VoiceCLAP-commercial embedding → per-dim MLP heads); EmoNet-40 (on (Empathic-Insight-Voice-Plus));
Parakeet-TDT for word + sentence timestamps; the vocalburst-locator (50 fps burst probability)
and a VoiceCLAP→MLP multi-label burst classifier (82 taxonomy classes + no_burst).
Caption composition.
• Global caption = Top-5 VoiceNet dims + Top-3 EmoNet emotions + genuineness + Age + Gender + Tempo.
Bursts are not placed at global level.
• Per-sentence caption = Top-3 VoiceNet dims + Top-3 emotions (no Age/Gender/Tempo — those are global only).
Two burst-insertion variants (compare them per clip below): Variant A — locator (precise position). Scan the whole clip with the burst locator at
threshold 0.7; each detected time-span's audio → classifier → if P(no_burst) < 0.5 take the
top-1 class and insert it as its own (Class) inline at that moment, between the
two ASR words nearest the burst time. Otherwise the span is discarded (no false alarm). Variant B — sentence-level. Run the classifier on each whole sentence segment; if
P(no_burst) ≥ 0.5 attach nothing, else weave the top-1 class into that sentence's caption with a
small procedurally-generated phrase (“… punctuated by a (Gasp).”).
Recommendation: use Variant A (locator) as the primary, Variant B as a fallback.
Variant A places the burst where it actually happens — the inline transcript reads like a real script
(“(Surprised Gasp) No, no, (Surprised Gasp) oh, it hurts.”), which is exactly the per-event, time-anchored
signal the voice-acting format wants, and its explicit P(no_burst)<0.5 gate suppresses false alarms
per-span. Variant B is coarser: it can only say a sentence contains some burst, it collapses multiple bursts
in one sentence into a single label, and running the classifier over a whole sentence dilutes a short burst among
seconds of speech (lower recall on brief events). Prefer Variant B only when word-level timestamps are missing
(non-speech screams, noisy ASR) or when a compact one-label-per-sentence summary is all that's needed — there,
Variant A has nowhere to anchor and Variant B still adds useful information.
ASR: No, no, oh, it hurts. Ah, Hang, get it off me.
Global caption
A voice that is extremely raw and vulnerable; extremely unpleasant and harsh; extremely effortful and strained; extremely jerky and uneven in rhythm; extremely choppy and fragmented; childlike; androgynous, gender-ambiguous in pitch; low and bassy in register; somewhat slow and deliberate in tempo; extremely carrying physical pain; extremely carrying submission; extremely carrying alleviation; genuine in delivery.
Per-sentence captions + Variant B (sentence-level bursts)
“No, no, oh, it hurts.”
Delivered extremely raw and vulnerable; extremely unpleasant and harsh; extremely audibly breathing and gasping; extremely carrying agony; extremely carrying powerlessness; extremely carrying worry.
Variant B: Delivered extremely raw and vulnerable; extremely unpleasant and harsh; extremely audibly breathing and gasping; extremely carrying agony; extremely carrying powerlessness; extremely carrying worry, carrying a (Surprised Gasp). (Surprised Gasp)
“Ah, Hang, get it off me.”
Delivered extremely audibly breathing and gasping; extremely raw and vulnerable; extremely unpleasant and harsh; extremely carrying alleviation; extremely carrying exhaustion; notably free of attention.
Variant B: Delivered extremely audibly breathing and gasping; extremely raw and vulnerable; extremely unpleasant and harsh; extremely carrying alleviation; extremely carrying exhaustion; notably free of attention, marked by a (Surprised Gasp). (Surprised Gasp)
Variant A — locator (precise inline bursts) 10 span(s) @0.7
(Surprised Gasp) No, no, (Surprised Gasp) oh, it hurts. (Surprised Gasp) Ah, (Slap Face) Hang, (Slap Face) get it off me.
A voice that is extremely one-dimensional in resonance; extremely unpleasant and harsh; extremely dialogic and outward; extremely thin in chest resonance; childlike; feminine and high-pitched; high in register; moderate in tempo; extremely carrying helplessness; extremely carrying terror; extremely carrying anxiety; genuine in delivery; interwoven with vocal bursts (laughs, gasps, sighs).
Per-sentence captions + Variant B (sentence-level bursts)
“No, no, Erk, somebody help me”
Delivered high in register; extremely one-dimensional in resonance; extremely dialogic and outward; extremely carrying desperation; extremely carrying anxiety; extremely carrying dread.
Variant B: Delivered high in register; extremely one-dimensional in resonance; extremely dialogic and outward; extremely carrying desperation; extremely carrying anxiety; extremely carrying dread, carrying a (Trembling Whimper). (Trembling Whimper)
Variant A — locator (precise inline bursts) 0 span(s) @0.7
No, no, Erk, somebody help me
no span ≥ 0.7
characterzombie512.4s · genu 2.22 · blend 3.6
ASR: So hungry. Hmm. Help me. I can't stop.
Global caption
A voice that is extremely raw and vulnerable; extremely unpleasant and harsh; extremely meek and submissive; extremely flat and monotone; extremely dissociated and inward; adult; clearly feminine and high-pitched; low-to-mid in register; very slow and deliberate in tempo; extremely carrying burnout; notably free of engrossment; extremely carrying sexual lust; genuine in delivery; interwoven with vocal bursts (laughs, gasps, sighs).
