Presentation | 2007-08-03 Evaluation of Artificial Speech Conversion System Using a Laryngectomee's Data Keigo NAKAMURA, Tomoki TODA, Yoshitaka NAKAJIMA, Hiroshi SARUWATARI, Kiyohiro SHIKANO, |
---|---|
PDF Download Page | ![]() |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | We are investigating speech communication-aid methods for laryngectomees whose vocal folds were removed due to laryngeal cancer and so on. In our previous work, a speaking-aid system was proposed. In this sytem, speech using small sound source would be converted into much natural voice by statistical voice conversion. The system was also evaluated using artificial speech of a normal speaker. In this paper, 170 sentences of newspaper read speech are recorded by one laryngectomee using two different small sound source units. Moreover, speech utterances converted by the proposed system are evaluated by objective and subjective experiments. Rate of speech and location of pauses differ remarkably in original and target speech. This is one of the reasons for the degradation of speech quality. To reveal influences of these factors on voice conversion, voice conversion experiment is conducted using target speech in a speech corpus or other target speech rerecorded to source speech. Moreover, another voice conversion experiment performed to compare the speech quality of the laryngectomee's converted speech and speech using an electrolarynx to verify the performance of our proposed system. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | Laryngectomees / Small Sound Source / Electrolarynx / Statistical Voice Conversion |
Paper # | WIT2007-21 |
Date of Issue |
Conference Information | |
Committee | WIT |
---|---|
Conference Date | 2007/7/27(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 | Well-being Information Technology(WIT) |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | Evaluation of Artificial Speech Conversion System Using a Laryngectomee's Data |
Sub Title (in English) | |
Keyword(1) | Laryngectomees |
Keyword(2) | Small Sound Source |
Keyword(3) | Electrolarynx |
Keyword(4) | Statistical Voice Conversion |
1st Author's Name | Keigo NAKAMURA |
1st Author's Affiliation | Graduate School of Information Science, Nara Institute of Science and Technology() |
2nd Author's Name | Tomoki TODA |
2nd Author's Affiliation | Graduate School of Information Science, Nara Institute of Science and Technology |
3rd Author's Name | Yoshitaka NAKAJIMA |
3rd Author's Affiliation | Graduate School of Information Science, Nara Institute of Science and Technology |
4th Author's Name | Hiroshi SARUWATARI |
4th Author's Affiliation | Graduate School of Information Science, Nara Institute of Science and Technology |
5th Author's Name | Kiyohiro SHIKANO |
5th Author's Affiliation | Graduate School of Information Science, Nara Institute of Science and Technology |
Date | 2007-08-03 |
Paper # | WIT2007-21 |
Volume (vol) | vol.107 |
Number (no) | 179 |
Page | pp.pp.- |
#Pages | 6 |
Date of Issue |