Indonesian Voice Cloning Text-to-Speech System With Vall-E-Based Model and Speech Enhancement

In recent years, Text-to-Speech (TTS) technology has advanced, with research focusing on multi-speaker TTS capable of voice cloning. In 2023, Wang et al. introduced Vall-E, a Transformer-based neural codec language model, achieving state-of-the-art results in voice cloning. However, limited research...

Full description

Saved in:
Bibliographic Details
Main Authors: Hizkia Raditya Pratama Roosadi, Rizki Rivai Ginanjar, Dessi Puji Lestari
Format: Article
Language:English
Published: IEEE 2024-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10806715/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850242842648117248
author Hizkia Raditya Pratama Roosadi
Rizki Rivai Ginanjar
Dessi Puji Lestari
author_facet Hizkia Raditya Pratama Roosadi
Rizki Rivai Ginanjar
Dessi Puji Lestari
author_sort Hizkia Raditya Pratama Roosadi
collection DOAJ
description In recent years, Text-to-Speech (TTS) technology has advanced, with research focusing on multi-speaker TTS capable of voice cloning. In 2023, Wang et al. introduced Vall-E, a Transformer-based neural codec language model, achieving state-of-the-art results in voice cloning. However, limited research has applied such models to the Indonesian language, leaving room for improvement in speech synthesis. This paper explores the development a TTS system using Vall-E and explores enhancements of the speech synthesis. The dataset, comprising audio-transcript pairs, was sourced from previous Indonesian speech processing research. Data preparation involved converting audio into codec tokens and transcripts into phoneme tokens. Following Wang et al., a neural codec language model was built and trained using open-source tools. Additionally, this paper explores the integration VoiceFixer tool for speech enhancement. The inclusion of VoiceFixer improved the naturalness MOS score from 3.34 to 3.95, demonstrating its effectiveness in enhancing speech quality. Overall, the TTS system achieved a naturalness MOS score of 3.489 and a similarity MOS score of 3.521, with a WER of 19.71% and speaker embedding vector similarity visualization. These results indicate that the Vall-E model can produce Indonesian speech with high speaker similarity. The development also emphasizes the importance of factors like the number of speakers, data selection, processing components, modeling, and speech duration during training for synthesis quality.
format Article
id doaj-art-b24cc2d4fb7e4393b3615ca4c246f79a
institution OA Journals
issn 2169-3536
language English
publishDate 2024-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-b24cc2d4fb7e4393b3615ca4c246f79a2025-08-20T02:00:10ZengIEEEIEEE Access2169-35362024-01-011219313119314010.1109/ACCESS.2024.351987010806715Indonesian Voice Cloning Text-to-Speech System With Vall-E-Based Model and Speech EnhancementHizkia Raditya Pratama Roosadi0https://orcid.org/0009-0003-0010-3034Rizki Rivai Ginanjar1Dessi Puji Lestari2School of Electrical Engineering and Informatics, Institut Teknologi Bandung, Bandung, Jawa Barat, IndonesiaSpeech TTS and Paralinguistics Division, Prosa.ai, Bandung, Jawa Barat, IndonesiaSchool of Electrical Engineering and Informatics, Institut Teknologi Bandung, Bandung, Jawa Barat, IndonesiaIn recent years, Text-to-Speech (TTS) technology has advanced, with research focusing on multi-speaker TTS capable of voice cloning. In 2023, Wang et al. introduced Vall-E, a Transformer-based neural codec language model, achieving state-of-the-art results in voice cloning. However, limited research has applied such models to the Indonesian language, leaving room for improvement in speech synthesis. This paper explores the development a TTS system using Vall-E and explores enhancements of the speech synthesis. The dataset, comprising audio-transcript pairs, was sourced from previous Indonesian speech processing research. Data preparation involved converting audio into codec tokens and transcripts into phoneme tokens. Following Wang et al., a neural codec language model was built and trained using open-source tools. Additionally, this paper explores the integration VoiceFixer tool for speech enhancement. The inclusion of VoiceFixer improved the naturalness MOS score from 3.34 to 3.95, demonstrating its effectiveness in enhancing speech quality. Overall, the TTS system achieved a naturalness MOS score of 3.489 and a similarity MOS score of 3.521, with a WER of 19.71% and speaker embedding vector similarity visualization. These results indicate that the Vall-E model can produce Indonesian speech with high speaker similarity. The development also emphasizes the importance of factors like the number of speakers, data selection, processing components, modeling, and speech duration during training for synthesis quality.https://ieeexplore.ieee.org/document/10806715/Neural codec language modelspeech enhancementtransformertext-to-speechVall-Evoice cloning
spellingShingle Hizkia Raditya Pratama Roosadi
Rizki Rivai Ginanjar
Dessi Puji Lestari
Indonesian Voice Cloning Text-to-Speech System With Vall-E-Based Model and Speech Enhancement
IEEE Access
Neural codec language model
speech enhancement
transformer
text-to-speech
Vall-E
voice cloning
title Indonesian Voice Cloning Text-to-Speech System With Vall-E-Based Model and Speech Enhancement
title_full Indonesian Voice Cloning Text-to-Speech System With Vall-E-Based Model and Speech Enhancement
title_fullStr Indonesian Voice Cloning Text-to-Speech System With Vall-E-Based Model and Speech Enhancement
title_full_unstemmed Indonesian Voice Cloning Text-to-Speech System With Vall-E-Based Model and Speech Enhancement
title_short Indonesian Voice Cloning Text-to-Speech System With Vall-E-Based Model and Speech Enhancement
title_sort indonesian voice cloning text to speech system with vall e based model and speech enhancement
topic Neural codec language model
speech enhancement
transformer
text-to-speech
Vall-E
voice cloning
url https://ieeexplore.ieee.org/document/10806715/
work_keys_str_mv AT hizkiaradityapratamaroosadi indonesianvoicecloningtexttospeechsystemwithvallebasedmodelandspeechenhancement
AT rizkirivaiginanjar indonesianvoicecloningtexttospeechsystemwithvallebasedmodelandspeechenhancement
AT dessipujilestari indonesianvoicecloningtexttospeechsystemwithvallebasedmodelandspeechenhancement