An Unsupervised Method to Select a Speaker Subset from Large Multi-Speaker Speech Synthesis Datasets

被引：2

作者：

Gallegos, Pilar Oplustil ^{[1
]}

Williams, Jennifer ^{[1
]}

Rownicka, Joanna ^{[1
]}

King, Simon ^{[1
]}

机构：

[1] Univ Edinburgh, Ctr Speech Technol Res, Edinburgh, Midlothian, Scotland

来源：

INTERSPEECH 2020 | 2020年

基金：

英国工程与自然科学研究理事会;

关键词：

speech synthesis; data; clustering; speaker representation; sequence-to-sequence models; multi-speaker;

D O I：

10.21437/Interspeech.2020-2567

中图分类号：

R36 [病理学]; R76 [耳鼻咽喉科学];

学科分类号：

100104 ; 100213 ;

摘要：

Large multi-speaker datasets for TTS typically contain diverse speakers, recording conditions, styles and quality of data. Although one might generally presume that more data is better, in this paper we show that a model trained on a carefully-chosen subset of speakers from LibriTTS provides significantly better quality synthetic speech than a model trained on a larger set. We propose an unsupervised methodology to find this subset by clustering per-speaker acoustic representations.

引用

页码：1758 / 1762

页数：5

共 50 条

[1] TOWARDS MULTI-SPEAKER UNSUPERVISED SPEECH PATTERN DISCOVERY
Zhang, Yaodong
Glass, James R.
[J]. 2010 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING, 2010, : 4366 - 4369
[2] Unsupervised Speaker and Expression Factorization for Multi-Speaker Expressive Synthesis of Ebooks
Chen, Langzhou
Braunschweiler, Norbert
[J]. 14TH ANNUAL CONFERENCE OF THE INTERNATIONAL SPEECH COMMUNICATION ASSOCIATION (INTERSPEECH 2013), VOLS 1-5, 2013, : 1041 - 1045
[3] Unsupervised Discovery of Phoneme Boundaries in Multi-Speaker Continuous Speech
Armstrong, Tom
Antetomaso, Stephanie
[J]. 2011 IEEE INTERNATIONAL CONFERENCE ON DEVELOPMENT AND LEARNING (ICDL), 2011,
[4] Speaker Clustering with Penalty Distance for Speaker Verification with Multi-Speaker Speech
Das, Rohan Kumar
Yang, Jichen
Li, Haizhou
[J]. 2019 ASIA-PACIFIC SIGNAL AND INFORMATION PROCESSING ASSOCIATION ANNUAL SUMMIT AND CONFERENCE (APSIPA ASC), 2019, : 1630 - 1635
[5] PHONEME DEPENDENT SPEAKER EMBEDDING AND MODEL FACTORIZATION FOR MULTI-SPEAKER SPEECH SYNTHESIS AND ADAPTATION
Fu, Ruibo
Tao, Jianhua
Wen, Zhengqi
Zheng, Yibin
[J]. 2019 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP), 2019, : 6930 - 6934
[6] Speech Rhythm-Based Speaker Embeddings Extraction from Phonemes and Phoneme Duration for Multi-Speaker Speech Synthesis
Fujita, Kenichi
Ando, Atsushi
Ijima, Yusuke
[J]. IEICE TRANSACTIONS ON INFORMATION AND SYSTEMS, 2024, E107D (01) : 93 - 104
[7] Deep Gaussian process based multi-speaker speech synthesis with latent speaker representation
Mitsui, Kentaro
Koriyama, Tomoki
Saruwatari, Hiroshi
[J]. SPEECH COMMUNICATION, 2021, 132 : 132 - 145
[8] DNN based multi-speaker speech synthesis with temporal auxiliary speaker ID embedding
Lee, Junmo
Song, Kwangsub
Noh, Kyoungjin
Park, Tae-Jun
Chang, Joon-Hyuk
[J]. 2019 INTERNATIONAL CONFERENCE ON ELECTRONICS, INFORMATION, AND COMMUNICATION (ICEIC), 2019, : 61 - 64
[9] Phoneme Duration Modeling Using Speech Rhythm-Based Speaker Embeddings for Multi-Speaker Speech Synthesis
Fujita, Kenichi
Ando, Atsushi
Ijima, Yusuke
[J]. INTERSPEECH 2021, 2021, : 3141 - 3145
[10] Automatic speaker clustering from multi-speaker utterances
McLaughlin, J
Reynolds, D
Singer, E
O'Leary, GC
[J]. ICASSP '99: 1999 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING, PROCEEDINGS VOLS I-VI, 1999, : 817 - 820

← 1 2 3 4 5 →