Per-sentence captions + Variant B (sentence-level bursts)
“So hungry.”
Delivered extremely collapsing in energy; extremely blurry and mumbled; extremely meek and submissive; extremely carrying burnout; extremely carrying suffering; extremely carrying desperation.
Variant B: Delivered extremely collapsing in energy; extremely blurry and mumbled; extremely meek and submissive; extremely carrying burnout; extremely carrying suffering; extremely carrying desperation, interrupted by a (Contented Sigh). (Contented Sigh)
“Hmm.”
Delivered extremely blurry and mumbled; extremely flat and unemphasized; extremely collapsing in energy; extremely carrying fatigue; extremely carrying alleviation; notably free of concentration.
Variant B: Delivered extremely blurry and mumbled; extremely flat and unemphasized; extremely collapsing in energy; extremely carrying fatigue; extremely carrying alleviation; notably free of concentration, with an audible (Wistful Sigh). (Wistful Sigh)
“Help me.”
Delivered extremely flat and unemphasized; extremely blurry and mumbled; extremely collapsing in energy; extremely carrying submission; extremely carrying sorrow; extremely carrying ache.
Variant B: Delivered extremely flat and unemphasized; extremely blurry and mumbled; extremely collapsing in energy; extremely carrying submission; extremely carrying sorrow; extremely carrying ache, marked by a (Surprised Gasp). (Surprised Gasp)
“I can't stop.”
Delivered extremely flat and unemphasized; extremely collapsing in energy; extremely flat and monotone; extremely carrying exhaustion; extremely carrying submission; extremely carrying melancholy.
Variant B: Delivered extremely flat and unemphasized; extremely collapsing in energy; extremely flat and monotone; extremely carrying exhaustion; extremely carrying submission; extremely carrying melancholy. — no burst
Variant A — locator (precise inline bursts) 11 span(s) @0.7
So hungry. Hmm. (Relief Sigh) Help (Lip Smack) me. (Relief Sigh) I can't (Slap Face)(Lip Smack) stop. (Chuckle)
ASR: You should not have come here, little one. This is where your journey ends.
Global caption
A voice that is extremely full and cinematic; extremely pristine in recording quality; very smooth and clean; very one-dimensional in resonance; very warm and enveloping; adult; strongly masculine, deep and low-pitched; low and bassy in register; somewhat slow and deliberate in tempo; notably free of attention; notably carrying pride; notably carrying desire to harm; only slightly genuine, somewhat performed.
Per-sentence captions + Variant B (sentence-level bursts)
“You should not have come here, little one.”
Delivered extremely seamless and inaudible in breathing; extremely smooth and clean; extremely pristine in recording quality; notably free of deep focus; notably carrying ribbing; notably carrying reverence.
Variant B: Delivered extremely seamless and inaudible in breathing; extremely smooth and clean; extremely pristine in recording quality; notably free of deep focus; notably carrying ribbing; notably carrying reverence. — no burst
“This is where your journey ends.”
Delivered extremely seamless and inaudible in breathing; extremely pristine in recording quality; very crisply articulated.
Variant B: Delivered extremely seamless and inaudible in breathing; extremely pristine in recording quality; very crisply articulated. — no burst
Variant A — locator (precise inline bursts) 0 span(s) @0.7
You should not have come here, little one. This is where your journey ends.
no span ≥ 0.7
characterranting511.9s · genu 1.5 · blend 9.54
ASR: Enough. I am sick to death of this. You pathetic, spineless fool. How dare you how dare you waste my time again and again?
Global caption
A voice that is extremely raw and vulnerable; extremely noisy and aperiodic; extremely denasal and clear; extremely distressed and negative; extremely darkening in mood; adult; strongly masculine, deep and low-pitched; low and bassy in register; moderate in tempo; notably carrying jealousy; notably carrying infatuation; notably carrying ache; only slightly genuine, somewhat performed; interwoven with vocal bursts (laughs, gasps, sighs).
Per-sentence captions + Variant B (sentence-level bursts)
“Enough.”
Delivered extremely jerky and uneven in rhythm; extremely flat and unemphasized; extremely collapsing in energy; notably free of absorption; extremely carrying butterflies in the stomach; notably free of intrigue.
Variant B: Delivered extremely jerky and uneven in rhythm; extremely flat and unemphasized; extremely collapsing in energy; notably free of absorption; extremely carrying butterflies in the stomach; notably free of intrigue. — no burst
“I am sick to death of this.”
Delivered extremely audibly breathing and gasping; extremely rough and raspy; extremely denasal and clear; extremely carrying adoration; extremely carrying ache; extremely carrying submission.
Variant B: Delivered extremely audibly breathing and gasping; extremely rough and raspy; extremely denasal and clear; extremely carrying adoration; extremely carrying ache; extremely carrying submission, broken by a (Snort). (Snort)
“You pathetic, spineless fool.”
Delivered extremely denasal and clear; extremely rough and raspy; extremely distressed and negative; extremely carrying repulsion; extremely carrying scorn; extremely carrying schadenfreude.
Variant B: Delivered extremely denasal and clear; extremely rough and raspy; extremely distressed and negative; extremely carrying repulsion; extremely carrying scorn; extremely carrying schadenfreude, carrying a (Spitting). (Spitting)
“How dare you how dare you waste my time again and again?”
