AVE Speech: Mandarin Speech Corpus with Audio, Video, and EMG Signals

Mandarin speech data from 100 native speakers, each reading 100 sentences ten times. The dataset provides over 55 hours of synchronized audio, lip video, and surface electromyography recordings per modality. It was created by MML-Group and last updated on the platform in October 2025.

Use Cases

• Train multimodal speech recognition models based on synchronized audio, video, and EMG signals.
• Research silent speech interfaces based on surface electromyography recordings.
• Develop lip-reading models based on the provided lip video data.
• Study speaker-independent speech patterns based on data from 100 native Mandarin speakers.

Strengths

• Contains data from 100 native Mandarin speakers, providing speaker diversity.
• Includes over 55 hours of recordings per modality (audio, video, EMG).
• Features synchronized multi-modal signals (audio, video, EMG) for each utterance.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Data may reflect geographic or linguistic bias inherent to its single-language, lab-recorded collection.

Provenance

Source

MML-Group

Collection Method

Recordings of 100 native speakers reading a set of 100 sentences, each repeated ten times.

Freshness

Last updated 2025-10-11 07:09:21; freshness should be verified.

Geography

Mandarin-speaking region(s); specific location is unknown.