nodes/SuperBeastsAI-ComfyUI-SuperBeasts
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
59 Commits 2 Branches 0 Tags
1bb7dca3cb1f7c902292158f77d8ee115b40c40a
T
Clone
Open with VS Code Open with VSCodium Open with Intellij IDEA
Download ZIP Download TAR.GZ Download BUNDLE

README.md

SuperBeastsAI Nodes

This repository contains custom nodes for ComfyUI created and used by SuperBeasts.AI (@SuperBeasts.AI on Instagram)

Updates

  • 19/07/25: Super Pop Colour Adjustment model and node release
  • 31/07/24: Resolved bugs with dynamic input thanks to @Amorano. Add Github Action for Publishing to Comfy Registry thanks to @haohaocreates
  • 30/07/24: Moved Deflicker & PixelDeflicker to Experimental labels (this will require readding them in your WF but I wanted this to be clearer)
  • 30/04/24:
    • Updated readme with documentation
    • Added Deflicker and PixelDeflicker nodes for reducing flickering artifacts in image sequences.
    • Introduced CrossFadeImageBatches node for smooth cross-fade transitions between image batches
    • Enhanced ImageBatchManagement and MaskBatchManagement nodes with resizing, cropping, and reordering capabilities
  • 27/03/24: Released and also pushed an update to resolve batched images for videos etc

Super Beasts Custom Models (and nodes)

SPCA-Header

Super Pop Color Adjustment (SuperBeasts.AI)

Bring the SuperBeasts “pop” to any image in one click. Months of solo R&D, hundreds of hand‑tuned colour‑grade examples, and a custom residual CNN trained to push vibrant palette, deep blacks, crisp highlights, and HDR bite with ease.

Why I built it

Everything I post gets graded: lift blacks, pull whites, bend curves, punch colour. Existing auto tools (and many AI outputs — looking at you, muddy yellow “whites”) just weren’t landing, especially across batches or iterative generations. Training my own lightweight correction model let me reclaim hours of post and lock in a consistent SuperBeasts look I could reuse — and share.

Personal project, shared as‑is.
Works great across my SuperBeasts production flow and a wide mix of test art, but it’s not exhaustively validated on every style, lighting condition, or colour space. Expect edge cases, occasional overcorrection, and behaviour that varies with resolution/patch overlap. Please experiment and let me know what you think!

If you want to show support please take a momenet to follow me on Instagram @SuperBeasts.AI

Happy generating!!

Same settings multiple situations corrected

Highlights/Shadows Correction

Original SPCA Original SPCA
Original 1 S.P.C.A 1 Original 2 S.P.C.A 2
Original 3 S.P.C.A 3 Original 4 S.P.C.A 4

Dull/Washed Out/Muted Correction

Original SPCA Original SPCA
Original 1 S.P.C.A 1 Original 2 S.P.C.A 2
Original 3 S.P.C.A 3 Original 4 S.P.C.A 4

Black/White Levels & Contrast Correction

Original SPCA Original SPCA
Original 1 S.P.C.A 1 Original 2 S.P.C.A 2
Original 3 S.P.C.A 3 Original 4 S.P.C.A 4

Color Correction

Original SPCA Original SPCA
Original 1 S.P.C.A 1 Original 2 S.P.C.A 2
Original 3 S.P.C.A 3 Original 4 S.P.C.A 4

Vibrancy Improvements

Original SPCA Original SPCA
Original 1 S.P.C.A 1 Original 2 S.P.C.A 2
Original 3 S.P.C.A 3 Original 4 S.P.C.A 4

Restoration

Potential uses in old photo restorations and corrections.

Original SPCA Original SPCA
Original 1 S.P.C.A 1 Original 2 S.P.C.A 2
Original 3 S.P.C.A 3 Original 4 S.P.C.A 4

Color balance correction

Original SPCA Original SPCA
Original 1 S.P.C.A 1 Original 2 S.P.C.A 2
Original 3 S.P.C.A 3 Original 4 S.P.C.A 4

How it works

  1. SB Load Model node loads the ONNX-formatted colour-adjustment network (models/SuperBeasts_ColorAdjustment_512px_V1.onnx).
  2. The model analyses the image in 512 × 512 patches, predicting a residual colour grade.
  3. Patches are seamlessly stitched back with configurable strength, count (variant batching) and overlap for large-resolution output.
  4. A 64 × 64 context thumbnail guides the global colour mapping – leave it empty for natural results or experiment with creative transfers.

Recommended quick-start

Parameter Suggested Notes
Max strength 1.0 0.5-2.0 for subtle ➜ dramatic
Count 1 Increase for automatic strength batching up to your max strength value
Overlap 0.4 Up to 0.9 if GPU / Compute permits
Initial context for batch False True (Experimental - may be useful for tone-locked video frames)

Known limitations / quirks

  • Patch size sensitivity: The network works on 512 px tiles – with images above ~2 K you may notice slight local exposure shifts. Pushing Overlap up to 0.8-0.9 blends these out much better for higher quality results at the cost of extra compute/time.
  • Atmospheric overlays: Because the model corrects blacks / whites aggressively it can dial back matte-style colour washes or cinematic tints. Keep Strength around 0.7, or increase count for variation options or blend the result back with the original.
  • Very high strengths: Values above 2.0 can clip highlights or oversaturate; use sparingly for stylised effects.

