Redundant tags

[ww408]

Downloaded photos from a science gallery and saw that the caption was copied into three different tags. What is your opinion about that practice? This is ChatGPT's opinion:

It's not necessarily overkill, but you're correct that including multiple metadata standards for the same information can feel redundant and might create extra work when managing files. The intention behind embedding multiple metadata tags (e.g., IPTC, XMP, Exif) is often to ensure maximum compatibility across various platforms, software, and users, especially when the file might be used in different contexts or by different types of applications.


Why Multiple Metadata Tags Might Be Used:

1. Cross-Platform Compatibility:

• Different applications and systems prioritize different types of metadata.

• IPTC is widely supported in image management tools and photo editors (especially for journalism and photography workflows).

• XMP (including Dublin Core) is often preferred in cross-media environments and is used in Adobe tools, web platforms, and archival systems.

• Exif is generally used for camera-specific metadata and may be recognized by basic image viewers.


2. Ensuring Access to Metadata Across Software:

• Different apps may support or display different metadata formats. Some software might only read IPTC tags, while others may favor XMP or Exif.

• By embedding multiple tags, the creators may have wanted to ensure compatibility regardless of the software being used, especially in cases where users might open the file in an application that prioritizes a specific standard.


3. Purpose-Driven Use:

• IPTC is generally used for descriptions, captions, copyright, and creator data, so it's highly useful in publishing and journalism.

• XMP is more versatile and allows embedding of complex metadata for broader applications (like Dublin Core).

• Exif is mainly used for camera settings and technical image data but is still used for some descriptive fields.


Is It Overkill?

It depends on your perspective:

• If you're working in a specific workflow (e.g., photojournalism or content creation), it's understandable to include all three to ensure that the metadata is universally accessible across platforms.

• From a simplicity standpoint, though, it could be seen as redundant. Many image tools will read IPTC and XMP equally, so embedding both might feel like extra work with limited practical benefit.


Alternatives or Best Practices:

• Minimalist Approach: If you're only working with one specific software or platform, it's often best to use the metadata standard most compatible with that platform. For instance:

• If you're focusing on image editing and management, IPTC may be all you need.

• If you're aiming for cross-media compatibility, XMP might be the best option.
 
• Keep It Simple: If your goal is just ensuring basic compatibility, you don't always need all of them. Prioritize the one most relevant to your use case.
 
• Automating Metadata Entry: If you're concerned about embedding multiple metadata standards, you can use tools like ExifTool to automate the process, ensuring the metadata stays consistent across formats without manual tagging each time.


Conclusion:

While it's possible to use multiple metadata standards, it's important to evaluate your specific needs and the compatibility of the software or platforms you're targeting.

[StarGeek]

The best starting place is the Metadata Working Group's Guidelines For Handling Image Metadata. Specifically, section 4 "Metadata Management". Even though the MWG is defunct, this standard is still followed for the most part by the better Digital Asset Management (DAM) programs.

The IPTC Photo Metadata Standard is also worth reading, though I don't think it gives any specific advice on whether to fill multiple tags or not.

The MWG standard suggests that data should be written to the XMP and EXIF groups. The IPTC IIM tags should only be written if the IPTC block already exists.

In the case of the description (caption), the MWG says that it should be written to the XMP:Description, EXIF:ImageDescription, and, if the IPTC block exists, the IPTC:Caption-Abstract.

Now, my opinion. Most of the EXIF tags are set by the camera. There's only a few of these that I will write to. Those being the GPS tags, EXIF:Copyright, and EXIF:Artist (which maps to the Creator of the image). I don't write to the EXIF:ImageDescription because I haven't seen much support for it in a lot of programs, though I think Adobe programs will write it. The other two tags have much better support. There's not much reason to write it when not many programs will specifically read it.

Writing the IPTC tags is a more nuanced situation. Technically, if you want maximum compatibility, then you would want to write to them. But the IPTC standard is a really old standard with some problems. It's not very expandable and the tags have limits on how long they are supposed to be, though many programs will ignore that limitiation. Exiftool will enforce those limits unless the -m (-ignoreMinorErrors) option is used.

