Paper Detail
Sang-Hoon Lee, Ha-Yeong Choi
This work proposes WaveGen, a fully end-to-end diffusion-based text-to-waveform framework based on Spectral Forcing, which jointly learns an internal spectral trajectory and the target waveform trajectory within a single model, enabling direct end-to-end text-to-waveform generation. Unlike recent TTS pipelines, WaveGen does not rely on separately trained or externally decoded intermediate representations, such as neural audio codecs, Mel-spectrogram vocoders, or VAE-based latent models. In addition, WaveGen performs internal text-speech alignment within a single model, eliminating the need for external alignment modules such as duration predictors. To preserve a truly end-to-end formulation, our framework further avoids dependence on self-supervised representations, such as BERT-based text embeddings and Wav2Vec 2.0-based semantic representations, as well as semantic distillation methods for training optimization. To achieve this, we carefully design Spectral Forcing, a model architecture with disentangled diffusion heads, and a training objective based on a multi-scale spectral v-loss. We further validate the scalability of end-to-end TTS models by scaling the model size from 0.1B to 0.6B parameters. Using a single architecture, open-source data, and one-stage training, without external modules such as duration predictor or REPA, WaveGen demonstrates fully end-to-end waveform generation achieving state-of-the-art performance on the Emilia training dataset. In particular, WaveGen achieves a WER of 1.47 and a SIM of 0.75 on the Seed-en benchmark. Source code and checkpoints are available at https://github.com/sh-lee-prml/WaveGen
No structured notes yet. Add `summary_sections`, `why_relevant`, `claim_impact`, or `next_action` in `papers.jsonl` to enrich this view.
No ranking explanation is available yet.
No tags.
@misc{lee2026wavegen,
title = {WaveGen: Truly End-to-End Waveform Generation via Internal Spectral Trajectory Learning},
author = {Sang-Hoon Lee and Ha-Yeong Choi},
year = {2026},
abstract = {This work proposes WaveGen, a fully end-to-end diffusion-based text-to-waveform framework based on Spectral Forcing, which jointly learns an internal spectral trajectory and the target waveform trajectory within a single model, enabling direct end-to-end text-to-waveform generation. Unlike recent TTS pipelines, WaveGen does not rely on separately trained or externally decoded intermediate representations, such as neural audio codecs, Mel-spectrogram vocoders, or VAE-based latent models. In addit},
url = {https://huggingface.co/papers/2611.12527},
keywords = {diffusion-based text-to-waveform, Spectral Forcing, end-to-end text-to-waveform generation, TTS pipelines, neural audio codecs, Mel-spectrogram vocoders, VAE-based latent models, internal text-speech alignment, duration predictors, self-supervised representations, BERT-based text embeddings, Wav2Vec 2.0-based semantic representations, semantic distillation, disentangled diffusion heads, multi-scale spectral v-loss, REPA, code available, huggingface daily},
eprint = {2611.12527},
archiveprefix = {arXiv},
}
{}