{
  "id": "humanness-annotator-audio-review-v1",
  "scale": "Integer 0–5: 0 unusable/wrong; 1 major errors; 2 partly supported; 3 useful with notable errors; 4 accurate with minor errors; 5 fully correct and useful",
  "composite_weights": {
    "global_caption_fidelity": 0.15,
    "global_emotion_fit": 0.10,
    "sentence_emotion_fit": 0.20,
    "utterance_talking_style_fit": 0.15,
    "sentence_delivery_fit": 0.10,
    "burst_description_fit": 0.10,
    "burst_timing_fit": 0.10,
    "event_precision": 0.05,
    "event_coverage": 0.05
  },
  "definitions": {
    "global_caption_fidelity": "Does GENERAL describe the audible voice, performance and environment accurately, without invented demographic/identity/psychological claims?",
    "global_emotion_fit": "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.",
    "sentence_emotion_fit": "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.",
    "utterance_talking_style_fit": "Do GENERAL's speaking-style claims match audible tempo, articulation, breathiness, pitch movement, register, emphasis and performance style?",
    "sentence_delivery_fit": "Do local delivery/voice tags fit the corresponding spoken segment rather than merely repeating a global trait?",
    "burst_description_fit": "Does each free-form burst caption match the actual nonverbal vocal sound in its selected interval? Reward acoustic accuracy, not verbosity.",
    "burst_timing_fit": "Do absolute event starts/ends cover the audible events without unnecessary unrelated speech?",
    "event_precision": "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": "Are salient audible nonverbal events covered? This is a reviewer rating, not measured detection recall.",
    "transcript_fidelity": "Does the recognized word sequence match what is heard? This is diagnostic, not part of the weighted composite.",
    "speech_timing_fit": "Do speech budgets and pause intervals fit the audio? This is diagnostic, separate from mechanical formatting validation.",
    "readability": "Is the annotation understandable, concise and useful? This is diagnostic, not part of the weighted composite."
  },
  "aggregation": "Equal weight per recording; fixed weights above. Per-segment ratings are diagnostic and do not overweight recordings with more sentences.",
  "empty_event_policy": "If no event is audible and none is proposed, event ratings are 5. False proposals or omitted audible events reduce the corresponding ratings.",
  "important_limitations": "Single automated reviewer; 50 reused synthetic acting recordings, 45 English/5 German; no human reference labels, no inter-rater agreement. New rubric is not comparable numerically to the older MOSS/procedural composite. No settings are optimized using these new ratings."
}