Saved in:
Bibliographic Details
Main Authors: Shan, Siyuan, Li, Yang, Banerjee, Amartya, Oliva, Junier B.
Format: Preprint
Published: 2023
Subjects:
Online Access:https://arxiv.org/abs/2308.06382
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929195384307712
author Shan, Siyuan
Li, Yang
Banerjee, Amartya
Oliva, Junier B.
author_facet Shan, Siyuan
Li, Yang
Banerjee, Amartya
Oliva, Junier B.
contents Voice conversion (VC) aims at altering a person's voice to make it sound similar to the voice of another person while preserving linguistic content. Existing methods suffer from a dilemma between content intelligibility and speaker similarity; i.e., methods with higher intelligibility usually have a lower speaker similarity, while methods with higher speaker similarity usually require plenty of target speaker voice data to achieve high intelligibility. In this work, we propose a novel method \textit{Phoneme Hallucinator} that achieves the best of both worlds. Phoneme Hallucinator is a one-shot VC model; it adopts a novel model to hallucinate diversified and high-fidelity target speaker phonemes based just on a short target speaker voice (e.g. 3 seconds). The hallucinated phonemes are then exploited to perform neighbor-based voice conversion. Our model is a text-free, any-to-any VC model that requires no text annotations and supports conversion to any unseen speaker. Objective and subjective evaluations show that \textit{Phoneme Hallucinator} outperforms existing VC methods for both intelligibility and speaker similarity.
format Preprint
id arxiv_https___arxiv_org_abs_2308_06382
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Phoneme Hallucinator: One-shot Voice Conversion via Set Expansion
Shan, Siyuan
Li, Yang
Banerjee, Amartya
Oliva, Junier B.
Sound
Machine Learning
Audio and Speech Processing
Voice conversion (VC) aims at altering a person's voice to make it sound similar to the voice of another person while preserving linguistic content. Existing methods suffer from a dilemma between content intelligibility and speaker similarity; i.e., methods with higher intelligibility usually have a lower speaker similarity, while methods with higher speaker similarity usually require plenty of target speaker voice data to achieve high intelligibility. In this work, we propose a novel method \textit{Phoneme Hallucinator} that achieves the best of both worlds. Phoneme Hallucinator is a one-shot VC model; it adopts a novel model to hallucinate diversified and high-fidelity target speaker phonemes based just on a short target speaker voice (e.g. 3 seconds). The hallucinated phonemes are then exploited to perform neighbor-based voice conversion. Our model is a text-free, any-to-any VC model that requires no text annotations and supports conversion to any unseen speaker. Objective and subjective evaluations show that \textit{Phoneme Hallucinator} outperforms existing VC methods for both intelligibility and speaker similarity.
title Phoneme Hallucinator: One-shot Voice Conversion via Set Expansion
topic Sound
Machine Learning
Audio and Speech Processing
url https://arxiv.org/abs/2308.06382