Presentation | 2014/12/8 朗読音声-歌声音声の特徴量変換と話者適応を用いた歌詞認識の性能向上の検討(音声認識,第16回音声言語シンポジウム) , |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | As a first step, we consider Japanese lyrics recognition in monophonic singing that contains no musical instruments. To express singing well, we attempt to use an n-gram language model using a lyrics corpus, singing-adapted GMM-HMM-based acoustic models and plural pronunciation lexicons for vowel-lengthening. We attempted to adapt the read-speech AMs to sung-speech AMs using two approaches. One is MAP adaptation and the other is neural network-based feature transformation. For adapting to singing, we use 40 pieces of music sung by 40 male singers. For adapting to speaker, we use a piece of music sung by a male singer who is the same speaker as a singer of a test data. To deal with the property of singing offten involving lengthening the duration of each vowel, we augment the pronunciation variations. Evaluation is performed on a test set that contains 7 pieces of commercial music sung by 7 male singers. As a result of experiments, our system showed syllable accuracy of 46.1% (phoneme accuracy of 59.0%) and word accuracy of 25.9% in male monophonic Japanese singing. This result showed higher accuracy than a conventional system based on the newspaper LM and the read-speech AM. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | lyrics recognition / read-sung speech transformation / MAP adaptation / vowel-lengthening |
Paper # | Vol.2014-SLP-104 No.2 |
Date of Issue |
Conference Information | |
Committee | SP |
---|---|
Conference Date | 2014/12/8(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) | |
Sub Title (in English) | |
Keyword(1) | lyrics recognition |
Keyword(2) | read-sung speech transformation |
Keyword(3) | MAP adaptation |
Keyword(4) | vowel-lengthening |
1st Author's Name | |
1st Author's Affiliation | Toyohashi Uniersity of Technology() |
Date | 2014/12/8 |
Paper # | Vol.2014-SLP-104 No.2 |
Volume (vol) | vol.114 |
Number (no) | 365 |
Page | pp.pp.- |
#Pages | 6 |
Date of Issue |