Relations: Metadata Propagation

[JohnZeman]

According to

https://www.photools.com/help/imatch/#rel_versions.htm#a_prop

Metadata propagation from masters to versions should happen upon metadata writeback for propagation items indicated by an asterisk (*)

I would like the XMP Description* copied from the master to the version and it does that but not automatically upon writeback.  I have to manually trigger the propagation (F4,P) and then it copies the description to the version.

Am I missing something?
Thanks.

[Mario]

I have tried this.

I have a RAW -> PSD/JPG relation.
One version in the same folder as the RAW, 2 versions in a folder below that.
When I change the description for the RAW (master) and write back, the same description is written to all versions.
IMatch writes the master, then copies the propagated tags to the versions (in my test case, description and keywords).
This can be seen in the ExifTool Output panel as well.

There must be some condition in your files or with your settings which causes this.
Double check for other propagated data, maybe issues with exiting EXIF or IPTC descriptions which override the XMP description on import (!) etc.

[JohnZeman]

I have two sets of relations configured.

DNG Master/JPG Versioning
DNG Buddy Files

TIF Master/JPG Versioning
TIF Buddy Files

The buddy files for both versions are configured so the DxO .dop buddy files are moved with the masters when IMatch moves the files.  This works as expected.

The versions are configured to copy XMP Descriptions, XMP Instructions, and XMP Keywords from master to versions.

The masters and versions are in the same folder, that is all DNG, JPG, TIF images taken in 2019 are in a single folder named 2019.

When I make a change to a DNG or TIF XMP description, XMP keywords, and/or XMP Instructions then perform a writeback to file, afterwards if I use the IMatch Metadata Browser to search the JPG version for the changes I made no changes are found.  Until I manually do a Refresh Relations command, after I do that the metadata is propagated to the jpg versions.

I don't know if it will help but I've attached the logfile (I have full logging enabled) along with the contents of the ExifTool Output panel after doing a writeback to a master DNG.

Thanks Mario

[Mario]

If versions are updated, their file names appear in the output file.
The output panel data shows no trace of propagation.
Data is only written to the DNG file. No other files are updated.
This explains the results.

Make sure your versions are set up correctly, that the versions are found (check in Version panel) etc.

Also: You are using version chaining, which is to be avoided.
You propagate from DNG to TIF and JPEG, but also from TIF to JPEG. This may have unexpected consequences.

The log file mentions some ExifTool warnings about maker notes in the DNG and some invalid dates in the TIF. This should not harm, though. Just standard metadata mess.

[Carlo Didier]

Also: You are using version chaining, which is to be avoided.
You propagate from DNG to TIF and JPEG, but also from TIF to JPEG. This may have unexpected consequences.

That is not such an uncommon thing. I may have raw files from which I derive TIFFs, JPGs, etc. One of the TIFFs may be a stitched panorama from wich I may then derive JPG versions. Just one example. Or I derive JPGs for different uses from a TIFF or PSD which is derived from a raw file.

[Mario]

Version chaining is not explicitly disabled in IMatch. But if you use such complex setups you have to think very careful about how all this will work in the end.
I cannot handle anticipate or support every potential edge case, all possible variations users can come up with. If it works for you, great. Keeping multi-level chained versions separated in different folders is helpful.

[JohnZeman]

Sorry, I didn't explain it the best last time.

All of my original master files are either DNGs or TIFs.
All of my versions are JPGs.
So the DNGs propagate to JPGs with the same name, while the TIFs propagate to other JPGs that have the same name as the TIFs.

In other words I have one master (DNG or TIF) and one JPG version in the folder, 2 examples follow.

2000_0115_080000_000.tif
2000_0115_080000_000.tif.dop
2000_0115_080000_000.jpg

2019_1216_155231_722.dng
2019_1216_155231_722.dng.dop
2019_1216_155231_722.jpg

I've checked that the versions are found in the preferences Test Expressions section and IMatch always correctly matches up the masters to the versions when I add new images to my database.  Plus a manual propagation (F4,P) always propagates the metadata, it just doesn't happen when I change the metadata in the master then do a writeback.

Since no one else seems to have this problem I'll keep checking things here, it must be something on my end.

Thanks guys

[Mario]

I've tried it twice. Even setup DNG->JPEG and TIF-JPG relations, copied some sample files together etc.
Works here like a charm. Always.

Maybe try to isolate things. Create a new folder. Copy a DNG, JPEG into it. Make sure the version is detected. Change the description for the DNG. Write back.
The ExifTool output panel will list the .jpg file in some of the commands (where it copies from the master to the version after writing the master).

There must be a difference in setup or file layout / condition. Finding it should help solving this.