Spectral tilt modelling with GMMs for intelligibility enhancement of narrowband telephone speech.
Emma JokinenUlpu RemesMarko TakanenKalle J. PalomäkiMikko KurimoPaavo AlkuPublished in: INTERSPEECH (2014)
Keyphrases
- speech signal
- speech recognition
- spectral analysis
- linear prediction
- spectral features
- gaussian mixture model
- automatic speech recognition
- linear predictive coding
- cepstral coefficients
- speaker identification
- speaker recognition
- noisy environments
- speech quality
- speech synthesis
- speech recognizer
- hidden markov models
- image processing
- language model
- signal to noise ratio
- vocal tract
- image enhancement
- multi modal
- speaker adaptation
- call center
- emotion recognition
- audio visual
- feature vectors
- spectral images
- maximum likelihood
- non stationary