Multilingual TTS Accent Impressions for Accented ASR

No Thumbnail Available

Access rights

openAccess

URL

Journal Title

Journal ISSN

Volume Title

A4 Artikkeli konferenssijulkaisussa

Date

2023

Major/Subject

Mcode

Degree programme

Language

en

Pages

11
317-327

Series

Text, Speech, and Dialogue - 26th International Conference, TSD 2023, Proceedings, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Volume 14102 LNAI

Abstract

Automatic Speech Recognition (ASR) for high-resource languages like English is often considered a solved problem. However, most high-resource ASR systems favor socioeconomically advantaged dialects. In the case of English, this leaves behind many L2 speakers and speakers of low-resource accents (a majority of English speakers). One way to mitigate this is to fine-tune a pre-trained English ASR model for a desired low-resource accent. However, collecting transcribed accented audio is costly and time-consuming. In this work, we present a method to produce synthetic L2-English speech via pre-trained text-to-speech (TTS) in an L1 language (target accent). This can be produced at a much larger scale and lower cost than authentic speech collection. We present initial experiments applying this augmentation method. Our results suggest that success of TTS augmentation relies on access to more than one hour of authentic training data and a diversity of target-domain prompts for speech synthesis.

Description

Publisher Copyright: © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Keywords

accented speech recognition, data augmentation, low-resource speech technologies, speech synthesis

Other note

Citation

Karakasidis, G, Robinson, N, Getman, Y, Ogayo, A, Al-Ghezi, R, Ayasi, A, Watanabe, S, Mortensen, D R & Kurimo, M 2023, Multilingual TTS Accent Impressions for Accented ASR . in K Ekštein, F Pártl & M Konopík (eds), Text, Speech, and Dialogue - 26th International Conference, TSD 2023, Proceedings . Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 14102 LNAI, Springer, pp. 317-327, International Conference on Text, Speech, and Dialogue, Pilsen, Czech Republic, 04/09/2023 . https://doi.org/10.1007/978-3-031-40498-6_28