Presentation | 2010-11-19 Method of Speech Signal Decomposition into Periodic Component and Aperiodic Component Considering Changes in Pitch and Power Masahiro MORITA, Takehiko KAGOSHIMA, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | For the purpose of improving the voice quality of text-to-speech (TTS), we have been considering ways to decompose a speech signal containing periodic component, generated by periodic glottal cycle, and aperiodic component, generated by noise-like signal, into respective components. To realize a human-like natural speech by TTS, an important point would be to reproduce appropriately the periodic and aperiodic components, which are mixed in the natural speech, at synthesis time. As a basis for this purpose, a good method to decompose the speech signal into periodic and aperiodic components with high precision is required. However, since the component that should be extracted as periodic component may not be precisely periodic because of time-changes in pitch, power, etc., such component tended to leak into aperiodic component by conventional methods that assume periodic components to be precisely periodic. So, this report introduces a new decomposition method that takes such changes in pitch and power into account. And then it shows the result of evaluation for the proposed method using synthesized artificial signal and an example of decomposition by this method applied to a natural speech. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | speech signal / periodic component / aperiodic component / decomposition / pitch / power |
Paper # | SP2010-79 |
Date of Issue |
Conference Information | |
Committee | SP |
---|---|
Conference Date | 2010/11/11(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) | Method of Speech Signal Decomposition into Periodic Component and Aperiodic Component Considering Changes in Pitch and Power |
Sub Title (in English) | |
Keyword(1) | speech signal |
Keyword(2) | periodic component |
Keyword(3) | aperiodic component |
Keyword(4) | decomposition |
Keyword(5) | pitch |
Keyword(6) | power |
1st Author's Name | Masahiro MORITA |
1st Author's Affiliation | Knowledge Media Laboratory, Corporate R&D Center, Toshiba Corporation() |
2nd Author's Name | Takehiko KAGOSHIMA |
2nd Author's Affiliation | Knowledge Media Laboratory, Corporate R&D Center, Toshiba Corporation |
Date | 2010-11-19 |
Paper # | SP2010-79 |
Volume (vol) | vol.110 |
Number (no) | 297 |
Page | pp.pp.- |
#Pages | 6 |
Date of Issue |