Delivered extremely denasal and clear; extremely rough and raspy; extremely intensely focused; notably carrying butterflies in the stomach; notably carrying fury; notably carrying feeling horny.
Variant B: Delivered extremely denasal and clear; extremely rough and raspy; extremely intensely focused; notably carrying butterflies in the stomach; notably carrying fury; notably carrying feeling horny, punctuated by a (Exasperated Sigh). (Exasperated Sigh)
Variant A — locator (precise inline bursts) 3 span(s) @0.7
Enough. (Breathy Giggle) I am sick to death of this. You pathetic, (Slap Face) spineless fool. How dare you how dare you waste my time again and again?
ASR: I am absolutely sick of this! How dare you waste my time again and again!
Global caption
A voice that is extremely strongly emphasized; extremely wide in dynamic swell; extremely collapsing in energy; extremely tight and tense; very dramatic and theatrical; young and youthful; masculine and deep-pitched; low-to-mid in register; somewhat fast and rapid in tempo; notably carrying wrath; notably carrying restlessness; notably carrying letdown; only slightly genuine, somewhat performed.
Per-sentence captions + Variant B (sentence-level bursts)
“I am absolutely sick of this!”
Delivered extremely accelerating in pace; extremely tight and tense; extremely bright in oral resonance; notably carrying abhorrence; notably free of deep focus; notably carrying pleasure.
Variant B: Delivered extremely accelerating in pace; extremely tight and tense; extremely bright in oral resonance; notably carrying abhorrence; notably free of deep focus; notably carrying pleasure. — no burst
“How dare you waste my time again and again!”
Delivered extremely wide in dynamic swell; extremely tight and tense; extremely accelerating in pace; notably carrying enragement; notably carrying exasperation; notably carrying disappointment.
Variant B: Delivered extremely wide in dynamic swell; extremely tight and tense; extremely accelerating in pace; notably carrying enragement; notably carrying exasperation; notably carrying disappointment. — no burst
Variant A — locator (precise inline bursts) 4 span(s) @0.7
I am absolutely sick of this! How dare you waste my time again and again! (Hiccups)
ASR: You should not have come here. This house feeds on the living.
Global caption
A voice that is extremely rough and raspy; extremely full and cinematic; elderly-sounding; extremely warm and enveloping; extremely distressed and negative; strongly masculine, deep and low-pitched; low and bassy in register; somewhat slow and deliberate in tempo; notably carrying malevolence; notably carrying nervousness; notably carrying disgust; only slightly genuine, somewhat performed; interwoven with vocal bursts (laughs, gasps, sighs).
Per-sentence captions + Variant B (sentence-level bursts)
“You should not have come here.”
Delivered extremely full and cinematic; extremely rough and raspy; extremely intensely focused; notably free of attention; notably carrying schadenfreude; notably carrying rage.
Variant B: Delivered extremely full and cinematic; extremely rough and raspy; extremely intensely focused; notably free of attention; notably carrying schadenfreude; notably carrying rage. — no burst
“This house feeds on the living.”
Delivered extremely full and cinematic; extremely rough and raspy; very thin in head resonance; notably carrying feeling turned on; notably carrying longing; notably carrying dejection.
Variant B: Delivered extremely full and cinematic; extremely rough and raspy; very thin in head resonance; notably carrying feeling turned on; notably carrying longing; notably carrying dejection, broken by a (Heavy Breathing). (Heavy Breathing)
Variant A — locator (precise inline bursts) 1 span(s) @0.7
(Exasperated Sigh) You should not have come here. This house feeds on the living.
0.2–1.1s → Exasperated Sigh (0.80)
charactergoblin58.3s · genu 1.74 · blend 6.55
ASR: Huh. What have we here? A lost little traveller with shiny pockets?
Global caption
A voice that is extremely warm and enveloping; extremely decelerating in pace; extremely flat and monotone; extremely soft and gentle in onset; extremely meek and submissive; middle-aged to mature; strongly masculine, deep and low-pitched; low and bassy in register; very slow and deliberate in tempo; extremely carrying longing; notably free of concentration; extremely carrying helplessness; genuine in delivery; interwoven with vocal bursts (laughs, gasps, sighs).
Per-sentence captions + Variant B (sentence-level bursts)
“Huh.”
Delivered extremely flat and unemphasized; extremely collapsing in energy; extremely noisy in the background; notably free of engrossment; extremely carrying weariness; notably free of curiosity.
Variant B: Delivered extremely flat and unemphasized; extremely collapsing in energy; extremely noisy in the background; notably free of engrossment; extremely carrying weariness; notably free of curiosity. — no burst
“What have we here?”
Delivered extremely flat and unemphasized; extremely flat and monotone; extremely blurry and mumbled; extremely carrying powerlessness; extremely carrying melancholy; notably free of concentration.
Variant B: Delivered extremely flat and unemphasized; extremely flat and monotone; extremely blurry and mumbled; extremely carrying powerlessness; extremely carrying melancholy; notably free of concentration. — no burst
“A lost little traveller with shiny pockets?”
Delivered extremely warm and enveloping; extremely soft and gentle in onset; extremely collapsing in energy; notably carrying fondness; notably carrying feeling horny; notably carrying altered perception.
