ComfyUI Node
SF EXR to Image
Turn a scene-linear layer into the 0-1 image a diffusion model expects. Pair with SF Image to EXR to put the result back into the EXR. With a mask, 'crop to mask' hands the model only the masked region at the size it works best at; the return pastes it back into the full-resolution plate.
SF EXR to Image
- exr
- mask
- image
- context
- alpha
- crop mask
◄layerauto►
◄source_spaceauto (from file)►
◄encodingsRGB (ACES SDR view)►
◄exposure0.0►
◄crop▾►
CategorySF EXR
Inputs (7)
| Name | Type | Default | Description |
|---|---|---|---|
| exr | SF_EXR | — | |
| layer | COMBO | auto | 'auto' is the beauty. |
| source_space | COMBO | auto (from file) | 28 options: auto (from file), CIE XYZ-D65 - Display-referred, sRGB - Display, Gamma 2.2 Rec.709 - Display, Display P3 - Display, Display P3 HDR - Display, +22 |
| encoding | COMBO | sRGB (ACES SDR view) | ACEScct keeps more highlight range; sRGB looks like a normal photo. |
| exposure | FLOAT | 0.0-10–10 | Stops; undone on return. |
| crop | COMBO | 'crop to mask' needs the crop mask input. One crop covers it on every frame. | |
| maskopt | MASK | The area to edit. Crop to mask crops to it, it comes out matching 'image' for inpainting, and SF Image to EXR blends through it. |
Outputs (4)
| Name | Type | Description |
|---|---|---|
| image | IMAGE | — |
| context | SF_EXR_MODEL_CONTEXT | — |
| alpha | MASK | The layer's own alpha channel. |
| crop mask | MASK | The crop mask matching 'image' (cropped and scaled with it); empty when none is connected. |