PPG-based singing voice conversion with adversarial representation learning

TL;DR

This paper introduces a novel end-to-end PPG-based singing voice conversion model that leverages adversarial and mel-regressive modules to enhance naturalness, melody, and voice similarity, demonstrating robustness to noise.

Contribution

It presents a new architecture with separate encoders and specialized modules for improved singing voice conversion performance.

Findings

Significant improvement in naturalness, melody, and voice similarity over baselines.

Robustness of the method to noisy source inputs.

Effective use of adversarial and mel-regressive modules in singing voice conversion.

Abstract

Singing voice conversion (SVC) aims to convert the voice of one singer to that of other singers while keeping the singing content and melody. On top of recent voice conversion works, we propose a novel model to steadily convert songs while keeping their naturalness and intonation. We build an end-to-end architecture, taking phonetic posteriorgrams (PPGs) as inputs and generating mel spectrograms. Specifically, we implement two separate encoders: one encodes PPGs as content, and the other compresses mel spectrograms to supply acoustic and musical information. To improve the performance on timbre and melody, an adversarial singer confusion module and a mel-regressive representation learning module are designed for the model. Objective and subjective experiments are conducted on our private Chinese singing corpus. Comparing with the baselines, our methods can significantly improve the conversion performance in terms of naturalness, melody, and voice similarity. Moreover, our PPG-based method is proved to be robust for noisy sources.