weepiess2383/Spectral_Forcing

Show the Signal, Hide the Noise: Spectral Forcing for Pixel-Space Diffusion

This repository contains the weights for Spectral Forcing, a parameter-free, time-conditional low-pass operator that makes the coarse-to-fine structure of diffusion explicit at the input of a pixel-space model.

• Paper: Show the Signal, Hide the Noise: Spectral Forcing for Pixel-Space Diffusion
• Github Repository: https://github.com/WeichenFan/Spectral_Forcing

Introduction

Spectral Forcing (SF) is based on the observation that under rectified-flow diffusion and natural-image power-law spectra, a moving frequency front separates a signal-bearing low-frequency region from a noise-dominated high-frequency region at each timestep. SF imposes this boundary directly via a 2D-DCT mask whose cutoff expands with the diffusion time and becomes the identity at the data endpoint, showing the network the signal and hiding the noise.

BibTex

@misc{fan2026show,
  title         = {{Show the Signal, Hide the Noise: Spectral Forcing for Pixel-Space Diffusion}},
  author        = {Fan, Weichen and Diao, Haiwen and Wu, Penghao and Liu, Ziwei},
  year          = {2026},
  eprint        = {2606.15236},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CV},
  doi           = {10.48550/arXiv.2606.15236},
  url           = {https://arxiv.org/abs/2606.15236}
}