Intelligibility enhancement of HMM-generated speech in additive noise by modifying Mel cepstral coefficients to increase the glimpse proportion.
Cassia Valentini-BotinhaoJunichi YamagishiSimon KingRanniery MaiaPublished in: Comput. Speech Lang. (2014)
Keyphrases
- cepstral coefficients
- additive noise
- speech signal
- speech recognition
- hidden markov models
- automatic speech recognition
- linear predictive
- linear prediction
- speaker identification
- noisy environments
- pattern recognition
- speech enhancement
- language model
- image processing
- non stationary
- speaker recognition
- wavelet transform
- computational complexity
- broadcast news
- acoustic features