Variant B: Delivered extremely warm and enveloping; extremely soft and gentle in onset; extremely collapsing in energy; notably carrying fondness; notably carrying feeling horny; notably carrying altered perception. — no burst
Variant A — locator (precise inline bursts) 7 span(s) @0.7
(Clears Throat) Huh. (Deep Breath) What have (Purr) we here? (Slap Face) A lost little traveller with (Lip Smack) shiny pockets?
ASR: He what have we here? A lost little traveller with shiny pockets?
Global caption
A voice that is extremely heavily nasal in resonance; extremely bright in oral resonance; extremely narrator-like and storybook; very crisply articulated; very effortless and fluent; adult; feminine and high-pitched; low-to-mid in register; moderate in tempo; notably carrying provoking lightly; notably free of engrossment; notably carrying playfulness; only slightly genuine, somewhat performed.
Per-sentence captions + Variant B (sentence-level bursts)
“He what have we here?”
Delivered very accelerating in pace; very clean and quiet in background; very choppy and fragmented; extremely carrying disorientation; extremely carrying silliness; notably free of engrossment.
Variant B: Delivered very accelerating in pace; very clean and quiet in background; very choppy and fragmented; extremely carrying disorientation; extremely carrying silliness; notably free of engrossment. — no burst
“A lost little traveller with shiny pockets?”
Delivered very bright in oral resonance; very effortless and fluent; very narrator-like and storybook; notably carrying covetousness; notably free of attention; notably carrying envy.
Variant B: Delivered very bright in oral resonance; very effortless and fluent; very narrator-like and storybook; notably carrying covetousness; notably free of attention; notably carrying envy. — no burst
Variant A — locator (precise inline bursts) 0 span(s) @0.7
He what have we here? A lost little traveller with shiny pockets?
no span ≥ 0.7
charactermouse511.8s · genu 0.97 · blend 2.89
ASR: Eee wait wait ooh did you see it? Did you see how teeny tiny it was?
Global caption
A voice that is extremely rough and raspy; extremely unpleasant and harsh; extremely tight and tense; extremely choppy and fragmented; extremely dominant and commanding; young and youthful; strongly masculine, deep and low-pitched; low and bassy in register; somewhat slow and deliberate in tempo; notably carrying wonder; notably carrying joy; notably free of absorption; only slightly genuine, somewhat performed.
Per-sentence captions + Variant B (sentence-level bursts)
Variant B: Delivered extremely rough and raspy; extremely unpleasant and harsh; extremely emotionally volatile; notably carrying awestruck; notably carrying amazement; notably carrying happiness, punctuated by a (Hiss). (Hiss)
“Did you see how teeny tiny it was?”
Delivered extremely dominant and commanding; very wide in dynamic swell; very dialogic and outward; notably carrying wonder; notably free of attention; notably carrying startlement.
Variant B: Delivered extremely dominant and commanding; very wide in dynamic swell; very dialogic and outward; notably carrying wonder; notably free of attention; notably carrying startlement. — no burst
Variant A — locator (precise inline bursts) 10 span(s) @0.7
Eee wait wait ooh did you see it? Did you see how teeny tiny it was? (Slap Face)(Lip Smack)(Lip Smack)(Lip Smack)(Lip Smack)(Slap Face)
ASR: You dare enter my lair, mortal. Your bones will join the others.
Global caption
A voice that is extremely rough and raspy; extremely full and cinematic; extremely deep in throat resonance; extremely thin in mask resonance; very denasal and clear; middle-aged to mature; masculine and deep-pitched; low and bassy in register; somewhat slow and deliberate in tempo; notably carrying fondness; notably carrying lust; notably carrying spite; only slightly genuine, somewhat performed; interwoven with vocal bursts (laughs, gasps, sighs).
Per-sentence captions + Variant B (sentence-level bursts)
“You dare enter my lair, mortal.”
Delivered extremely rough and raspy; extremely full and cinematic; extremely thin in mask resonance; notably carrying butterflies in the stomach; notably carrying sexual lust; notably carrying malice.
Variant B: Delivered extremely rough and raspy; extremely full and cinematic; extremely thin in mask resonance; notably carrying butterflies in the stomach; notably carrying sexual lust; notably carrying malice, interrupted by a (Wolf Whistle). (Wolf Whistle)
“Your bones will join the others.”
Delivered extremely rough and raspy; extremely one-dimensional in resonance; extremely thin in mask resonance; notably carrying sadism; notably carrying adoration; notably carrying carnal desire.
Variant B: Delivered extremely rough and raspy; extremely one-dimensional in resonance; extremely thin in mask resonance; notably carrying sadism; notably carrying adoration; notably carrying carnal desire, with an audible (Displeased Grunt). (Displeased Grunt)
Variant A — locator (precise inline bursts) 6 span(s) @0.7
You dare enter my lair, mortal. (Slap Face) Your bones will (Displeased Grunt) join the others. (Wolf Whistle)
ASR: Oh hello there little one what a wonderful day to make a new friend.
Global caption
A voice that is high in register; extremely one-dimensional in resonance; very thin in chest resonance; very dialogic and outward; clearly feminine and high-pitched; young and youthful; moderate in tempo; extremely carrying devotion; extremely carrying contentment; extremely carrying exhilaration; genuine in delivery; interwoven with vocal bursts (laughs, gasps, sighs).
Per-sentence captions + Variant B (sentence-level bursts)
“Oh hello there little one what a wonderful day to make a new friend.”
