Maximum likelihood voice conversion based on GMM with STRAIGHT mixed excitation.
Yamato OhtaniTomoki TodaHiroshi SaruwatariKiyohiro ShikanoPublished in: INTERSPEECH (2006)
Keyphrases
- maximum likelihood
- gaussian mixture model
- em algorithm
- voice activity detection
- expectation maximization
- mixture model
- parameter estimation
- gaussian distribution
- likelihood function
- speaker recognition
- maximum a posteriori
- line segments
- additive noise
- maximum likelihood estimation
- gaussian mixture
- variational bayes
- mel frequency cepstral coefficients
- gaussian mixture modeling
- exponential family
- background subtraction
- density estimation
- speaker identification
- log likelihood function
- feature selection
- hyperparameters
- data sets
- image segmentation
- multiscale
- text to speech