Source-filter models for time-scale pitch-scale modification of speech
Author:
- 27 November 2002
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 2, 881-884
Abstract
This paper presents two time-scale pitch-scale modification techniques to be used in speech synthesis systems. They have been applied to Microsoft's Whistler system, which is based on concatenative synthesis. Both methods are based on a source-filter model, one of them using LPC parameters and the other one using cepstral parameters. The proposed methods achieve high quality prosody modification, retain the characteristics of the donor speaker, allow for spectral manipulation (to reduce spectral discontinuities at unit boundaries), yield compact acoustic inventories and improved voiced fricativesKeywords
This publication has 10 references indexed in Scilit:
- Whistler: a trainable text-to-speech system2002
- Recent improvements on Microsoft's trainable text-to-speech system-Whistler2002
- Unit selection in a concatenative speech synthesis system using a large speech database1996
- Improvements in an HMM-based speech synthesiser1995
- Transformation and decomposition of the speech signal for codingIEEE Signal Processing Letters, 1994
- Efficient vector quantization of LPC parameters at 24 bits/frameIEEE Transactions on Speech and Audio Processing, 1993
- Pitch-synchronous waveform processing techniques for text-to-speech synthesis using diphonesSpeech Communication, 1990
- Two-channel speech analysisIEEE Transactions on Acoustics, Speech, and Signal Processing, 1986
- Maximum likelihood pitch estimationIEEE Transactions on Acoustics, Speech, and Signal Processing, 1976
- Discrete representation of signalsProceedings of the IEEE, 1972