A frequency-weighted post-filtering transform for compensation of the over-smoothing effect in HMM-based speech synthesis.
Florian EybenYannis AgiomyrgiannakisPublished in: ICASSP (2014)
Keyphrases
- speech synthesis
- speech recognition
- text to speech
- high pass
- hidden markov models
- vocal tract
- filtering process
- prosodic features
- filtering method
- filtering algorithm
- image filtering
- gaussian filter
- noise removal
- low pass
- bilateral filtering
- bandpass
- directional filtering
- speech corpus
- speech signal
- image acquisition
- adaptive filtering
- neural network
- order statistics
- denoising
- weighted averaging
- pattern recognition
- smoothing filter
- image processing
- directional filters
- information retrieval