Delivered high in register; extremely one-dimensional in resonance; extremely dialogic and outward; extremely carrying forgiveness; extremely carrying gladness; extremely carrying joy.
Variant B: Delivered high in register; extremely one-dimensional in resonance; extremely dialogic and outward; extremely carrying forgiveness; extremely carrying gladness; extremely carrying joy. — no burst
Variant A — locator (precise inline bursts) 0 span(s) @0.7
Oh hello there little one what a wonderful day to make a new friend.
no span ≥ 0.7
charactersad_man59.7s · genu 1.78 · blend 5.5
ASR: It's my fault. I I should have been there I should have saved her.
Global caption
A voice that is extremely denasal and clear; extremely decelerating in pace; extremely soft and gentle in onset; extremely thin in chest resonance; very organic and soft; childlike; clearly feminine and high-pitched; low-to-mid in register; very slow and deliberate in tempo; extremely carrying remorse; extremely carrying solace; extremely carrying sorrow; genuine in delivery; interwoven with vocal bursts (laughs, gasps, sighs).
Per-sentence captions + Variant B (sentence-level bursts)
“It's my fault.”
Delivered very organic and soft; very slow and deliberate in tempo; clearly feminine and high-pitched; extremely carrying guilt; extremely carrying desperation; extremely carrying sorrow.
Variant B: Delivered very organic and soft; very slow and deliberate in tempo; clearly feminine and high-pitched; extremely carrying guilt; extremely carrying desperation; extremely carrying sorrow. — no burst
“I I should have been there I should have saved her.”
Delivered extremely decelerating in pace; extremely denasal and clear; extremely soft and gentle in onset; extremely carrying respite; extremely carrying sadness; extremely carrying disappointment.
Variant B: Delivered extremely decelerating in pace; extremely denasal and clear; extremely soft and gentle in onset; extremely carrying respite; extremely carrying sadness; extremely carrying disappointment, marked by a (Surprised Gasp). (Surprised Gasp)
Variant A — locator (precise inline bursts) 5 span(s) @0.7
(Slap Face) It's my fault. I I should have been there (Clears Throat) I should have saved her.
ASR: No no please Don't leave me to I can't I can't lose anyone else
Global caption
A voice that is extremely collapsing in energy; extremely denasal and clear; extremely thin in chest resonance; extremely decelerating in pace; extremely distressed and negative; childlike; clearly feminine and high-pitched; mid register; notably slow and deliberate in tempo; notably carrying submission; notably carrying humiliation; notably carrying suffering; only slightly genuine, somewhat performed.
Per-sentence captions + Variant B (sentence-level bursts)
“No no please Don't leave me to I can't I can't lose anyone else”
Delivered extremely denasal and clear; extremely collapsing in energy; extremely thin in chest resonance; notably carrying humiliation; notably carrying suffering; notably carrying submission.
Variant B: Delivered extremely denasal and clear; extremely collapsing in energy; extremely thin in chest resonance; notably carrying humiliation; notably carrying suffering; notably carrying submission, interrupted by a (Exasperated Sigh). (Exasperated Sigh)
Variant A — locator (precise inline bursts) 3 span(s) @0.7
No no please Don't leave me to (Deep Breath) I can't I can't lose anyone else (Surprised Gasp)
ASR: There now, my love, close your eyes. I have you, safe and warm.
Global caption
A voice that is extremely collapsing in energy; extremely blurry and mumbled; extremely raw and vulnerable; extremely denasal and clear; extremely choppy and fragmented; middle-aged to mature; masculine and deep-pitched; low and bassy in register; notably slow and deliberate in tempo; notably carrying comfort; notably carrying tenderness; notably carrying burnout; only slightly genuine, somewhat performed.
Per-sentence captions + Variant B (sentence-level bursts)
“There now, my love, close your eyes.”
Delivered extremely warm and enveloping; extremely full and cinematic; extremely beautiful and pleasing; notably carrying infatuation; notably carrying feeling horny; notably carrying clemency.
Variant B: Delivered extremely warm and enveloping; extremely full and cinematic; extremely beautiful and pleasing; notably carrying infatuation; notably carrying feeling horny; notably carrying clemency. — no burst
“I have you, safe and warm.”
Delivered extremely full and cinematic; extremely flat and unemphasized; extremely warm and enveloping; notably carrying warmth; notably carrying lust; notably carrying butterflies in the stomach.
Variant B: Delivered extremely full and cinematic; extremely flat and unemphasized; extremely warm and enveloping; notably carrying warmth; notably carrying lust; notably carrying butterflies in the stomach. — no burst
Variant A — locator (precise inline bursts) 3 span(s) @0.7
(Hiss) There now, my love, close your eyes. I have you, safe and warm. (Breathy Giggle)
ASR: Just relax now. Let it all go. You are safe here. Close your eyes.
Global caption
A voice that is extremely collapsing in energy; extremely meek and submissive; extremely soft and gentle in onset; extremely thin in chest resonance; extremely relaxed and loose; middle-aged to mature; clearly feminine and high-pitched; mid register; notably slow and deliberate in tempo; notably carrying tranquility; notably carrying trust; notably carrying feeling turned on; only slightly genuine, somewhat performed.
Per-sentence captions + Variant B (sentence-level bursts)
“Just relax now.”
Delivered extremely effortless and fluent; extremely meek and submissive; very relaxed and loose; notably carrying relief; notably carrying fulfillment; notably free of engrossment.
