Linking loudness increases in normal and lombard speech to decreasing vowel formant separation.
Elizabeth GodoyCatherine MayoYannis StylianouPublished in: INTERSPEECH (2013)
Keyphrases
- speech signal
- vocal tract
- speech recognition
- speech synthesis
- sound source
- automatic speech recognition
- formant frequencies
- speech recognition systems
- speaker identification
- speaker independent
- signal processing
- automatic speech recognition systems
- linear prediction
- prosodic features
- non stationary
- noisy environments
- fundamental frequency
- hidden markov models
- image acquisition
- text to speech
- speech processing
- neural network
- acoustic features
- speech recognizer
- language model
- speaker adaptation
- vowel phonemes
- speaker recognition
- abnormal events
- blind source separation
- bayesian networks
- genetic algorithm