YingMusic-Singer-Plus: Controllable Singing Voice Synthesis with Flexible Lyric Manipulation and Annotation-free Melody Guidance

TL;DR

YingMusic-Singer-Plus is a diffusion-based singing voice synthesis model that allows flexible lyric editing and melody control without manual alignment, outperforming existing methods in melody preservation and lyric adherence.

Contribution

It introduces a novel diffusion model for controllable singing synthesis with lyric manipulation and presents a new benchmark for evaluating melody-preserving lyric modifications.

Findings

YingMusic-Singer-Plus outperforms Vevo2 in melody preservation and lyric adherence.

The model supports lyric editing without manual alignment.

Code and benchmark are publicly available at the provided GitHub link.

Abstract

Regenerating singing voices with altered lyrics while preserving melody consistency remains challenging, as existing methods either offer limited controllability or require laborious manual alignment. We propose YingMusic-Singer-Plus, a fully diffusion-based model enabling melody-controllable singing voice synthesis with flexible lyric manipulation. The model takes three inputs: an optional timbre reference, a melody-providing singing clip, and modified lyrics, without manual alignment. Trained with curriculum learning and Group Relative Policy Optimization, YingMusic-Singer-Plus achieves stronger melody preservation and lyric adherence than Vevo2, the most comparable baseline supporting melody control without manual alignment. We also introduce LyricEditBench, the first benchmark for melody-preserving lyric modification evaluation. The code, weights, benchmark, and demos are publicly available at https://github.com/ASLP-lab/YingMusic-Singer-Plus.