# Wave-Trainer-Fit: Neural Vocoder with Trainable Prior and Fixed-Point Iteration towards High-Quality Speech Generation from SSL features

Source: [Wave-Trainer-Fit: Neural Vocoder with Trainable Prior and Fixed-Point Iteration towards High-Quality Speech Generation from SSL features](<https://www.sota2.com/research/paper/6985758ac97a0b3e3b5d5c70>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-02-07T05:00:04.109Z

Authors: Hien Ohnaka, Yuma Shirahata, Masaya Kawamura

Published: 2026-02-05T08:33:54.000Z

## Abstract

We propose WaveTrainerFit, a neural vocoder that performs high-quality waveform generation from data-driven features such as SSL features. WaveTrainerFit builds upon the WaveFit vocoder, which integrates diffusion model and generative adversarial network. Furthermore, the proposed method incorporates the following key improvements: 1. By introducing trainable priors, the inference process starts from noise close to the target speech instead of Gaussian noise. 2. Reference-aware gain adjustment is performed by imposing constraints on the trainable prior to matching the speech energy. These improvements are expected to reduce the complexity of waveform modeling from data-driven features, enabling high-quality waveform generation with fewer inference steps. Through experiments, we showed that WaveTrainerFit can generate highly natural waveforms with improved speaker similarity from data-driven features, while requiring fewer iterations than WaveFit. Moreover, we showed that the proposed method works robustly with respect to the depth at which SSL features are extracted. Code and pre-trained models are available from https://github.com/line/WaveTrainerFit.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Waveform Generation on LibriTTS-R clean (test)](<https://www.sota2.com/research/sota/waveform-generation-on-libritts-r-clean-test.md>): Waveform Generation; LibriTTS-R clean (test); Speech BERT Score: 95.28.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2602.05443>)

[PDF](<https://arxiv.org/pdf/2602.05443.pdf>)
