# WaveNeXt 2: ConvNeXt-Based Fast Neural Vocoders With Residual Denoising and Sub-Modeling for GAN and Diffusion Models

arXiv: https://arxiv.org/abs/2605.25506 · HTML: https://mlsanity.com/paper/2605.25506/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Audio, speech and music, one of 14 directions in Diffusion models.
- Builds on 23 of its 30 references that are in the map.
- First posted on arXiv 25 May 2026.

## Placed in

- Diffusion models / [Audio, speech and music](https://mlsanity.com/diffusion/audio-speech-and-music/)

## Builds on (most cited first, 23 of 23)

- [PyTorch: An Imperative Style, High-Performance Deep Learning Library](https://mlsanity.com/paper/1912.01703/) (2019)
- [A ConvNet for the 2020s](https://mlsanity.com/paper/2201.03545/) (2022)
- [DiffWave: A Versatile Diffusion Model for Audio Synthesis](https://mlsanity.com/paper/2009.09761/) (2020)
- [HiFi-GAN: Generative Adversarial Networks for Efficient and High Fidelity Speech Synthesis](https://mlsanity.com/paper/2010.05646/) (2020)
- [WaveGrad: Estimating Gradients for Waveform Generation](https://mlsanity.com/paper/2009.00713/) (2020)
- [UTMOS: UTokyo-SaruLab System for VoiceMOS Challenge 2022](https://mlsanity.com/paper/2204.02152/) (2022)
- [MelGAN: Generative Adversarial Networks for Conditional Waveform Synthesis](https://mlsanity.com/paper/1910.06711/) (2019)
- [BigVGAN: A Universal Neural Vocoder with Large-Scale Training](https://mlsanity.com/paper/2206.04658/) (2022)
- [NISQA: A Deep CNN-Self-Attention Model for Multidimensional Speech Quality Prediction with Crowdsourced Datasets](https://mlsanity.com/paper/2104.09494/) (2021)
- [Vocos: Closing the gap between time-domain and Fourier-based neural vocoders for high-quality audio synthesis](https://mlsanity.com/paper/2306.00814/) (2023)
- [FastDiff: A Fast Conditional Diffusion Model for High-Quality Speech Synthesis](https://mlsanity.com/paper/2204.09934/) (2022)
- [LibriTTS-R: A Restored Multi-Speaker Text-to-Speech Corpus](https://mlsanity.com/paper/2305.18802/) (2023)
- [PriorGrad: Improving Conditional Denoising Diffusion Models with Data-Dependent Adaptive Prior](https://mlsanity.com/paper/2106.06406/) (2021)
- [UnivNet: A Neural Vocoder with Multi-Resolution Spectrogram Discriminators for High-Fidelity Waveform Generation](https://mlsanity.com/paper/2106.07889/) (2021)
- [BDDM: Bilateral Denoising Diffusion Models for Fast and High-Quality Speech Synthesis](https://mlsanity.com/paper/2203.13508/) (2022)
- [Multi-band MelGAN: Faster Waveform Generation for High-Quality Text-to-Speech](https://mlsanity.com/paper/2005.05106/) (2020)
- [SpecGrad: Diffusion Probabilistic Model based Neural Vocoder with Adaptive Noise Spectral Shaping](https://mlsanity.com/paper/2203.16749/) (2022)
- [iSTFTNet: Fast and Lightweight Mel-Spectrogram Vocoder Incorporating Inverse Short-Time Fourier Transform](https://mlsanity.com/paper/2203.02395/) (2022)
- [WaveFit: An Iterative and Non-autoregressive Neural Vocoder based on Fixed-Point Iteration](https://mlsanity.com/paper/2210.01029/) (2022)
- [Fixed Point Strategies in Data Science](https://mlsanity.com/paper/2008.02260/) (2020)
- [FreGrad: Lightweight and Fast Frequency-aware Diffusion Vocoder](https://mlsanity.com/paper/2401.10032/) (2024)
- [Hierarchical Diffusion Models for Singing Voice Neural Vocoder](https://mlsanity.com/paper/2210.07508/) (2022)
- [SpecDiff-GAN: A Spectrally-Shaped Noise Diffusion GAN for Speech and Music Synthesis](https://mlsanity.com/paper/2402.01753/) (2024)

## Abstract (arXiv)

Most neural vocoders are limited to one type: either GAN or diffusion-based. While state-of-the-art models like Vocos and WaveNeXt use powerful ConvNeXt-based generators, they have only been used in GAN frameworks and have limited performance in multi-speaker settings. Moreover, diffusion models, despite training faster than GANs, have slow CPU inference. In this paper, we introduce WaveNeXt 2, a unified ConvNeXt-based framework compatible with both GAN and diffusion vocoders. Its core innovation is residual denoising and sub-modeling, where each sub-model progressively refines the waveform. Experimental results in the multi-speaker dataset demonstrate the effectiveness of our approach: (1) GAN-WaveNeXt 2 is much faster than HiFi-GAN and WaveFit, and (2) Diff-WaveNeXt 2 also delivers much faster inference and competitive synthesis quality compared with FastDiff with 4 steps. The Diff-WaveNeXt 2 is very training-efficient, training in only 32 hours, making it ideal for resource-constrained applications.