High-Fidelity Music Vocoder using Neural Audio Codecs
Abstract
While neural vocoders have made significant progress in high-fidelity speech synthesis, their application on polyphonic music has remained underexplored. In this work, we propose DisCoder, a neural vocoder that leverages a generative adversarial encoder-decoder architecture informed by a neural audio codec to reconstruct high-fidelity 44.1 kHz audio from mel spectrograms. Our approach first transforms the mel spectrogram into a lower-dimensional representation aligned with the Descript Audio Codec (DAC) latent space before reconstructing it to an audio signal using a fine-tuned DAC decoder. DisCoder achieves state-of-the-art performance in music synthesis on several objective metrics and in a MUSHRA listening study. Our approach also shows competitive performance in speech synthesis, highlighting its potential as a universal vocoder.
Keywords
Cite
@article{arxiv.2502.12759,
title = {High-Fidelity Music Vocoder using Neural Audio Codecs},
author = {Luca A. Lanzendörfer and Florian Grötschla and Michael Ungersböck and Roger Wattenhofer},
journal= {arXiv preprint arXiv:2502.12759},
year = {2025}
}
Comments
Accepted at ICASSP 2025