Reducing resolution in a png file

[Bwip877]

Select Topic Area

Question

Body

Hello. Would someone please tell me if using the windows 'Paint' to reduce a jpeg's resolution and also transferring it from a jpeg to a png effects how comfyui renders it. It seems to me that most of the Png's that I use after I convert them, Comfy goes a little crazy as far as the prompts don't always work and the videos come out sort of lousy and if this is the case, what do people use to manipulate images beforehand? Just a shot in the dark. I thought I'd ask. Thanks!

[yusufkecer]

When you use Windows Paint to convert jpeg to pngand resize them, it's likely messing up your Comfyui results because Paint isn't really designed for this kind of work. It doesn't handle color profiles properly, keeps all those jpegcompression issues but locks them into the pngfile, and throws away important image data during the conversion. The resizing algorithm it uses is pretty crude too, which doesn't help.
You'd be better off using something like IrfanView or gimpif you want free options theyre much more careful with how they handle images and won't destroy important information. Honestly though, if you're working with ComfyUI, you might want to just keep your images as jpegs until you actually need them as png and let Comfyui handle the resizing through its own nodes. That way you're not introducing problems before the aieven gets to work with your images.

[Bwip877]

I figured something was going on. Thanks very much for that info. I appreciate it!

[openvaibhav]

Converting a jpeg to png and resizing it in paint can definitely impact how ComfyUI interprets the image. Jpeg is a lossy format so any detail lost during compression is permanent and saving it as PNG doesn’t restore that data it just preserves the existing artifacts in a lossless container and paint uses very basic resizing algorithms which can soften edges, blur fine details, and degrade texture information that diffusion models rely on for prompt accuracy and stable renders especially in video workflows. It may also mishandle color profiles or strip metadata. For better results it’s recommended to resize or convert images using tools like GIMP, Krita, Photoshop, or even ComfyUI own upscale/resize nodes which preserve more useful image information for AI processing.

[Bwip877]

Thank you, I'll do that.

[Bwip877]

One more quick question before I close this. Where is the setting to disable the little rectangular box above a node when you left click on the top of the node?

[openvaibhav]

That little rectangular box is the node header context menu or quick actions popup that appears when you click the title bar of a node and as far as current ComfyUI builds go there isnt a dedicated setting in the UI to disable it. Its part of the default LiteGraph node interface that ComfyUI uses. If its getting in the way the only real workarounds are modifying the frontend css/js or using a custom fork or theme that changes node UI behavior otherwise its expected behavior and cant be toggled off from the standard settings panel.

[Bwip877]

Thanks!