VoiceBank-DEMAND
The Edinburgh release provides paired noisy and clean utterances, plus text for speech enhancement and noise-robust speech workflows.
Visit official source (opens in a new tab)Research use
Compare enhancement pipelines against clean targets and test how preprocessing changes recognition on noisy speech.
Limitations
The noisy signals use constructed mixtures. Performance on natural conversational recordings may differ.
License and access
- Dataset license
- CC BY 4.0 (opens in a new tab)
- Access status
- Direct access
Select matching clean, noisy and text archives from the university repository.