Sparse representation of phonetic features for voice conversion with and without parallel data

Berrak Sisman; Haizhou Li; Kay Chen Tan

doi:10.1109/asru.2017.8269002

Sparse representation of phonetic features for voice conversion with and without parallel data

Berrak Sisman, Haizhou Li, Kay Chen Tan

Department of Computer Science

Research output: Chapters, Conference Papers, Creative and Literary Works › RGC 32 - Refereed conference paper (with host publication) › peer-review

37 Citations (Scopus)

Abstract

This paper presents a voice conversion framework that uses phonetic information in an exemplar-based voice conversion approach. The proposed idea is motivated by the fact that phone-dependent exemplars lead to better estimation of activation matrix, therefore, possibly better conversion. We propose to use the phone segmentation results from automatic speech recognition (ASR) to construct a sub-dictionary for each phone. The proposed framework can work with or without parallel training data. With parallel training data, we found that phonetic sub-dictionary outperforms the state-of-the-art baseline in objective and subjective evaluations. Without parallel training data, we use Phonetic PosteriorGrams (PPGs) as the speaker-independent exemplars in the phonetic sub-dictionary to serve as a bridge between speakers. We report that such technique achieves a competitive performance without the need of parallel training data.

Original language	English
Title of host publication	2017 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU) - Proceedings
Publisher	IEEE
Pages	677-684
ISBN (Electronic)	978-1-5090-4788-8
DOIs	https://doi.org/10.1109/asru.2017.8269002
Publication status	Published - Dec 2017
Event	2017 IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2017 - Okinawa, Japan Duration: 16 Dec 2017 → 20 Dec 2017

Conference

Conference	2017 IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2017
Place	Japan
City	Okinawa
Period	16/12/17 → 20/12/17

Bibliographical note

Full text of this publication does not contain sufficient affiliation information. With consent from the author(s) concerned, the Research Unit(s) information for this record is based on the existing academic department affiliation of the author(s).

Research Keywords

phonetic exemplars
Phonetic PosteriorGrams
sparse representation
Voice conversion

Access Output

10.1109/asru.2017.8269002

Other Access Options

Check CityUHK Library

Permanent Link

Right click to copy link

Cite this

@inproceedings{fd1450faadfe4f7d8616cf594d615b86,

title = "Sparse representation of phonetic features for voice conversion with and without parallel data",

abstract = "This paper presents a voice conversion framework that uses phonetic information in an exemplar-based voice conversion approach. The proposed idea is motivated by the fact that phone-dependent exemplars lead to better estimation of activation matrix, therefore, possibly better conversion. We propose to use the phone segmentation results from automatic speech recognition (ASR) to construct a sub-dictionary for each phone. The proposed framework can work with or without parallel training data. With parallel training data, we found that phonetic sub-dictionary outperforms the state-of-the-art baseline in objective and subjective evaluations. Without parallel training data, we use Phonetic PosteriorGrams (PPGs) as the speaker-independent exemplars in the phonetic sub-dictionary to serve as a bridge between speakers. We report that such technique achieves a competitive performance without the need of parallel training data.",

keywords = "phonetic exemplars, Phonetic PosteriorGrams, sparse representation, Voice conversion",

author = "Berrak Sisman and Haizhou Li and Tan, {Kay Chen}",

note = "Full text of this publication does not contain sufficient affiliation information. With consent from the author(s) concerned, the Research Unit(s) information for this record is based on the existing academic department affiliation of the author(s).; 2017 IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2017 ; Conference date: 16-12-2017 Through 20-12-2017",

year = "2017",

month = dec,

doi = "10.1109/asru.2017.8269002",

language = "English",

pages = "677--684",

booktitle = "2017 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU) - Proceedings",

publisher = "IEEE",

address = "United States",

}

Sisman, B, Li, H & Tan, KC 2017, Sparse representation of phonetic features for voice conversion with and without parallel data. in 2017 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU) - Proceedings. IEEE, pp. 677-684, 2017 IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2017, Okinawa, Japan, 16/12/17. https://doi.org/10.1109/asru.2017.8269002

Sparse representation of phonetic features for voice conversion with and without parallel data. / Sisman, Berrak; Li, Haizhou; Tan, Kay Chen.
2017 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU) - Proceedings. IEEE, 2017. p. 677-684.

Research output: Chapters, Conference Papers, Creative and Literary Works › RGC 32 - Refereed conference paper (with host publication) › peer-review

TY - GEN

T1 - Sparse representation of phonetic features for voice conversion with and without parallel data

AU - Sisman, Berrak

AU - Li, Haizhou

AU - Tan, Kay Chen

N1 - Full text of this publication does not contain sufficient affiliation information. With consent from the author(s) concerned, the Research Unit(s) information for this record is based on the existing academic department affiliation of the author(s).

PY - 2017/12

Y1 - 2017/12

N2 - This paper presents a voice conversion framework that uses phonetic information in an exemplar-based voice conversion approach. The proposed idea is motivated by the fact that phone-dependent exemplars lead to better estimation of activation matrix, therefore, possibly better conversion. We propose to use the phone segmentation results from automatic speech recognition (ASR) to construct a sub-dictionary for each phone. The proposed framework can work with or without parallel training data. With parallel training data, we found that phonetic sub-dictionary outperforms the state-of-the-art baseline in objective and subjective evaluations. Without parallel training data, we use Phonetic PosteriorGrams (PPGs) as the speaker-independent exemplars in the phonetic sub-dictionary to serve as a bridge between speakers. We report that such technique achieves a competitive performance without the need of parallel training data.

AB - This paper presents a voice conversion framework that uses phonetic information in an exemplar-based voice conversion approach. The proposed idea is motivated by the fact that phone-dependent exemplars lead to better estimation of activation matrix, therefore, possibly better conversion. We propose to use the phone segmentation results from automatic speech recognition (ASR) to construct a sub-dictionary for each phone. The proposed framework can work with or without parallel training data. With parallel training data, we found that phonetic sub-dictionary outperforms the state-of-the-art baseline in objective and subjective evaluations. Without parallel training data, we use Phonetic PosteriorGrams (PPGs) as the speaker-independent exemplars in the phonetic sub-dictionary to serve as a bridge between speakers. We report that such technique achieves a competitive performance without the need of parallel training data.

KW - phonetic exemplars

KW - Phonetic PosteriorGrams

KW - sparse representation

KW - Voice conversion

UR - http://www.scopus.com/inward/record.url?scp=85046631146&partnerID=8YFLogxK

UR - https://www.scopus.com/record/pubmetrics.uri?eid=2-s2.0-85046631146&origin=recordpage

U2 - 10.1109/asru.2017.8269002

DO - 10.1109/asru.2017.8269002

M3 - RGC 32 - Refereed conference paper (with host publication)

SP - 677

EP - 684

BT - 2017 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU) - Proceedings

PB - IEEE

T2 - 2017 IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2017

Y2 - 16 December 2017 through 20 December 2017

ER -

Sparse representation of phonetic features for voice conversion with and without parallel data

Abstract

Conference

Bibliographical note

Research Keywords

Access Output

Other Access Options

Permanent Link

Other Files and Links

Fingerprint

Cite this