Presentation | 2014-02-28 Prediction of pronunciation distances based on structural representation for clustering World Englishes Shun KASAHARA, Nobuaki MINEMATSU, Han-Ping SHEN, Takehiko MAKINO, Daisuke SAITO, Keikichi HIROSE, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | The term of World Englishes is often used to indicate the current state of English as international language. It claims that English does not have the standard pronunciation and that every country, region, and even individual uses different pronunciations. From the viewpoint of World Englishes, it will be much more important to let each speaker know how his/her pronunciation is located in the diversity of World Englishes pronunciations, not how his/her pronunciation is incorrect compared to native pronunciations. This study tries to predict inter-speaker pronunciation distances only by speech analysis to examine the possibility of individual-basis pronunciation clustering of World Englishes. Speech features are often altered by non-linguistic factors such as age and gender differences. Considering this, the pronunciation structure, known as speaker-invariant feature, and support vector regression were applied for prediction. In the experiments, two conditions of a speaker-pair-open mode and a speaker-open mode were examined for training and testing the SVR. As a result, although a striking performance was obtained in the speaker-pair-open mode, only insufficient performances were found in the speaker-open mode. To predict pronunciation distances between unknown speakers, a further investigation is required. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | World Englishes / pronunciation clustering / structural representation / support vector regression / speaker-pair-open / speaker-open |
Paper # | SP2013-109 |
Date of Issue |
Conference Information | |
Committee | SP |
---|---|
Conference Date | 2014/2/21(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) | Prediction of pronunciation distances based on structural representation for clustering World Englishes |
Sub Title (in English) | |
Keyword(1) | World Englishes |
Keyword(2) | pronunciation clustering |
Keyword(3) | structural representation |
Keyword(4) | support vector regression |
Keyword(5) | speaker-pair-open |
Keyword(6) | speaker-open |
1st Author's Name | Shun KASAHARA |
1st Author's Affiliation | The university of Tokyo() |
2nd Author's Name | Nobuaki MINEMATSU |
2nd Author's Affiliation | The university of Tokyo |
3rd Author's Name | Han-Ping SHEN |
3rd Author's Affiliation | National Cheng Kung University |
4th Author's Name | Takehiko MAKINO |
4th Author's Affiliation | Chuo University |
5th Author's Name | Daisuke SAITO |
5th Author's Affiliation | The university of Tokyo |
6th Author's Name | Keikichi HIROSE |
6th Author's Affiliation | The university of Tokyo |
Date | 2014-02-28 |
Paper # | SP2013-109 |
Volume (vol) | vol.113 |
Number (no) | 452 |
Page | pp.pp.- |
#Pages | 6 |
Date of Issue |