P: Store AI related data in seperate sidecar file (not XMP)

@DdeGannes 

Yes, like many other users, I activated that option to automatically also store metadata in XMP sidecar files (not only in the database).

And if one activates it, the AI-related data is stored in the XMP files, too. (By design.) Pumping its size up (from usually 16-17 kB or less then 1 MB with AI masks) to several MB each.

Additionally to the lrcat.data storage. (By design.)

And seemingly also in the lrcat storage. (As discussed in the bug report, linked in my previous comment.)

My request is, to only store the AI development parameters in XMP (e.g. denoise on/off & intensity-level) but to not store the AI-data in the XMP.

Instead the data could be stored in a seperate, second, more efficient binary sidecar file.

And the reasons why I belive this to be better can be found in my opening post and in the linked discussions.

@robert36972564 , I am not sure I understand the request, LrC by default stores the data in the LrC Catalog file by default there is an option to also save to XMP but while working in LrC the app is reading and writing to the Catalog. Data for th AI functions are stored in the .lrcat.data file so there is not a necessary for the .XMP files,

I just found out about another, still active bug. This idea might also be part of a solution for this issue:

https://community.adobe.com/t5/lightroom-classic-bugs/p-automatically-write-changes-to-xmp-can-t-be-used-with-denoise-on-large-numbers-of-photos/idi-p/15393188

It's at least related closly related and both could be considered together when working on each other.

"Other file formats can save 50%+ storage capacity (as they allow for more efficient binary storing of that data instead of text-encoded storing in [XMP])."

A minor correction that doesn't change the substance of your proposal: Binary data is encoded in XMP using Ascii85, which increases the size of the stored data by 25%, not 50%. (Ascii85 uses 5 bytes to represent 4 binary bytes.)