CRW + THM in help file vs metadata settings

[Ferdinand]

Before I file a change for the help file using the link at the bottom of the page, I wanted to check that my understanding is correct.

In the help page "Metadata Storage", under "Canon CRW and THM Files" it says "If you look at the CRW file in the metadata browser you see the data from both files [CRW + THM] merged. When IMatch writes back data to CRW files, it automatically updates the EXIF/IPTC data in the THM and the XMP data in both files". 

Yet when I go to Metadata2 preferences and look at the settings for CRW,  the default is not to write IPTC.  Can I take it that the help file was not updated after those big changes to file formats in Metadata2?

[Mario]

I don't know. THM are so old, I don't check on them with every build.
Did you try it? Does the THM get updated? Error messages? What does the ExifTool output panel show?

I can do these tests for you but this will take a while.

[Ferdinand]

Well, the help file says one thing and the metadata2 prefs say something else.

I did a test and I think the help file is wrong and the prefs are right.  That is, I enter something in City for the CRW and execute writeback and the CRW is unchanged (correct), the THM gets XMP (correct) and the thumb does not get IPTC (correct according to the prefs, but not to the help file).

I will submit a help file change.

I know THMs are a pain, but there are a lot out there a lot out there in archives and we're stuck with them.

[Mario]

Why do you need THM? This was a clutch from Canon to allow applications 10 years ago to display thumbnails for their CRW format. Since then lots has changed and THM files are no longer needed. They may contain IPTC/EXIF data, but IMatch imports most of it into the XMP record so maybe the XMP route can rid you of THM files altogether.

THM was always a mess and a kludge. I had to implement special cases all over in order to support them, and I still regret doing it. Getting rid of the "drag THM along" code would be a good thing in my book...

[clpratt]

AVI files from the Canon cameras I use have an accompanying THM file which carries the EXIF Date, Time, Exposure, Camera model information etc.
As far as I know there is no such information in the AVI file and no way of inserting it as EXIF or XMP data, so the THM is definitely needed.

[Mario]

Why not use the ExifTool Command Processor to copy the metadata in your THM files into an XMP file, and get rid of the THM alltogether? This needs to be done only once and then all your metadata will be in a modern metadata format (XMP). You can then throw away the THM files, unless Windows is unable to produce a thumbnail for your AVI files IMatch can use. In this case the THM may serve as a visual proxy but not as a hard-coded data source/target IMatch somehow has to drag along...

[Ferdinand]

Why do you need THM? This was a clutch from Canon to allow applications 10 years ago to display thumbnails for their CRW format. Since then lots has changed and THM files are no longer needed.

You may well  be right Mario, I'll give it some thought.

When I started using IMatch ten years ago I was shooting CRW+THM, and in addition to the thumbnail the THM was also a store of IPTC.  V3.6 handled the substitution of the THM for the CRW seamlessly, so the whole workflow was smooth, at least from my point of view.  Remember that this was in the pre-XMP era.  Even when XMP arrived, IMatch 3.6 was still very IPTC centric.  An XMP workflow was not easy whereas an IPTC workflow using the THM was simple. 

At the moment I am doing my utmost to migrate my production image collection to V5.  I'm migrating certain categories to keywords, removing the keywords from IPTC for RAW files, removing XMP from RAW files, and generally cleaning up.   So I have to deal with issues like this.  I've got to ensure I strip these things out of the THMs.  I may then remove them and back them up somewhere, but I'm going to clean them up first.

Question:  Do you recall if there is any EXIF in the THM which cannot be obtained from the CRW? 

[Mario]

Question:  Do you recall if there is any EXIF in the THM which cannot be obtained from the CRW? 

When I recall it correctly, CRW files contain only proprietary Canon maker notes, some of which duplicate certain EXIF fields. This has changed several times, depending on the CRW format variant used. 'Standard' EXIF is only in the THM file. And legacy IPTC data.

IMatch 5 merges all that to form the input stream for XMP and to do the IPTC/EXIF/GPS => XMP mapping on import.

[Ferdinand]

Question:  Do you recall if there is any EXIF in the THM which cannot be obtained from the CRW? 

When I recall it correctly, CRW files contain only proprietary Canon maker notes, some of which duplicate certain EXIF fields.

That's what it looks like to me too.  The problem is that without doing a forensic analysis, it's hard to be certain that in discarding the THM you're not discarding the only source of some piece of metadata.  The XMP doesn't store everything, as I understand it.

[Mario]

You can always look at the XMP data and check if it contains what you need. Most of the Canon maker notes is rubbish anyway and does neither fit in EXIF nor XMP. I'm not sure how important this data is for you, for such old files. XMP has no notion of maker notes (which is positive) and thus cannot import maker notes.