Turns out LTX-2 makes a very good video upscaler for WAN

I have had a lot of fun with LTX but for a lot of usecases it is useless for me. for example this usecase where I could not get anything proper with LTX no matter how much I tried (mild nudity):
https://aurelm.com/portfolio/ode-to-the-female-form/
The video may be choppy on the site but you can download it locally. Looks quite good to me and also gets rid of the warping and artefacts from wan and the temporal upscaler also does a damn good job.
First 5 shots were upscaled from 720p to 1440p and the rest are from 440p to 1080p (that's why they look worse). No upscaling outside Comfy was used.

workwlow. I could not get a proper link of the 2 steps in one run (OOM) so the first group is for wan, second you load the wan video and run with only the second group active.

https://aurelm.com/upload/ComfyWorkflows/Wan\_22\_IMG2VID\_3\_STEPS\_TOTAL\_LTX2Upsampler.json

This are the kind of videos I could get from LTX only, sometimes with double faces, twisted heads and all in all milky, blurry.
https://aurelm.com/upload/ComfyUI\_01500-audio.mp4
https://aurelm.com/upload/ComfyUI\_01501-audio.mp4

Denoising should normally not go above 0.15 otherwise you run into ltx-related issues like blur, distort, artefacts. Also for wan you can set for both samplers the number of steps to 3 for faster iteration.

https://redd.it/1rb4ms7
@rStableDiffusion

How would you go about generating video with a character ref sheet?

I've generated a character sheet for a character that I want to use in a series of videos. I'm struggling to figure out how to properly use it when creating videos. Specifically Titmouse style DnD animation of a fight sequence that happened in game.

Would appreciate an workflow examples you can point to or tutorial vids for making my own.

https://preview.redd.it/kpallbyckxkg1.png?width=1024&format=png&auto=webp&s=d0fe33baeabeee6d356020ea81c0bae707cad638

https://preview.redd.it/805h1eyckxkg1.png?width=1024&format=png&auto=webp&s=42ef42bde1edee800e25210bf471831c93290726



https://redd.it/1rb5n9h
@rStableDiffusion

A single diffusion pass is enough to fool SynthID

I've been digging into invisible watermarks, SynthID, StableSignature, TreeRing — the stuff baked into pixels by Gemini, DALL-E, etc. Can't see them, can't Photoshop them out, they survive screenshots. Got curious how robust they actually are, so I threw together noai-watermark over a weekend. It runs a watermarked image through a diffusion model and the output looks the same but the watermark is gone. A single pass at low strength fools SynthID. There's also a CtrlRegen mode for higher quality. Strips all AI metadata too.

Mostly built this for research and education, wanted to understand how these systems work under the hood. Open source if anyone wants to poke around.

github: https://github.com/mertizci/noai-watermark

https://redd.it/1rbb24f
@rStableDiffusion

I Combined Wan Animate 2.2 Complete Ecosystem Workflow | SCAIL + SteadyDancer + One-to-All Workflows Into ONE Ultimate Multi-Character Animation Setup (Now on CivitAI)
https://redd.it/1rbftee
@rStableDiffusion

[Final Update] Anima 2B Style Explorer: 20,000+ Danbooru Artists, Swipe Mode, and Uniqueness Rank

https://redd.it/1rh2890
@rStableDiffusion

Control your AI for art and design

https://redd.it/1rgu5yw
@rStableDiffusion

For very low resolution videos restoration, SeedVR2 is better than FlashVSR+ like 256px to 1024px

https://redd.it/1rgovde
@rStableDiffusion

Fast Flux2K inpainting on 8+ mp images without upscale

https://redd.it/1rh8aza
@rStableDiffusion

Z-Image-Turbo Controlnet Union 2.1 version 2602 just released

https://preview.redd.it/je2zyojhf9mg1.png?width=917&format=png&auto=webp&s=7eb32d6dca2a129acde4b1137275aabf116c7505

[2026.02.26\] Update to version 2602, with support for Gray Control.

Personally I had much better results with the Lite versions BTW (the full versions really produced very bad quality outputs, for some reason)

Download: https://huggingface.co/alibaba-pai/Z-Image-Turbo-Fun-Controlnet-Union-2.1/tree/main

https://redd.it/1rh6nwr
@rStableDiffusion

Z-Image-Fun-Controlnet-Union v2.1 Tile available

https://preview.redd.it/rovv9lwrj8mg1.png?width=946&format=png&auto=webp&s=073edea7da210bf08f9b4329608fa8f052c41fab

DOWNLOAD

https://redd.it/1rh2gck
@rStableDiffusion

My LTX2 Night of the Living Dead Submission

https://redd.it/1rh9wu2
@rStableDiffusion

Act step 1.5 M2M best practices - do we have them?

Love ace step 1.5. Amazing and fast for text to music. But music to music, it's terrible. At medium noise, it changes the songs completely. Essentially the same as t2m but lower quality. At low denoise it just messes up audio quality.

Anyone manged to get decent results out of music to music? E.g. tweaking genre, replacing some words in lyrics, or similar?

https://redd.it/1rh6lmz
@rStableDiffusion