| Presentation | 2007/3/19 Speaker conversion system based on vowels An implementation of voice texture mapping Toru TAKAHASHI, Masanori MORISE, Ryuichi NISIMURA, Toshio IRINO, Hideki BANNO, Hideki KAWAHARA, |
|---|---|
| PDF Download Page | PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | A simple and high-quality voice conversion procedure only depends on vowel information is proposed. It is based on a framewise conversion of frequency axis, fundamental frequency and global spectral and aperiodicity information using posterior probability as weighting function for calculating mapping function. The proposed method is an implementation of a concept called "speech texture mapping" that was proposed by one of the authors. The key idea which yields advantages of the proposed method is that the role and the relevant mapping function of detailed structure (refer as "texture") and global structure (refer as "framework") are different from each other. This clear distinction of "texture" and "framework" enabled a high-quality voice conversion requiring only a very small amount of training data. This distinction also provides a way to alleviate degradations due to "averaging" or "learning" processes which are indispensable in conventional voice conversion methods. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | STRAIGHT / MFCC / Bayes' method / voice texture / speaker conversion |
| Paper # | SP2006-162 |
| Date of Issue |
| Conference Information | |
| Committee | SP |
|---|---|
| Conference Date | 2007/3/19(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) | Speaker conversion system based on vowels An implementation of voice texture mapping |
| Sub Title (in English) | |
| Keyword(1) | STRAIGHT |
| Keyword(2) | MFCC |
| Keyword(3) | Bayes' method |
| Keyword(4) | voice texture |
| Keyword(5) | speaker conversion |
| 1st Author's Name | Toru TAKAHASHI |
| 1st Author's Affiliation | Wakayama University() |
| 2nd Author's Name | Masanori MORISE |
| 2nd Author's Affiliation | Wakayama University |
| 3rd Author's Name | Ryuichi NISIMURA |
| 3rd Author's Affiliation | Wakayama University |
| 4th Author's Name | Toshio IRINO |
| 4th Author's Affiliation | Wakayama University |
| 5th Author's Name | Hideki BANNO |
| 5th Author's Affiliation | Meijo University |
| 6th Author's Name | Hideki KAWAHARA |
| 6th Author's Affiliation | Wakayama University |
| Date | 2007/3/19 |
| Paper # | SP2006-162 |
| Volume (vol) | vol.106 |
| Number (no) | 613 |
| Page | pp.pp.- |
| #Pages | 6 |
| Date of Issue | |