DDS: A new device-degraded speech dataset for speech enhancement.
Haoyu LiJunichi YamagishiPublished in: INTERSPEECH (2022)
Keyphrases
- speech enhancement
- noise reduction
- speech signal
- noisy environments
- signal to noise ratio
- vocal tract
- single channel
- noisy speech
- speech recognition
- linear prediction
- smoothing algorithm
- speaker identification
- additive noise
- background noise
- wiener filter
- image processing
- edge preserving
- acoustic features
- speech synthesis
- denoising
- hidden markov models
- automatic speech recognition
- spatial domain
- feature set
- probabilistic model