PAAPLoss: A Phonetic-Aligned Acoustic Parameter Loss for Speech Enhancement.
Muqiao YangJoseph KonanDavid BickYunyang ZengShuo HanAnurag KumarShinji WatanabeBhiksha RajPublished in: CoRR (2023)
Keyphrases
- speech enhancement
- noisy speech
- speech recognition
- speech sounds
- speech signal
- vocal tract
- noisy environments
- signal to noise ratio
- noise reduction
- single channel
- automatic speech recognition
- emotional speech
- linear prediction
- background noise
- hidden markov models
- wiener filter
- acoustic features
- speech synthesis
- language model
- pattern recognition
- edge detection
- image acquisition
- multiscale