TokSing: Singing Voice Synthesis based on Discrete Tokens

TL;DR

TokSing is a novel singing voice synthesis system that uses discrete tokens for efficient and expressive melody generation, overcoming challenges in discretization and enhancing performance over traditional spectrogram methods.

Contribution

The paper introduces TokSing, a discrete-token based SVS system with a melody enhancement strategy, improving melody expression and efficiency over existing methods.

Findings

Outperforms Mel spectrogram baselines in quality.

Offers lower storage and faster convergence.

Effectively enhances melody expression in discrete token space.

Abstract

Recent advancements in speech synthesis witness significant benefits by leveraging discrete tokens extracted from self-supervised learning (SSL) models. Discrete tokens offer higher storage efficiency and greater operability in intermediate representations compared to traditional continuous Mel spectrograms. However, when it comes to singing voice synthesis(SVS), achieving higher levels of melody expression poses a great challenge for utilizing discrete tokens. In this paper, we introduce TokSing, a discrete-based SVS system equipped with a token formulator that offers flexible token blendings. We observe a melody degradation during discretization, prompting us to integrate a melody signal with the discrete token and incorporate a specially-designed melody enhancement strategy in the musical encoder. Extensive experiments demonstrate that our TokSing achieves better performance against the Mel spectrogram baselines while offering advantages in intermediate representation space cost and convergence speed.