Saved in:
Bibliographic Details
Main Authors: Kamble, Anand, Tathe, Aniket, Kumbharkar, Suyash, Bhandare, Atharva, Mitra, Anirban C.
Format: Preprint
Published: 2023
Subjects:
Online Access:https://arxiv.org/abs/2311.14836
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929205532426240
author Kamble, Anand
Tathe, Aniket
Kumbharkar, Suyash
Bhandare, Atharva
Mitra, Anirban C.
author_facet Kamble, Anand
Tathe, Aniket
Kumbharkar, Suyash
Bhandare, Atharva
Mitra, Anirban C.
contents This paper proposes two innovative methodologies to construct customized Common Voice datasets for low-resource languages like Hindi. The first methodology leverages Bark, a transformer-based text-to-audio model developed by Suno, and incorporates Meta's enCodec and a pre-trained HuBert model to enhance Bark's performance. The second methodology employs Retrieval-Based Voice Conversion (RVC) and uses the Ozen toolkit for data preparation. Both methodologies contribute to the advancement of ASR technology and offer valuable insights into addressing the challenges of constructing customized Common Voice datasets for under-resourced languages. Furthermore, they provide a pathway to achieving high-quality, personalized voice generation for a range of applications.
format Preprint
id arxiv_https___arxiv_org_abs_2311_14836
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Custom Data Augmentation for low resource ASR using Bark and Retrieval-Based Voice Conversion
Kamble, Anand
Tathe, Aniket
Kumbharkar, Suyash
Bhandare, Atharva
Mitra, Anirban C.
Sound
Computation and Language
Audio and Speech Processing
This paper proposes two innovative methodologies to construct customized Common Voice datasets for low-resource languages like Hindi. The first methodology leverages Bark, a transformer-based text-to-audio model developed by Suno, and incorporates Meta's enCodec and a pre-trained HuBert model to enhance Bark's performance. The second methodology employs Retrieval-Based Voice Conversion (RVC) and uses the Ozen toolkit for data preparation. Both methodologies contribute to the advancement of ASR technology and offer valuable insights into addressing the challenges of constructing customized Common Voice datasets for under-resourced languages. Furthermore, they provide a pathway to achieving high-quality, personalized voice generation for a range of applications.
title Custom Data Augmentation for low resource ASR using Bark and Retrieval-Based Voice Conversion
topic Sound
Computation and Language
Audio and Speech Processing
url https://arxiv.org/abs/2311.14836