# Singing voice conversion with non-parallel data

## TL;DR

This paper introduces a novel non-parallel data singing voice conversion method that uses a phonetic posterior feature and RNN with DBLSTM to convert singing voices between different singers, demonstrating effective subjective results.

## Contribution

First to utilize non-parallel data for singing voice conversion, combining ASR-based features with RNN modeling for accurate voice transformation.

## Key findings

- Effective voice conversion demonstrated through subjective evaluation.
- First application of non-parallel data approach in singing voice conversion.
- Uses phonetic posterior features and RNN with DBLSTM for modeling.

## Abstract

Singing voice conversion is a task to convert a song sang by a source singer to the voice of a target singer. In this paper, we propose using a parallel data free, many-to-one voice conversion technique on singing voices. A phonetic posterior feature is first generated by decoding singing voices through a robust Automatic Speech Recognition Engine (ASR). Then, a trained Recurrent Neural Network (RNN) with a Deep Bidirectional Long Short Term Memory (DBLSTM) structure is used to model the mapping from person-independent content to the acoustic features of the target person. F0 and aperiodic are obtained through the original singing voice, and used with acoustic features to reconstruct the target singing voice through a vocoder. In the obtained singing voice, the targeted and sourced singers sound similar. To our knowledge, this is the first study that uses non parallel data to train a singing voice conversion system. Subjective evaluations demonstrate that the proposed method effectively converts singing voices.