Time-domain deterministic plus noise model based hybrid source modeling for statistical parametric speech synthesis.
N. P. NarendraK. Sreenivasa RaoPublished in: Speech Commun. (2016)
Keyphrases
- speech synthesis
- speech recognition
- data driven
- text to speech
- statistical analysis
- vocal tract
- hybrid models
- genetic algorithm and simulated annealing
- frequency domain
- prosodic features
- statistical modeling
- noisy data
- noise reduction
- noise level
- random noise
- noisy environments
- noise model
- data mining
- black box
- statistical models
- missing data
- computer vision