Advanced Tagging options for Images?

[Lynx_TWO]

Hello!

I've been doing a lot of Ai generated art and I noticed that often, the output image will have the prompt or description as a caption or title tag. However, once I upscale the image, the output will often be lacking the tags, so if I upload it to social media I then have to re-enter the captions.

Would it be possible to edit the EXIF data in images like we edit tags in audio? It would also be cool to export the tag info as an XMP file (minus the resolution or size of course) and then import that XMP data to new files (just place in the same folder, the upscaled images with the same name and replace the originals)

Thanks for all you do!

P.S. - I noticed that WebP lossless compression option is broken - it converts to a much smaller file than even the lossy option at 100%

[Spoon]

Re: Advanced Tagging options for Images?

The latest version does have a tag editor, right click >> Edit Tags

[Lynx_TWO]

Re: Advanced Tagging options for Images?

Outstanding Spoon, Thanks!!!

[Spoon]

Re: Advanced Tagging options for Images?

About the lossless option, it can produce smaller files than the lossy, depends on what you are compressing, if text then lossless will likely be smaller.

[Lynx_TWO]

Re: Advanced Tagging options for Images?

About the lossless option, it can produce smaller files than the lossy, depends on what you are compressing, if text then lossless will likely be smaller.

That makes sense, and I have noticed when using the RIOT software, lossless is often a bit smaller than the nearly lossless option.

So I dug into this a bit more and the issue I'm running into is, that if I convert a PNG image using the lossless option and the Photo preset, the PNG image goes from 15MB to 750kB. I was blown away as obviously, that would be a MASSIVE increase in efficiency for WebP over PNG. But I got suspicious so I opened both images side by side, and there was obvious macro blocking. I tried the Lossy option with the photo preset at 100% and the output was 4MB.

Odd right? So, I did it again but selected None as the preset, and the Lossless option. The output was a WebP image at 12MB. So basically, from a GUI standpoint it looks like if the user selects the Lossless option, the presets should default to none and the user should not be able to select any other preset. Does that make sense?