bloc97

Yes, MadVR has extensive de-noising processes. The goal of Anime4K was never really to denoise low quality encodes. High quality denoising/deblocking is inherently slow (there's no around it). You need...

@Jules-A Here's a small comparison, let's say you have a very noisy image/video:  Here's what it looks like after one pass of Denoise-Bilateral-Mode:  And here's what it looks...

@Jules-A Just like in MadVR NGU denoising, the Anime4K_Upscale_x2_Denoise shaders have in-built denoising (for free, the neural network upscales and denoises at the same time). You can consider them to...

@Jules-A Ah... I was unaware that Crunchyroll content was compressed to that point. Maybe a stronger denoiser would be warranted in the next version. But since you are watching 1080p...

Thanks for the in-depth comparison pictures! I think that it's simply by chance the M shader didn't converge correctly during training. I will try to fix this before the official...

In fact I think I can see very tiny (1 pixel) artifacts in the L version too. There might be broader convergence issues from the adversarial training. I will have...

That's because GAN x2-UL was trained with small amounts of denoising, I will see if we should support 6 GAN+Denoise upscalers, or let the GAN-Restore handle denoising.

The optimal case would be to delegate all degradations to RestoreGAN shaders, and let the SRGAN work with pristine images. The current problem is that Restore shaders sometimes introduce ringing,...

That's basically what we're trying to do with feature-wise transformations. If the shader can actually know how much compression is in an entire image, it can learn to not remove...

Upon further examination of the image you provided, it is a very good example of having all kinds of artistic degradation + unintended noise from compression. There are four things:...