Audio-Visual (AV) Convo
Dataset Description
Audio-Visual (AV) Convo is a multimodal dataset of natural conversations between pairs of friends, captured with two simultaneous camera angles and a separate audio recording. Each sample is a side-by-side video combining both camera views with the original audio track.
Key Characteristics
- Conversations between friends talking about casual/life topics
- Diverse demographics: participants range in age (18–75), race, ethnicity, and gender
- Rich paralinguistic features: interruptions, backchanneling, laughter, and other natural speech phenomena
- Range of emotions: sadness, laughter, frustration, and more
Technical Specs
Note: The samples in this dataset viewer are compressed side-by-side MP4 previews. The actualy dataset would contain the original files at the specifications below.
Video
| Metric | Value |
|---|---|
| Resolution | 3840 x 2160 (4K) per camera |
| Codec | H.264 (High profile) |
| Frame rate | 30 fps |
| Pixel format | yuv420p |
| Color space | BT.709 |
| Bit depth | 8-bit |
| Average bitrate | 44.95 Mbps |
| Scan type | Progressive |
| Cameras per session | 2 |
Audio
| Metric | Value |
|---|---|
| File format | WAV |
| Codec | PCM (signed 24-bit little-endian) |
| Sample rate | 48,000 Hz |
| Bit depth | 24-bit |
| Channels | 2 (stereo) |
| Recording setup | Channel-separated, each speaker wears an individual lavalier mic |
- Downloads last month
- 58