Request for Visual/Interface Enhancements

[hyupuunos]

​Request for Visual/Interface Enhancements​

Hello,

I must once again seek your expert guidance. This time, my request focuses on visual/interface improvements:





1. ​Integrated Image Input Ports in Image Filter​

• ​Proposal: Integrate 1-3 image input ports directly into the image filter.
• ​Purpose:
• Eliminate the need for an external image batch node, simplifying the workflow.
• Enable image border customization (e.g., color-coded borders) to visually distinguish inputs from different filters.
• ​Impact: During intensive 8-hour sessions with dozens of images, this would drastically reduce eye strain from tracking unmarked inputs.
  

2. ​Zoom Functionality for Image Selection​

• ​Request: Allow zooming during image selection.
• ​Purpose: Facilitate inspection of subtle details (e.g., noise patterns, edge artifacts).
• ​Impact: Prevent physical exhaustion caused by squinting at small differences.
  
  

3. ​Larger Font for Tips Display​

• ​Request: Display tips in a larger font at the top of the screen.
• ​Purpose: Clarify the current processing stage when using multiple image filters.
• ​Impact: Avoid confusion during repetitive daily workflows.
  
  

4. ​Tips and Extra Fields for Mask Image Filter​

• ​Request: Add tips and at least one extra field (e.g., denoise value adjustment) to the mask image filter.
• ​Purpose: Enable real-time fine-tuning after applying masks.
• ​Impact: Streamline post-mask adjustments without workflow interruptions.
  

---


These suggestions are born from high-intensity usage. Your expertise has always elevated this tool — might we refine these details to protect the eyes of dedicated users? Thank you for considering!

[hyupuunos]

After an 8-hour pixel-hell marathon, my eyeballs are filing for workers' comp and threatening to unionize.

[hyupuunos]

Could we make the red selection border pop more?

[chrisgoringe]

the one when you hover the mouse over? It's already pretty intrusive, I reckon!

[hyupuunos]

During prolonged sorting sessions with large image batches — or when eyes are exhausted from fatigue — [the element] still isn't prominent enough for quick visual recognition

[chrisgoringe]

OK, so taking this one at a time :)

1. This would eliminate the need for an extra batch node, but at the expense of making the number of image inputs limited, and making the node itself more complicated. I'm not convinced. Not sure exactly what you are after for 'image border customisation' - do you mean that in the Image Chooser, images could be coded to indicate where they came from? That's worth exploring, but doesn't need multiple inputs (maybe some sort of field that is like : 1-4 red, 5-8 green)

2. Are you aware that when you mouse over an image, you can press the space bar to enlarge it? And you can then move through the images with the left and right arrows (space bar again to unzoom, Up Arrow to 'click' that image)?

3. I'm open to making the tips clearer, although I tend to use a lot of text in mine, so it's a tradeoff. Did you know that if you rename the node, the name gets shown where you've written "Tips"?

4. Yeah, I'd like to be able to do this. But because I'm using the built in Comfy Mask Editor it isn't accessible without a lot of hacking which would result in a fragile outcome. I tend to feed the Mask Filter output straight into a Text Image Filter (you can feed the mask as well as the image and it superimposes them). Not perfect, but I've learned the hard way not to connect too deeply into the Comfy code (because they keep changing it!)

[hyupuunos]

Thank you for your response:​​

1. I believe it won't limit the number of image inputs, as the first default input port remains unchanged and still accepts external image batches via standard nodes. The 3-4 additional input ports added later are the ones requiring custom borders. These borders serve purely as visual distinctions.
   ​Example from my workflow:​​
   When chaining three Image Filter nodes:
   The first Image Filter receives an "Original Reference" image (e.g., gold-bordered) for side-by-side comparison.
   The second Image Filter imports a color-adjusted version (e.g., metallic-bordered) for tonal analysis.
   The third follows the same logic.
   While custom borders don't affect functionality, they are critically important for instantly identifying reference images during repetitive tasks — especially when handling multiple auxiliary inputs that would otherwise require tedious counting to differentiate.

2. I was genuinely unaware of the ​Spacebar zoom feature. After testing it, this resolves my initial concern. My apologies for the oversight! That said, synchronized mouse-wheel zoom across all images would still be a welcome enhancement for rapid multi-image inspection (though it's no longer urgent).

3. For workflows with multiple Image Filter nodes performing distinct tasks, I utilize the Tips field to label each node's purpose. Any dedicated field with ​large, eye-catching text​ would suffice, as clarity is paramount for quick recognition.

4. As explained in 3, adding Tips to the Mask Image Filter is mission-critical for multi-node chains. Parameter passing (e.g., denoise values) can be prioritized later — thank you for considering this exploration!

​In summary, as a retoucher, I now regard your plugin as an ​essential tool​ and regret not discovering it sooner. For heavy users like myself who engage in prolonged, high-intensity editing sessions, maintaining visual comfort is vital. It took extensive experimentation to articulate these needs, but I trust this feedback honors the brilliance of your work. Once again, thank you — your plugin is a game-changer, and I wholeheartedly commend your innovation!

[chrisgoringe]

For the "tips" - the title of the node is currently displayed on all the nodes except the mask. In my workflows I retitle the nodes to describe what they are for, which sounds like your use case -

So maybe I could

• make that text a little larger, and
• superimpose it onto the mask editor window (I think I can do that without integrating into it)

For the other items in this thread, could I ask you to create issues for them separately - one request per issue - it's much easier to keep track of them that way. You can always link back to this discussion in the issue (just type #84) to include a link, so no need to retype everything.

[hyupuunos]

okok，thanks

[chrisgoringe]

Thanks for adding the issues, @hyupuunos.

I've had an idea that I'd be interested in your thoughts about. It works across all of the nodes (including mask) - and gives mask 'extras' and 'tip'

Basically I've moved the controls into a floating window that you can move around.

If you'd like to try it, it's on the dev branch:

cd cg-image-filter
git pull
git switch dev

Then restart comfy and reload the page.

Style changes are pretty easy, so please let me know.

git switch main to swap back.