Presentation | 2013-01-31 A Study on Style Control Based on Multiple-Regression HSMM for Synthesizing Singing Voices with Various Expressivity Takashi NOSE, Misa KANEMOTO, Tomoki KORIYAMA, Takao KOBAYASHI, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | This paper proposes a style control technique based on multiple regression HSMM (MRHSMM) for changing styles and their intensities appearing in synthetic singing voices. In the proposed technique, styles and their intensities are represented by low-dimensional vectors called style vectors and are modeled by an assumption that mean parameters of acoustic models are given as multiple regressions of the style vectors. When synthesizing speech, we can weaken or emphasize the intensity of each style by setting a desired style vector. In addition, the idea of pitch adaptive training is introduced into the MRHSMM to improve the modeling accuracy of F0 associated with musical notes. The novel vibrato modeling technique is also presented to extract vibrato parameters from singing voices that sometimes have unclear vibrato expressions. Subjective evaluations show that we can intuitively contorol styles and their intensities while keeping naturalness of synthetic speech. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | HMM-based singing voice synthesis / HMM-based speech synthesis / style control / multiple-regression HSMM / pitch adaptive training / vibrato modeling |
Paper # | SP2012-111 |
Date of Issue |
Conference Information | |
Committee | SP |
---|---|
Conference Date | 2013/1/23(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) | A Study on Style Control Based on Multiple-Regression HSMM for Synthesizing Singing Voices with Various Expressivity |
Sub Title (in English) | |
Keyword(1) | HMM-based singing voice synthesis |
Keyword(2) | HMM-based speech synthesis |
Keyword(3) | style control |
Keyword(4) | multiple-regression HSMM |
Keyword(5) | pitch adaptive training |
Keyword(6) | vibrato modeling |
1st Author's Name | Takashi NOSE |
1st Author's Affiliation | Interdisciplinary Graduate School of Science and Engineering, Tokyo Institute of Technology() |
2nd Author's Name | Misa KANEMOTO |
2nd Author's Affiliation | Interdisciplinary Graduate School of Science and Engineering, Tokyo Institute of Technology |
3rd Author's Name | Tomoki KORIYAMA |
3rd Author's Affiliation | Interdisciplinary Graduate School of Science and Engineering, Tokyo Institute of Technology / |
4th Author's Name | Takao KOBAYASHI |
4th Author's Affiliation | |
Date | 2013-01-31 |
Paper # | SP2012-111 |
Volume (vol) | vol.112 |
Number (no) | 422 |
Page | pp.pp.- |
#Pages | 6 |
Date of Issue |