Articulatory controllable speech modification based on Gaussian mixture models with direct waveform modification using spectrum differential.
Patrick Lumban TobingKazuhiro KobayashiTomoki TodaGraham NeubigSakriani SaktiSatoshi NakamuraPublished in: INTERSPEECH (2015)
Keyphrases
- gaussian mixture model
- speaker recognition
- speech recognition
- mixture model
- speaker identification
- em algorithm
- speech signal
- maximum likelihood
- pattern recognition
- vocal tract
- multi stream
- image processing
- frequency domain
- vector quantization
- variational bayes
- probabilistic model
- speech synthesis
- fundamental frequency