Whisper to normal speech conversion using pitch estimated from spectrum.
Hideaki KonnoMineichi KudoHideyuki ImaiMasanori SugimotoPublished in: Speech Commun. (2016)
Keyphrases
- formant frequencies
- fundamental frequency
- speech recognition
- speech signal
- acoustic features
- data mining
- automatic speech recognition
- audio visual
- real time
- broadcast news
- endpoint detection
- data sets
- multi stream
- speech processing
- machine learning
- english text
- speaker recognition
- multi modal
- spoken language
- accurate estimation
- dialogue system
- database
- pattern recognition
- maximum likelihood
- estimation error
- signal to noise ratio