| Presentation | 1999/5/20 A New Speech Synthesis Method based on Vocoder Preserving Fine Structure of Magnitude Spectrum Satoshi TAKANO, Masanobu ABE, |
|---|---|
| PDF Download Page | PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | The TD-PSOLA is the most popular synthesis algorithm in current TTS systems, it still has a problem in that range of modification wherein naturalness is retained is narrow, and large prosody modification introduces evident speech distortion. To solve the problem, this paper proposes a new speech modification method based on vocoder for high quality TTS system. We used STRAIGHT for synthesis part of vocoder. We found out that preserving fine structure of the magnitude spectrum by compensatory gaussian window FFT makes it possible to synthesize high quality speech. So we propose that harmonics should be modified not only to match the target F_0 value but also to preserve fine structure. Preference test showed that proposed method synthesizes higher quality speech than TD-PSOLA in large prosody modification, and spectral envelop from proposed method is superior to STRAIGHT-envelope especially when used for modifying frequency upward. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | text-to-speech synthesis / harmonics modification / fundamental frequency / vocoder |
| Paper # | SP99-5 |
| Date of Issue |
| Conference Information | |
| Committee | SP |
|---|---|
| Conference Date | 1999/5/20(1days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | |
| Chair | |
| Vice Chair | |
| Secretary | |
| Assistant | |
| Paper Information | |
| Registration To | Speech (SP) |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | A New Speech Synthesis Method based on Vocoder Preserving Fine Structure of Magnitude Spectrum |
| Sub Title (in English) | |
| Keyword(1) | text-to-speech synthesis |
| Keyword(2) | harmonics modification |
| Keyword(3) | fundamental frequency |
| Keyword(4) | vocoder |
| 1st Author's Name | Satoshi TAKANO |
| 1st Author's Affiliation | NTT Cyber Space Labs.() |
| 2nd Author's Name | Masanobu ABE |
| 2nd Author's Affiliation | NTT Cyber Space Labs. |
| Date | 1999/5/20 |
| Paper # | SP99-5 |
| Volume (vol) | vol.99 |
| Number (no) | 73 |
| Page | pp.pp.- |
| #Pages | 8 |
| Date of Issue | |