WEVSR: Video Diffusion Generators for Real-World Video Super‑Resolution with Wavelet-Enhanced VAE Encoder

Recent advances in video diffusion models have demonstrated remarkable generative capability, yet adapting these large pretrained text-to-video (T2V) models to video super‑resolution (VSR) typically encounters challenges, such as artifacts introduced by complex degradations in real-world scenarios and compromised fidelity due to the strong generative capacity of the powerful T2V models. We present WEVSR, a novel approach that adapts a pretrained flow-matching video diffusion transformer to VSR. First, we design a task-oriented adaptation strategy that leverages timestep sampling and noise augmentation to enhance detail restoration while preserving structural stability. Second, we propose a lightweight multi-level discrete wavelet transform (DWT) front-end for the VAE encoder, injecting explicit frequency priors into the latent space without modifying the pretrained decoder. Extensive experiments across multiple VSR benchmarks demonstrate that WEVSR achieves state-of-the-art performance against existing approaches. Code and models will be released here.

This paper introduces WEVSR, a new method designed to make low-quality videos look clearer, sharper, and more natural. Instead of trying to guess missing details frame by frame, WEVSR builds on modern video‑generation technology, which is very good at creating realistic motion and textures. The method adds two key ideas: it teaches the system to better understand fine details like edges and textures, and it guides the model to focus on the parts of the video that matter most for clarity. As a result, WEVSR produces videos that look more detailed and stable over time, without the flickering or artificial look that many enhancement methods create. Tests on a wide range of real-world videos show that WEVSR consistently improves both visual quality and realism, making it a strong step forward for practical video enhancement.