2🐕Saturation migration
Borrow just how vivid another image looks
- source_image
- target_image
- result_image
This is the narrowest of ergouzi's color-migration family. Where the pack's "regular" color migration node moves the whole color relationship between two images, Saturation migration touches exactly one thing: how vivid the target image looks, matched to a reference. Hue stays put, brightness stays put, only the intensity of the color shifts.
That's a genuinely useful scalpel to have. One of the most repeated pieces of advice for making AI output read as photographic rather than "AI-generated" is that generations tend to sit oversaturated and high-contrast by default, and pulling saturation back a notch is a cheap fix for it. Usually people do that with a flat percentage slider and eyeball it. This node does the same job differently: instead of guessing "pull it down 8%," you point it at a reference image whose saturation level you actually like - a real photo, a previous generation you were happy with - and it matches toward that, which tends to land better than a blind percentage when your reference and your target don't start from the same baseline.
Inputs and output
Just two required inputs: source_image (the one supplying the saturation level you want) and target_image (the one that gets adjusted). No strength dial on this one, unlike the pack's broader "Regular color migration" node - this is a fixed, all-or-nothing transfer of that single property. The output is result_image, the target with its saturation matched.
Installing it
Search "Comfyui-ergouzi-Nodes" in ComfyUI Manager and install, or clone it manually:
cd ComfyUI/custom_nodes
git clone https://github.com/11dogzi/Comfyui-ergouzi-Nodes.git
Restart ComfyUI. No model downloads needed - it's pixel-level color math, nothing to fetch, nothing that adds meaningfully to your load time. One thing worth knowing regardless of which node in this pack you're using: the author says this English repo isn't where new development happens anymore. There's a separate, more actively maintained Chinese-first sibling repo; this branch is stable, just frozen.
Common issues & troubleshooting
Because there's no strength control, the entire result depends on picking a sensible source_image. If your source is itself oversaturated (a lot of stock "vibrant" reference photos are), matching toward it will push your target further into the oversaturated territory you were probably trying to escape - check your reference image's own saturation before assuming the node did something wrong.
If the two images have very different subject matter, results can look plausible but slightly off, since the migration is working on statistics across the whole frame rather than understanding what's in it. If the effect looks too strong or too subtle and you need finer control than an all-or-nothing swap, the pack's "Regular color migration" sibling node adds an adjustable strength input and is worth switching to for that reason alone.
Inputs (2)
| Name | Type | Default | Description |
|---|---|---|---|
| source_image | IMAGE | — | |
| target_image | IMAGE | — |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| result_image | IMAGE | — |