Presentation | 2009-10-30 An Evaluation of Statistical Voice Conversion in Speaking-Aid Systems Using External Source Signals Keigo NAKAMURA, Tomoki TODA, Hiroshi SARUWATARI, Kiyohiro SHIKANO, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | This paper presents experimental evaluations of statistical voice conversion from artificial speech spoken by a laryngectomee using some external sound source units. The laryngectomee uses three kinds of the external sound source units; 1) a sound source unit that generates quite small source signals that cannot be heard by listeners; 2) a conventional electrolarynx that generates signals with monotone F_0; and 3) an electrolarynx using an air-pressure sensor that enables the laryngectomee to modulate the F_0 of the sound source signal using the air-pressure output from his/her tracheostoma. The generated artificial speech is detected with a headset microphone or Non-audible murmur microphone, and then, it is converted into whispered voice or normal speech uttered by a nonlaryngectomee. The experimental results demonstrate that 1) the use of the air-pressure sensor is effective to improve the voice conversion accuracy, 2) the modulated F_0 using an air-pressure sensor is less effective in F_0 estimation, but it doesn't cause any significant degradation of the voice conversion accuracy, and 3) voice conversion yields significant improvements in naturalness so that the converted speech is much more preferred to the original artificial speech by listeners. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | Laryngectomees / Electrolarynx / Air-pressure Sensor / Small Sound Source / Statistical Voice Conversion |
Paper # | SP2009-57,WIT2009-63 |
Date of Issue |
Conference Information | |
Committee | WIT |
---|---|
Conference Date | 2009/10/22(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) | An Evaluation of Statistical Voice Conversion in Speaking-Aid Systems Using External Source Signals |
Sub Title (in English) | |
Keyword(1) | Laryngectomees |
Keyword(2) | Electrolarynx |
Keyword(3) | Air-pressure Sensor |
Keyword(4) | Small Sound Source |
Keyword(5) | 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 | Hiroshi SARUWATARI |
3rd Author's Affiliation | Graduate School of Information Science, Nara Institute of Science and Technology |
4th Author's Name | Kiyohiro SHIKANO |
4th Author's Affiliation | Graduate School of Information Science, Nara Institute of Science and Technology |
Date | 2009-10-30 |
Paper # | SP2009-57,WIT2009-63 |
Volume (vol) | vol.109 |
Number (no) | 260 |
Page | pp.pp.- |
#Pages | 6 |
Date of Issue |