Spectral Features for Perceptually Natural Phoneme Replacement by Another Speaker's Speech.
Reiko TakouHiroyuki SegiTohru TakagiNobumasa SeiyamaPublished in: IEICE Trans. Fundam. Electron. Commun. Comput. Sci. (2012)
Keyphrases
- spectral features
- speech recognition
- mel frequency cepstral coefficients
- speaker dependent
- automatic speech recognition
- spectral analysis
- automatic speech recognition systems
- speech signal
- speech synthesis
- remote sensing images
- speaker identification
- phoneme recognition
- structural features
- speech sounds
- hyperspectral images
- hidden markov models
- speaker recognition
- prosodic features
- speech recognizer
- visual speech
- speaker verification
- vocal tract
- speaker independent
- text to speech
- change detection
- speaker diarization
- language model
- pattern recognition
- information content
- computer vision
- higher level
- broadcast news
- speaker adaptation
- audio visual
- noisy environments