Creative workflow ideas

  1. HDR Sandwich – Run Super Pop, then feed the output into HDR Effects (with gentle settings) to amplify micro-contrast.
  2. Colour remap – Plug a palette or reference frame into the Context input to nudge the final toning toward that colour space.
  3. Video – Yes it works quite well across video frames! Try experiement with Initial context for batch to subtely lock colour across all the processed frames or leave as is for each frame to determine the correction independantly.

Understanding the Context input

The context socket takes a 64 × 64 image that acts as a colour-map reference while the network predicts its residual. If you leave it unplugged the node automatically downsamples your source image – this mirrors how the model was trained and is the safest default.

When might you override it?

  • Palette steer – pass in a small palette swatch or another artwork to pull the grade towards what would likely be the correction those hues and levels.
  • Shot matching – feed a hero frame from earlier in a sequence to keep later frames corrections more tonally consistent.
  • Creative grading – drop in a totally different style image (e.g. a sunset or neon sign) for experimental colour correction shifts.

Remember: the context influences global colour balance, not fine detail, so even heavily down-scaled or abstract images work great but this only plays a small role in the final correction - ultimatley the actual 512x512px patch being corrected has the greated weight in the adjustment task.

System Requirements

Works out-of-the-box on CPU via onnxruntime. For 10× speed improvement ensure ComfyUI can use the CUDAExecutionProvider by using onnxruntime-gpu with the hardware that allows this. This node print out the currently available providers into the ComfyUI server logs if you need to check this.

Licensing & weight downloads

This model is released under the SPCA-RAIL licence. Full text lives in the model repository: https://github.com/SuperBeastsAI/SuperBeastsAI-Models/blob/main/SuperPopColorAdjustment/LICENSE.txt • Free for personal & commercial output use.
• Redistribution / hosting of the weights is not allowed without permission.

Weights are automatically downloaded the first time the SB Load Model node runs. If you prefer manual download, grab them from the GitHub release page:

https://github.com/SuperBeastsAI/SuperBeastsAI-Models/tree/main/SuperPopColorAdjustment/downloads/
and drop the .onnx file into custom_nodes/ComfyUI-SuperBeasts/models/.

Image Effects

HDR Effects (SuperBeasts.AI)

The HDR Effects is an image processing application that enhances the dynamic range and visual appeal of input images. It provides a set of adjustable parameters to fine-tune the HDR effect according to user preferences.

Features

  • Adjusts the intensity of shadows, highlights, and overall HDR effect
  • Applies gamma correction to control the overall brightness and contrast
  • Enhances contrast and color saturation for more vibrant results
  • Preserves color accuracy by processing the image in the LAB color space
  • Utilizes luminance-based masks for targeted adjustments
  • Blends the adjusted luminance with the original luminance for a balanced effect

Examples

Example 1

Example 2

Parameter details

The application provides the following adjustable parameters:

  • hdr_intensity (default: 0.5, range: 0.0 to 5.0, step: 0.01):

    • Controls the overall intensity of the HDR effect
    • Higher values result in a more pronounced HDR effect

  • shadow_intensity (default: 0.25, range: 0.0 to 1.0, step: 0.01):

    • Adjusts the intensity of shadows in the image
    • Higher values darken the shadows and increase contrast

  • highlight_intensity (default: 0.75, range: 0.0 to 1.0, step: 0.01):

    • Adjusts the intensity of highlights in the image
    • Higher values brighten the highlights and increase contrast

  • gamma_intensity (default: 0.25, range: 0.0 to 1.0, step: 0.01):

    • Controls the gamma correction applied to the image
    • Higher values increase the overall brightness and contrast

  • contrast (default: 0.1, range: 0.0 to 1.0, step: 0.01):

    • Enhances the contrast of the image
    • Higher values result in more pronounced contrast

  • enhance_color (default: 0.25, range: 0.0 to 1.0, step: 0.01):

    • Enhances the color saturation of the image
    • Higher values result in more vibrant colors

Image & Mask Batch Management

Keep your image and masks sized, cropped and ordered how ever you like without having to recreate the masks or mess with connections. Note the resizing capability doesn't provide any settings, we simply crop and resize to maximise the image size with your provided width/height.

Batch Management

ImageBatchManagement

The ImageBatchManagement node provides functionality to resize, crop, and reorder a batch of images. It ensures that all images in the batch have the same specified dimensions.

Features

  • Resizes and crops images to the specified width and height
  • Allows reordering the images in the batch based on a provided order
  • Supports unlimited input images

MaskBatchManagement

The MaskBatchManagement node is similar to the ImageBatchManagement node but works with mask inputs. It resizes, crops, and reorders a batch of masks to match the specified dimensions.

Features

  • Resizes and crops masks to the specified width and height
  • Allows reordering the masks in the batch based on a provided order
  • Supports unlimited input masks

Deprecated / Removed

Deflicker & PixelDeflicker

The Deflicker nodes have been marked as "Experimental" as they were experiments that were released but not really fit for purpose / doing what you would expect of a deflicker node. At some point I will be removing these nodes as I won't be supporting them at this stage.

MakeResizedMaskBatch (Deprecated please use MaskBatchManagement)

The MakeResizedMaskBatch node creates a batch of masks from multiple individual masks or batches. It resizes and crops the input masks to match the specified width and height.

Cross Fade Image Batches (SuperBeasts.AI)

There is another preexisting node that does this and more so I've removed this function.

Reference in New Issue View Git Blame Copy Permalink
S
Description
No description provided
Readme MIT
417 MiB
Languages
Python 95.7%
JavaScript 4.3%