Variant B: Delivered extremely effortless and fluent; extremely meek and submissive; very relaxed and loose; notably carrying relief; notably carrying fulfillment; notably free of engrossment. — no burst
“Let it all go.”
Delivered extremely flat and unemphasized; extremely collapsing in energy; extremely blurry and mumbled; notably carrying melancholy; notably carrying chagrin; notably carrying existential void.
Variant B: Delivered extremely flat and unemphasized; extremely collapsing in energy; extremely blurry and mumbled; notably carrying melancholy; notably carrying chagrin; notably carrying existential void. — no burst
“You are safe here.”
Delivered extremely effortless and fluent; extremely meek and submissive; very organic and soft; notably carrying alleviation; notably carrying clemency; notably carrying lightness.
Variant B: Delivered extremely effortless and fluent; extremely meek and submissive; very organic and soft; notably carrying alleviation; notably carrying clemency; notably carrying lightness. — no burst
“Close your eyes.”
Delivered extremely collapsing in energy; extremely soft and gentle in onset; extremely flat and unemphasized; notably carrying carnal desire; notably carrying fondness; notably carrying comfort.
Variant B: Delivered extremely collapsing in energy; extremely soft and gentle in onset; extremely flat and unemphasized; notably carrying carnal desire; notably carrying fondness; notably carrying comfort, punctuated by a (Wistful Sigh). (Wistful Sigh)
Variant A — locator (precise inline bursts) 3 span(s) @0.7
Just relax now. Let it all go. You are safe here. Close your eyes. (Lip Smack)(Slap Face)
A voice that is extremely blurry and mumbled; extremely soft and gentle in onset; extremely dark and muffled; extremely flat and monotone; extremely collapsing in energy; adult; masculine and deep-pitched; low and bassy in register; notably slow and deliberate in tempo; extremely carrying lethargy; extremely carrying feeling turned on; extremely carrying infatuation; genuine in delivery; interwoven with vocal bursts (laughs, gasps, sighs).
Per-sentence captions + Variant B (sentence-level bursts)
“Tiva.”
Delivered extremely collapsing in energy; extremely blurry and mumbled; extremely flat and unemphasized; extremely carrying fondness; notably free of concentration; extremely carrying carnal desire.
Variant B: Delivered extremely collapsing in energy; extremely blurry and mumbled; extremely flat and unemphasized; extremely carrying fondness; notably free of concentration; extremely carrying carnal desire, carrying a (Gulps). (Gulps)
“Tia Baya Gokufu Tsiwa.”
Delivered extremely collapsing in energy; very blurry and mumbled; very thin in head resonance; notably free of deep focus; notably carrying burnout; notably carrying indifference.
Variant B: Delivered extremely collapsing in energy; very blurry and mumbled; very thin in head resonance; notably free of deep focus; notably carrying burnout; notably carrying indifference. — no burst
Variant A — locator (precise inline bursts) 5 span(s) @0.7
Tiva. (Slap Face)(Deep Breath) Tia Baya Gokufu Tsiwa. (Clears Throat)
0.1–0.2s → discarded (P(no_burst)=1.00)0.9–1.1s → Slap Face (0.94)1.1–2.1s → Deep Breath (0.89)2.5–2.8s → discarded (P(no_burst)=0.60)3.6–3.8s → Clears Throat (0.80)
A voice that is extremely raw and vulnerable; extremely collapsing in energy; extremely audibly breathing and gasping; extremely thin in chest resonance; extremely meek and submissive; adult; feminine and high-pitched; low-to-mid in register; somewhat slow and deliberate in tempo; extremely carrying agony; extremely carrying pessimism; extremely carrying helplessness; deeply and genuinely felt; interwoven with vocal bursts (laughs, gasps, sighs).
Per-sentence captions + Variant B (sentence-level bursts)
“Ah Najulia Leba in Chuchu Chuchu”
Delivered extremely raw and vulnerable; extremely audibly breathing and gasping; extremely collapsing in energy; extremely carrying suffering; extremely carrying trepidation; extremely carrying desperation.
Variant B: Delivered extremely raw and vulnerable; extremely audibly breathing and gasping; extremely collapsing in energy; extremely carrying suffering; extremely carrying trepidation; extremely carrying desperation. — no burst
Variant A — locator (precise inline bursts) 4 span(s) @0.7
Ah (Slap Face)(Slap Face) Najulia Leba in Chuchu Chuchu
0.1–0.3s → discarded (P(no_burst)=0.60)0.3–0.5s → Slap Face (0.99)0.6–0.8s → Slap Face (0.97)4.3–4.5s → discarded (P(no_burst)=0.59)
A voice that is extremely full and cinematic; extremely warm and enveloping; extremely beautiful and pleasing; extremely rough and raspy; very thin in head resonance; middle-aged to mature; strongly masculine, deep and low-pitched; low and bassy in register; somewhat slow and deliberate in tempo; notably carrying misery; notably carrying desire to harm; notably carrying feeling turned on; only slightly genuine, somewhat performed; interwoven with vocal bursts (laughs, gasps, sighs).
Per-sentence captions + Variant B (sentence-level bursts)
“Hmm so the memento mori has fallen.”
Delivered extremely full and cinematic; extremely warm and enveloping; extremely beautiful and pleasing; notably carrying desire to harm; notably carrying carnal desire; notably carrying fondness.
