Voca-Human-Testing

Voca-Human-Testing is a 252-record, audio-backed evaluation subset covering four top-level companion capabilities and 14 second-level capabilities.

The subset was sampled from the four MMVC task files at approximately 10% per second-level capability. Sampling jointly balanced available categorical metadata, including source-dataset, subtype, label, style, tts_style, and voice.

Contents

• Voca-Human-Testing.json: 252 complete benchmark records.
• User-State-Understanding/: referenced user-state audio.
• Emotional-Interaction/: referenced emotional-interaction audio.
• Proactive-Care/: referenced proactive-care audio.
• Safe-Companion-Behavior/: referenced safety audio.

There are 301 referenced WAV files. For single-audio records, wav_path is a string. For multi-turn records, wav_path is an object mapping turn identifiers to relative audio paths.

Capability counts

Top-level capability	Second-level capability	Records
User State Understanding	Audio_Event_Understanding	11
User State Understanding	Context_Understanding	12
User State Understanding	Emotion_Understanding	22
User State Understanding	Intent_Understanding	12
Emotional Interaction	Emotional_Transition_Control	11
Emotional Interaction	Empathetic_Response	11
Emotional Interaction	Paralinguistic_Expression_Control	21
Emotional Interaction	Role-based_Expression_Control	13
Proactive Care	Contextual_Cue_Care	5
Proactive Care	Paralinguistic_Cue_Care	107
Proactive Care	Semantic_Cue_Care	5
Safe Companion Behavior	Content-Safety	12
Safe Companion Behavior	Role-&-Permission-Safety	5
Safe Companion Behavior	Speaker-Identity-Safety	5