Is GAN Necessary for Mel-Spectrogram-Based Neural Vocoder?
IEEE Signal Processing Letters · 2025 · 32 · 3485-3489
DOI: 10.1109/lsp.2025.3598729 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Signal Processing Letters, 2025. Verfasst von Hui-Peng Du und 4 weiteren Autor:innen. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 33 Literaturangaben. Gefördert von: National Natural Science Foundation of China, Natural Science Foundation of Anhui Province.
Abstract (Originalsprache)
Recently, mainstream mel-spectrogram-based neural vocoders rely on generative adversarial network (GAN) for high-fidelity speech generation, e.g., HiFi-GAN and BigVGAN. However, the use of GAN restricts training efficiency and model complexity. Therefore, this letter proposes a novel FreeGAN vocoder, aiming to answer the question of whether GAN is necessary for mel-spectrogram-based neural vocoders. The FreeGAN employs an amplitude-phase serial prediction framework, eliminating the need for GAN training. It incorporates amplitude prior input, SNAKE-ConvNeXt v2 backbone and frequency-weighted anti-wrapping phase loss to compensate for the performance loss caused by the absence of GAN. Experimental results confirm that the speech quality of FreeGAN is comparable to that of advanced GAN-based vocoders, while significantly improving training efficiency and complexity. Other explicit-phase-prediction-based neural vocoders can also work without GAN, leveraging our proposed methods.
Förderung
- National Natural Science Foundation of China
- Natural Science Foundation of Anhui Province
Zitieren
APA
Du, H., Ai, Y., Zheng, R., Lu, Y., & Ling, Z. (2025). Is GAN Necessary for Mel-Spectrogram-Based Neural Vocoder?. IEEE Signal Processing Letters, 32, 3485-3489. https://doi.org/10.1109/lsp.2025.3598729
Vancouver
Du H, Ai Y, Zheng R, Lu Y, Ling Z. Is GAN Necessary for Mel-Spectrogram-Based Neural Vocoder?. IEEE Signal Processing Letters. 2025;32:3485-3489. doi:10.1109/lsp.2025.3598729
BibTeX
@article{Du2025,
author = {Du, Hui-Peng and Ai, Yang and Zheng, Rui-Chen and Lu, Ye-Xin and Ling, Zhen-Hua},
title = {Is GAN Necessary for Mel-Spectrogram-Based Neural Vocoder?},
journal = {IEEE Signal Processing Letters},
year = {2025},
volume = {32},
pages = {3485--3489},
doi = {10.1109/lsp.2025.3598729}
}
EUVIMED