Variant B: Delivered extremely full and cinematic; extremely warm and enveloping; extremely beautiful and pleasing; notably carrying desire to harm; notably carrying carnal desire; notably carrying fondness. — no burst
Variant A — locator (precise inline bursts) 1 span(s) @0.7
ASR: Suis-je au paradis ? Sur l'île il est bien heureux.
Global caption
A voice that is extremely raw and vulnerable; extremely noisy and aperiodic; extremely rough and raspy; extremely audibly breathing and gasping; extremely darkening in mood; middle-aged to mature; masculine and deep-pitched; low and bassy in register; somewhat slow and deliberate in tempo; extremely carrying horror; extremely carrying worry; extremely carrying powerlessness; deeply and genuinely felt; interwoven with vocal bursts (laughs, gasps, sighs).
Per-sentence captions + Variant B (sentence-level bursts)
“Suis-je au paradis ?”
Delivered extremely blurry and mumbled; extremely collapsing in energy; extremely noisy and aperiodic; notably free of attention; extremely carrying powerlessness; extremely carrying lethargy.
Variant B: Delivered extremely blurry and mumbled; extremely collapsing in energy; extremely noisy and aperiodic; notably free of attention; extremely carrying powerlessness; extremely carrying lethargy, marked by a (Ahem). (Ahem)
“Sur l'île il est bien heureux.”
Delivered very dialogic and outward; very balanced and blended in resonance; very thin in head resonance; notably free of attention; very carrying nostalgia; very carrying helplessness.
Variant B: Delivered very dialogic and outward; very balanced and blended in resonance; very thin in head resonance; notably free of attention; very carrying nostalgia; very carrying helplessness. — no burst
Variant A — locator (precise inline bursts) 4 span(s) @0.7
Suis-je au paradis ? (Slap Face)(Slap Face) Sur l'île il est bien heureux.
2.5–2.7s → Slap Face (0.97)2.9–3.3s → Slap Face (0.94)3.6–3.7s → discarded (P(no_burst)=1.00)4.5–4.6s → discarded (P(no_burst)=1.00)
ASR: So that people go, What is that? I need to find out about that. So this is me making stuff up right in the moment, but I'm like, okay, I'm gonna use these oblique strategies as a way of
Global caption
A voice that is very bright in oral resonance; notably forward in mask resonance; notably brightening in mood; notably loud and non-whispered; notably disfluent and halting; young and youthful; masculine and deep-pitched; low and bassy in register; somewhat fast and rapid in tempo; extremely carrying joviality; notably carrying indifference; notably carrying impatience; deeply and genuinely felt; interwoven with vocal bursts (laughs, gasps, sighs).
Per-sentence captions + Variant B (sentence-level bursts)
“So that people go, What is that?”
Delivered extremely balanced and blended in resonance; very bright in oral resonance; very dialogic and outward; notably free of engrossment; very carrying provoking lightly; very carrying disdain.
Variant B: Delivered extremely balanced and blended in resonance; very bright in oral resonance; very dialogic and outward; notably free of engrossment; very carrying provoking lightly; very carrying disdain. — no burst
“I need to find out about that.”
Delivered extremely balanced and blended in resonance; very seamless and inaudible in breathing; very bright in oral resonance; notably free of deep focus; notably free of meditation; notably free of curiosity.
Variant B: Delivered extremely balanced and blended in resonance; very seamless and inaudible in breathing; very bright in oral resonance; notably free of deep focus; notably free of meditation; notably free of curiosity. — no burst
“So this is me making stuff up right in the moment, but I'm like, okay, I'm gonna use these oblique strategies as a way of”
Delivered extremely bright in oral resonance; notably joyful and positive; notably naturalistic and understated; extremely carrying ribbing; extremely carrying numbness; very carrying distrust.
Variant B: Delivered extremely bright in oral resonance; notably joyful and positive; notably naturalistic and understated; extremely carrying ribbing; extremely carrying numbness; very carrying distrust. — no burst
Variant A — locator (precise inline bursts) 0 span(s) @0.7
So that people go, What is that? I need to find out about that. So this is me making stuff up right in the moment, but I'm like, okay, I'm gonna use these oblique strategies as a way of
ASR: Soudain, on frappe à sa porte. Chloé sursaute. Elle se sent incapable de bouger. On frappe à nouveau. Chloé retient sa respiration. Elle a peur. Elle a l'impression de vivre un mauvais rêve.
Global caption
A voice that is elderly-sounding; extremely narrator-like and storybook; extremely news-anchor-like; extremely rough and raspy; very thin in mask resonance; masculine and deep-pitched; low and bassy in register; somewhat slow and deliberate in tempo; notably carrying relief; notably carrying alarm; notably carrying submission; only slightly genuine, somewhat performed; interwoven with vocal bursts (laughs, gasps, sighs).
Per-sentence captions + Variant B (sentence-level bursts)
“Soudain, on frappe à sa porte.”
Delivered extremely pristine in recording quality; very clean and quiet in background; notably thin in throat resonance; notably free of engrossment; notably carrying liberation; notably carrying surprise.
Variant B: Delivered extremely pristine in recording quality; very clean and quiet in background; notably thin in throat resonance; notably free of engrossment; notably carrying liberation; notably carrying surprise. — no burst
“Chloé sursaute.”
