# Self-Supervised Representations for Singing Voice Conversion

arXiv: https://arxiv.org/abs/2303.12197 · HTML: https://mlsanity.com/paper/2303.12197/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 21 Mar 2023.

## Built on by (most cited first, 8 of 8)

- [GTSinger: A Global Multi-Technique Singing Corpus with Realistic Music Scores for All Singing Tasks](https://mlsanity.com/paper/2409.13832/) (2024)
- [SONICS: Synthetic Or Not -- Identifying Counterfeit Songs](https://mlsanity.com/paper/2408.14080/) (2024)
- [FSD: An Initial Chinese Dataset for Fake Song Detection](https://mlsanity.com/paper/2309.02232/) (2023)
- [LDM-SVC: Latent Diffusion Model Based Zero-Shot Any-to-Any Singing Voice Conversion with Singer Guidance](https://mlsanity.com/paper/2406.05325/) (2024)
- [REF-VC: Robust, Expressive and Fast Zero-Shot Voice Conversion with Diffusion Transformers](https://mlsanity.com/paper/2508.04996/) (2025)
- [A Comparative Study of Voice Conversion Models with Large-Scale Speech and Singing Data: The T13 Systems for the Singing Voice Conversion Challenge 2023](https://mlsanity.com/paper/2310.05203/) (2023)
- [Singing Voice Conversion with Accompaniment Using Self-Supervised Representation-Based Melody Features](https://mlsanity.com/paper/2502.04722/) (2025)
- [Are Music Foundation Models Better at Singing Voice Deepfake Detection? Far-Better Fuse them with Speech Foundation Models](https://mlsanity.com/paper/2409.14131/) (2024)

## Abstract (arXiv)

A singing voice conversion model converts a song in the voice of an arbitrary source singer to the voice of a target singer. Recently, methods that leverage self-supervised audio representations such as HuBERT and Wav2Vec 2.0 have helped further the state-of-the-art. Though these methods produce more natural and melodic singing outputs, they often rely on confusion and disentanglement losses to render the self-supervised representations speaker and pitch-invariant. In this paper, we circumvent disentanglement training and propose a new model that leverages ASR fine-tuned self-supervised representations as inputs to a HiFi-GAN neural vocoder for singing voice conversion. We experiment with different f0 encoding schemes and show that an f0 harmonic generation module that uses a parallel bank of transposed convolutions (PBTC) alongside ASR fine-tuned Wav2Vec 2.0 features results in the best singing voice conversion quality. Additionally, the model is capable of making a spoken voice sing. We also show that a simple f0 shifting scheme during inference helps retain singer identity and bolsters the performance of our singing voice conversion model. Our results are backed up by extensive MOS studies that compare different ablations and baselines.