Loop Forge

Clowns on a city street

Works, from LTX 2.5 Creative Lab LoRAs, tested locally

Frame from the Out of focus result
Out of focus
Frame from the Deblur LoRA, 1536x864 result
Deblur LoRA, 1536x864
Watch it in the video at 10:30
LoRA
Deblur at 1.0
Output
1536×864, 161 frames at 30fps
Sampling
8 steps, euler_ancestral
Seed
42
Generated in
14 min
Peak VRAM
7.5 GiB
Ran on
RTX 3060, 12GB (local)

What was wired in

  • Clowns walking
    Guide video
    Clowns walkingBy Coverr on Pexels, Pexels License. First 161 frames, scaled to 1536×864Download MP44.4 MB
Works

This is phone footage that is out of focus from start to finish. The clowns' costumes, the cars, the police motorcycle and the brickwork come back sharp. The shop signs are redrawn as new lettering that does not say anything.

Reference shows a street scene in colour on an overcast day: two clowns in bright costumes walk from left to right along a city street, one in a red coat with blue and red checked trousers, a yellow wig, a small red hat and big red shoes, the other in a pink outfit behind, past a parked grey SUV and a black sedan, a white police motorcycle, and brick and stone shopfronts with red signs and lit windows, heavily out of focus with soft defocused blur and no fine detail. Edited shows the same scene in sharp focus with crisp detail and clean edges. DEBLUR a street scene in colour on an overcast day: two clowns in bright costumes walk from left to right along a city street, one in a red coat with blue and red checked trousers, a yellow wig, a small red hat and big red shoes, the other in a pink outfit behind, past a parked grey SUV and a black sedan, a white police motorcycle, and brick and stone shopfronts with red signs and lit windows Subject identity, framing, and background geometry are identical to the reference; only focus and sharpness differ between reference and edited.

Notes

I ran it at 1536×864, since the card suggests about 1536 wide for more detail. The trigger is DEBLUR, and it is meant for defocus only.

The Deblur LoRA on Hugging Face

Source footage: Clowns walking, by Coverr on Pexels, Pexels License

Run it yourself

Download everything for this test (.zip, 4.4 MB). It holds the workflow, every input it loads and a README. Copy the inputs into ComfyUI/input, put the models below in place and drag the workflow onto ComfyUI.

FilePut it inSize
ltx-2.5-22b-distilled-transformer-comfy-int8-convrot.safetensors gatedmodels/diffusion_models21.5 GB
gemma4-12b-with-proj-ltx-2.5-comfy-int8-convrot.safetensors gatedmodels/text_encoders15.4 GB
ltx-2.5-video-vae-bf16.safetensors gatedmodels/vae1.5 GB
ltx-2.5-audio-vae-bf16.safetensors gatedmodels/vae364.9 MB
ltx-2.5-22b-ic-lora-deblur-0.9.safetensors gatedmodels/loras906.1 MB

The Lightricks repositories (the LTX 2.5 base model and the LoRAs) are gated. Log in to Hugging Face and accept the terms on each repository page before downloading.

Custom nodes: ComfyUI-LTXVideo at commit 3bf3ca6. Everything else is part of ComfyUI 0.37.0 (commit 0f74f7fb, frontend 1.53.6).