Emotiontalk: 19,250 Multimodal Chinese Utterances for Dyadic Interaction

Emotiontalk contains 19,250 utterances totaling 23.6 hours of dyadic conversation speech recorded from 19 actors. Produced by BAAI and updated in 2025, it provides synchronized acoustic, visual, and textual data for Chinese emotional interaction analysis.

Use Cases

• Multimodal emotion recognition using the 7 utterance-level emotion categories
• Acoustic sentiment analysis using the 23.6 hours of speech data
• Visual gesture and expression modeling from the video modality

Strengths

• 19,250 annotated utterances
• 23.6 hours of speech data
• 7 utterance-level emotion categories
• Synchronized acoustic, visual, and textual modalities

Limitations

• Small sample size of 19 actors
• Language restricted to Chinese
• Actor-based performance may differ from spontaneous real-world emotion

Provenance

Source

BAAI (Beijing Academy of Artificial Intelligence)

Collection Method

Recorded dyadic conversations between 19 actors

Freshness

Updated August 2025

Geography

China