Delivered extremely choppy and fragmented; extremely collapsing in energy; very thin in head resonance; notably carrying disgust; notably free of concentration; notably carrying schadenfreude.
Variant B: Delivered extremely choppy and fragmented; extremely collapsing in energy; very thin in head resonance; notably carrying disgust; notably free of concentration; notably carrying schadenfreude, with an audible (Displeased Grunt). (Displeased Grunt)
“Elle se sent incapable de bouger.”
Delivered very denasal and clear; notably pristine in recording quality; notably full and cinematic; notably free of absorption; notably carrying solace; notably carrying helplessness.
Variant B: Delivered very denasal and clear; notably pristine in recording quality; notably full and cinematic; notably free of absorption; notably carrying solace; notably carrying helplessness. — no burst
“On frappe à nouveau.”
Delivered very jerky and uneven in rhythm; very unpleasant and harsh; very choppy and fragmented; notably free of deep focus; notably carrying anticipation; notably carrying bliss.
Variant B: Delivered very jerky and uneven in rhythm; very unpleasant and harsh; very choppy and fragmented; notably free of deep focus; notably carrying anticipation; notably carrying bliss. — no burst
“Chloé retient sa respiration.”
Delivered extremely full and cinematic; very thin in head resonance; very organic and soft; notably carrying liberation; notably free of engrossment; notably carrying agony.
Variant B: Delivered extremely full and cinematic; very thin in head resonance; very organic and soft; notably carrying liberation; notably free of engrossment; notably carrying agony, punctuated by a (Exasperated Sigh). (Exasperated Sigh)
“Elle a peur.”
Delivered extremely collapsing in energy; extremely choppy and fragmented; extremely unpleasant and harsh; notably carrying jealousy; notably carrying horror; notably carrying helplessness.
Variant B: Delivered extremely collapsing in energy; extremely choppy and fragmented; extremely unpleasant and harsh; notably carrying jealousy; notably carrying horror; notably carrying helplessness. — no burst
“Elle a l'impression de vivre un mauvais rêve.”
Delivered extremely balanced and blended in resonance; extremely denasal and clear; extremely effortless and fluent; notably carrying wonder; notably carrying panic; notably carrying pining.
Variant B: Delivered extremely balanced and blended in resonance; extremely denasal and clear; extremely effortless and fluent; notably carrying wonder; notably carrying panic; notably carrying pining. — no burst
Variant A — locator (precise inline bursts) 0 span(s) @0.7
Soudain, on frappe à sa porte. Chloé sursaute. Elle se sent incapable de bouger. On frappe à nouveau. Chloé retient sa respiration. Elle a peur. Elle a l'impression de vivre un mauvais rêve.
A voice that is extremely accelerating in pace; extremely pristine in recording quality; extremely warm and enveloping; very beautiful and pleasing; very effortless and fluent; adult; feminine and high-pitched; low-to-mid in register; notably slow and deliberate in tempo; notably carrying fondness; notably carrying pleasure; notably carrying ease; only slightly genuine, somewhat performed; interwoven with vocal bursts (laughs, gasps, sighs).
Per-sentence captions + Variant B (sentence-level bursts)
“The woman coos and sings a sweet song”
Delivered extremely accelerating in pace; extremely pristine in recording quality; extremely warm and enveloping; notably carrying butterflies in the stomach; notably carrying relaxation; notably carrying ecstasy.
Variant B: Delivered extremely accelerating in pace; extremely pristine in recording quality; extremely warm and enveloping; notably carrying butterflies in the stomach; notably carrying relaxation; notably carrying ecstasy. — no burst
Variant A — locator (precise inline bursts) 0 span(s) @0.7
ASR: Hue Shvung Sam Yiv Bush Sam Y forg. Some Y for some yung.
Global caption
A voice that is notably denasal and clear; notably bright in oral resonance; notably organic and soft; notably thin in mask resonance; somewhat authoritative and commanding; adult; strongly masculine, deep and low-pitched; low and bassy in register; somewhat fast and rapid in tempo; extremely carrying adoration; extremely carrying feeling turned on; extremely carrying alleviation; deeply and genuinely felt; interwoven with vocal bursts (laughs, gasps, sighs).
Per-sentence captions + Variant B (sentence-level bursts)
“Hue Shvung Sam Yiv Bush Sam Y forg.”
Delivered extremely explicit in content; very denasal and clear; very noisy and aperiodic; notably carrying infatuation; notably carrying tenderness; notably carrying sexual lust.
Variant B: Delivered extremely explicit in content; very denasal and clear; very noisy and aperiodic; notably carrying infatuation; notably carrying tenderness; notably carrying sexual lust. — no burst
“Some Y for some yung.”
Delivered very seamless and inaudible in breathing; very bright in oral resonance; notably clean and quiet in background; notably free of absorption; notably free of interest; notably free of pondering.
Variant B: Delivered very seamless and inaudible in breathing; very bright in oral resonance; notably clean and quiet in background; notably free of absorption; notably free of interest; notably free of pondering. — no burst
Variant A — locator (precise inline bursts) 1 span(s) @0.7
Hue Shvung Sam Yiv Bush Sam Y forg. Some Y for some yung.
3.0–3.1s → discarded (P(no_burst)=0.82)
Generated by build_burst_demo.py ·
pipeline in burst_captions.py. Bursts, emotions, dimensions and timestamps are all model predictions.