Burp Data: 1,500 English and Chinese Speech Clips with Burp Vocal Events

1,500 curated English and Chinese speech clips containing burp non-verbal vocal events, built by filtering and validating TTS-generated samples with Gemini 2.5 Pro audio understanding. The dataset was created by NVVSpeech-Challenge and last updated on July 11, 2026. It is intended exclusively for the NVVSpeech Challenge and not for commercial use or production deployment.

Use Cases

• Training models for non-verbal vocal event detection based on burp sounds.
• Benchmarking audio understanding models on TTS-generated speech clips.
• Studying cross-lingual characteristics of non-verbal sounds in English and Chinese speech.

Strengths

• Contains 1,500 curated speech clips.
• Includes both English and Chinese audio samples.
• Data was validated using Gemini 2.5 Pro audio understanding.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is known but total dataset size and file formats are unknown.
• Data is not intended for commercial use, limiting its application scope.

Provenance

Source

NVVSpeech-Challenge on Hugging Face.

Collection Method

Built by filtering and validating TTS-generated samples with Gemini 2.5 Pro audio understanding.

Freshness

Last updated 2026-07-11 12:29:01; freshness should be verified.