Presentation | 2022-06-17 Representation and analytical normalization for vocal-tract-length transformation by group theory Atsushi Miyashita, Tomoki Toda, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | In automatic speech recognition, a recognition result should be invariant with respect to acoustic changes caused by different speakers. Vocal Tract Length Transformation (VTLT) is one of the typical transformations to reproduce those acoustic changes. In this report, we propose an alternative representation of VTLT based on all-pass-warping by using the group theory and an analytical approach to normalizing the acoustic changes caused by it. The proposed approach is capable of extracting the normalized feature independently on training data. An experimental result of a frame-wise phoneme classification task using TIMIT has demonstrated that the proposed method yields classification performance improvements for pseudo speakers whose vocal tract lengths are different from those of the speakers in the training data. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | Automatic speech recognition / speaker normalization / vocal tract length transformation / invariant feature / group theory |
Paper # | SP2022-11 |
Date of Issue | 2022-06-10 (SP) |
Conference Information | |
Committee | SP / IPSJ-MUS / IPSJ-SLP |
---|---|
Conference Date | 2022/6/17(2days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | Online |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | |
Chair | Tomoki Toda(Nagoya Univ.) |
Vice Chair | |
Secretary | (NTT) / (Univ. of Electro-Comm.) |
Assistant | Ryo Aihara(Mitsubishi Electric) / Daisuke Saito(Univ. of Tokyo) |
Paper Information | |
Registration To | Technical Committee on Speech / Special Interest Group on Music and Computer / Special Interest Group on Spoken Language Processing |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | Representation and analytical normalization for vocal-tract-length transformation by group theory |
Sub Title (in English) | |
Keyword(1) | Automatic speech recognition |
Keyword(2) | speaker normalization |
Keyword(3) | vocal tract length transformation |
Keyword(4) | invariant feature |
Keyword(5) | group theory |
1st Author's Name | Atsushi Miyashita |
1st Author's Affiliation | Nagoya University(Nagoya Univ) |
2nd Author's Name | Tomoki Toda |
2nd Author's Affiliation | Nagoya University(Nagoya Univ) |
Date | 2022-06-17 |
Paper # | SP2022-11 |
Volume (vol) | vol.122 |
Number (no) | SP-81 |
Page | pp.pp.41-46(SP), |
#Pages | 6 |
Date of Issue | 2022-06-10 (SP) |