DNG metadata

DieRipps

Andrey
Hi Andrey, I understood. So I can explain my workflow.
I use Adobe LC for photo management. All photos are in RAW DNG.
I have an addon for tagging stuff automatically, but not faces. Currently I use the LR face recognition to tag faces. After the adjustments I use the LR publishing service to push the photos (as jpeg) to my Synology Photos. In the past a was using also the Synology Photos face recognition tool. (the result (it was a tag with the name) was synced back to LR). The problem here, the quality was not so good and it was a lot of work to assign faces in the app.
My hope is now that tonfotos, which is great in tagging faces, can help me to support my workflow.
Maybe I have my raw files first tagged in tonfotos and somehow I need the result (maybe as a tag) available in LR. From there my workflow continues like described and finally I write all tags int he jpeg. Synology Photos is able to display and filter along these tags.
What do you think?

Andrey

DieRipps I am not sure I totally understand the end goal here. It makes sense to me that in the end you are browsing your collection using Synology Photos. And if it can read embedded face tags from XMP as you say, then you can totally use Tonfotos for facail tagging, store results into metadata, and SP will pick that up, and you are all set for browsing your collection from the web.

However, I am not sure why you are trying to involve LR into facail recognition workflow, obviously it is not the best tool for that. It is great with RAW processing, but as far as I understand you can just ignore its facial recognition feature alltogether. Why not just doing RAW development in LR and then facial recognition in Tonfotos on developed photots? Why you try to pull facial tags back to LR? I guess I am missing some part of picture here, please comment.

DieRipps

Andrey
Hi Andrey, so first thanks for helping me to find a good solution.
So my situation is that currently LR is the the tool where I have the single point of truth. I want to have all information in LR. That means, all Photos in RAW, all adjustments, all tags.
From there I do the sync (exports) to other platform. I need all the tags in LR for the export, to have that info also in other tools, like Synology Photo.
My workaround was to tag the person in Synology Photo and sync the info back to LR as a tag (to keep the single point of truth alive). But this was a lot manual effort. And yes in LR the face / person recognition is not so good, but currently I have no better option. When tonfotos can help here, I’ll go all in with it ;-). Requirement is that I have the result available as a tag in LR for my RAW picture in the library before the export. Is it helping you to understand what I need ?

Andrey

DieRipps I guess it helps to understand, but it does not make this task any easier. As I said, there are several dozens of different RAW formats, and some of them even not capable of extended metadata. However, if LR can read sidecar XML, that can potentially be a solution. This needs additional testing.

DieRipps

Andrey
Hi Andrey, LR is able to create a .XMP sitecar file (kind of XML) to store all changes for RAW editing there. With that it’s possible to open the RAW picture with any other RAW editing tool and you were able to see the modifications you did in LR.

In that XMP file structure, there are several places you could insert the face recognition info. E.g. to make LR aware there is a face in the picture you can use this structure (but this is only for LR itself):

<mwg-rs:RegionList> <rdf:Bag> <rdf:li> <rdf:Description mwg-rs:Rotation="1.57080" mwg-rs:Name="Name 1" mwg-rs:Type="Face"> <mwg-rs:Area stArea:h="0.06305" stArea:w="0.04199" stArea:x="0.59033" stArea:y="0.47434"/> </rdf:Description> </rdf:li> <rdf:li> <rdf:Description mwg-rs:Rotation="1.57080" mwg-rs:Name="Name 2" mwg-rs:Type="Face"> <mwg-rs:Area stArea:h="0.06305" stArea:w="0.06199" stArea:x="0.89033" stArea:y="0.67434"/> </rdf:Description> </rdf:li>
So you have a list item for each face.

In addition you can add a tag with the person name. You can use this one as a filter in each other photo software which is able to read tags.
So maybe tonfotos is able to insert the face info in two different places later on.
Is this a way which is possible for to implement? We can check details later… if you think let’s do it.

e.g. for Synology Photos to be able to see the tagged face following path is used:

<dc:subject> <rdf:Bag> <rdf:li>Person 1</rdf:li> <rdf:li>Imported Tags</rdf:li> <rdf:li>Photo StatLr</rdf:li> <rdf:li>Synology Sync</rdf:li> </rdf:Bag> </dc:subject> <lr:weightedFlatSubject> <rdf:Bag> <rdf:li>Person 1</rdf:li> <rdf:li>Synology Sync</rdf:li> <rdf:li>Imported Tags</rdf:li> <rdf:li>Photo StatLr</rdf:li> </rdf:Bag> </lr:weightedFlatSubject> <lr:hierarchicalSubject> <rdf:Bag> <rdf:li>Photo StatLr|Imported Tags|Synology Sync|Person 1</rdf:li> </rdf:Bag> </lr:hierarchicalSubject>

Andrey

DieRipps Thank you for confirming LR is working with sidecar files. I thought that might be the case, I just was not 100% sure as I don’t use it myself.

Yes, RegionsList is most common place to put facail information, and this is where Tonfotos stores it. However, I am surprised Synology relies on Subject instead. Are you sure they cannot read face regions? This is not very smart of them…

DieRipps

Andrey
Yes Synology is using a own database for metadata and face recognition. That’s not good. But! there is this workaround using tags.

Regarding Lightroom:
Do you think it’s possible from tonfotos side to provide (write) the information into the XMP file.
Lightroom should be able to read that and the single point of truth is still LR.

Regarding Synology Photo…
Maybe this is also possible for tonfotos to provide an additional option to write the face/person info into the XMP like with the structure I have provided. You would make a lot of Synology Photos user happy ;-) (me included)
For other tools out there like Immich, I guess those are able to read the LR XMP info for face recognition.

Would be great, what do you think?
And I’m open to test of course…

Andrey

DieRipps Regarding Lightroom:
Do you think it’s possible from tonfotos side to provide (write) the information into the XMP file.
Lightroom should be able to read that and the single point of truth is still LR.

Yes, that is definitely doable and reasonable thing to do. I have added this to our backlog.

DieRipps Maybe this is also possible for tonfotos to provide an additional option to write the face/person info into the XMP like with the structure I have provided. You would make a lot of Synology Photos user happy ;-)

Improving compatibility with Synology Photos is potentially interesting, however I am afraid it is a bit more complicated than that. There at least several generations of that app, and more detailed research is needed first to establish what exactly it does understand. I am not sure I have resources to dedicate to that task. However if you are ready to do some research, that would definitely help.

DieRipps

Andrey
Hi Andrey,
let’s start with the first feature. When it’s done from your end, let me know. I’ll buy tonfotos immediately and will test it.

For the other feature, regarding Synology Photos, I can of course support testing!

DieRipps

Hello Andrey , any news, when the team is working on it?
Cheers

Andrey

DieRipps I’m sorry, I cannot disclouse our internal plans per our policy. So far this is not implemented.

Tenuki

I’m also interested in this approach. LrC is my main RAW editor and photo album manager (somewhere around 30k photos), but the facial recognition is terrible. Tonfotos facial recognition is great (testing the free version) but can’t export the facial metadatas to LrC.

Sign me up for the paid version if it allows me to have Tonfotos recognised faces inside LrC.

LaurentR

+1 for interest in writing people information in (existing or newly created) XMP sidecar files for RAW images.

I use LrC and ACDSee and both support sidecar XMP files.