Towards Naturalistic Voice Conversion: NaturalVoices Dataset with an Automatic Processing Pipeline

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Salman, Ali N., Du, Zongyang, Chandra, Shreeram Suresh, Ulgen, Ismail Rasim, Busso, Carlos, Sisman, Berrak
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909218677719040
author Salman, Ali N.
Du, Zongyang
Chandra, Shreeram Suresh
Ulgen, Ismail Rasim
Busso, Carlos
Sisman, Berrak
author_facet Salman, Ali N.
Du, Zongyang
Chandra, Shreeram Suresh
Ulgen, Ismail Rasim
Busso, Carlos
Sisman, Berrak
contents Voice conversion (VC) research traditionally depends on scripted or acted speech, which lacks the natural spontaneity of real-life conversations. While natural speech data is limited for VC, our study focuses on filling in this gap. We introduce a novel data-sourcing pipeline that makes the release of a natural speech dataset for VC, named NaturalVoices. The pipeline extracts rich information in speech such as emotion and signal-to-noise ratio (SNR) from raw podcast data, utilizing recent deep learning methods and providing flexibility and ease of use. NaturalVoices marks a large-scale, spontaneous, expressive, and emotional speech dataset, comprising over 3,800 hours speech sourced from the original podcasts in the MSP-Podcast dataset. Objective and subjective evaluations demonstrate the effectiveness of using our pipeline for providing natural and expressive data for VC, suggesting the potential of NaturalVoices for broader speech generation tasks.
format Preprint
id arxiv_https___arxiv_org_abs_2406_04494
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Towards Naturalistic Voice Conversion: NaturalVoices Dataset with an Automatic Processing Pipeline
Salman, Ali N.
Du, Zongyang
Chandra, Shreeram Suresh
Ulgen, Ismail Rasim
Busso, Carlos
Sisman, Berrak
Audio and Speech Processing
Voice conversion (VC) research traditionally depends on scripted or acted speech, which lacks the natural spontaneity of real-life conversations. While natural speech data is limited for VC, our study focuses on filling in this gap. We introduce a novel data-sourcing pipeline that makes the release of a natural speech dataset for VC, named NaturalVoices. The pipeline extracts rich information in speech such as emotion and signal-to-noise ratio (SNR) from raw podcast data, utilizing recent deep learning methods and providing flexibility and ease of use. NaturalVoices marks a large-scale, spontaneous, expressive, and emotional speech dataset, comprising over 3,800 hours speech sourced from the original podcasts in the MSP-Podcast dataset. Objective and subjective evaluations demonstrate the effectiveness of using our pipeline for providing natural and expressive data for VC, suggesting the potential of NaturalVoices for broader speech generation tasks.
title Towards Naturalistic Voice Conversion: NaturalVoices Dataset with an Automatic Processing Pipeline
topic Audio and Speech Processing
url https://arxiv.org/abs/2406.04494