IMO, if at all possible, you should avoid using IPTC tags, as it can lead to headaches trying to keep them in sync with the corrisponding XMP tags. Only really old programs are likely to need them.

[ww408]

Those being the GPS tags, EXIF:Copyright, and EXIF:Artist (which maps to the Creator of the image). I don't write to the EXIF:ImageDescription because I haven't seen much support for it in a lot of programs, though I think Adobe programs will write it. The other two tags have much better support.

So different tags in the same group have different levels of compatibility? Can someone make a list of all tags, graded with a compatibility rating, like 1–5 stars? ⭐


This is a CSV file that ChatGPT made for me based on the IPTC Photo Metadata Standard page. Supposedly, it's every tag mentioned in the document, organized by group, and it matches corresponding tags from different groups to each other. Am I to assume that all tags mentioned in that document are ⭐⭐⭐⭐⭐?

This is one of the images in question, by the way (redundant and IPTCDigest warning): https://www.sciencephoto.com/media/14094/view/ergot-fungus-sem  Download is blocked for me, but one can easily get around that by simply saving the page as Webpage, Complete and then accessing the resulting "_files" folder, which contains the image.

Code Select Expand

Primary Tag,Matching Tags,Comment
XMP-dc:Source,XMP-photoshop:Source,Exact match: Both define the source of the file.
XMP-dc:Creator,XMP-photoshop:AuthorsPosition,Exact match: Both describe the creator or author role.
XMP-dc:Rights,XMP-photoshop:Credit,Exact match: Rights and credit to the creator or copyright holder.
XMP-dc:Title,XMP-photoshop:Headline,Exact match: Both describe the title or headline of the file.
XMP-dc:Subject,XMP-iptcCore:Scene,Potential match: Subject keywords vs. scene descriptions.
XMP-iptcCore:CountryCode,XMP-photoshop:Country,Potential match: Country code vs. country name.
XMP-dc:Description,XMP-photoshop:CaptionWriter,Potential match: Description text vs. caption author.
XMP-photoshop:Instructions,XMP-iptcCore:IntellectualGenre,Potential match: Instructions vs. intellectual genre.
XMP-iptcCore:AltTextAccessibility,,Used for accessibility text descriptions.
XMP-photoshop:City,XMP-iptcCore:Location,Potential match: City vs. specific sub-location.
XMP-dc:Title,XMP-iptcCore:Location,Potential match: General title vs. specific location.
XMP-photoshop:Country,XMP-iptcCore:CountryCode,Exact match: Name and code for the country.
XMP-photoshop:DateCreated,XMP-iptcCore:DateCreated,Exact match: The date the file was created.
XMP-iptcExt:Event,XMP-iptcCore:Scene,Potential match: Event name vs. depicted scene.
XMP-iptcCore:Scene,XMP-dc:Subject,Potential match: Scene details vs. subject keywords.
XMP-dc:Rights,XMP-xmpRights:UsageTerms,Exact match: Terms for usage rights and copyright.
XMP-xmpRights:WebStatement,XMP-iptcExt:DigitalSourceType,Potential match: Web statement vs. source type.
XMP-iptcExt:LocationCreated,XMP-photoshop:City,Potential match: Created location vs. city details.
XMP-iptcExt:PersonInImage,,Describes people depicted in the resource.
XMP-iptcExt:PersonInImageWDetails,,Structured details of people in the resource.
XMP-iptcExt:ArtworkOrObject,,Details about artwork or objects depicted.
XMP-iptcExt:Genre,XMP-iptcCore:IntellectualGenre,Exact match: Genre of the intellectual content.
XMP-iptcCore:IntellectualGenre,,Defines the genre of the intellectual work.
XMP-photoshop:Headline,XMP-dc:Title,Exact match: Title vs. headline representation.
XMP-photoshop:TransmissionReference,XMP-dc:Source,Potential match: Source vs. transmission reference.
XMP-plus:ImageSupplier,XMP-plus:CopyrightOwner,Potential match: Supplier vs. copyright holder.
XMP-plus:Licensor,XMP-plus:ImageSupplier,Potential match: Licensor of rights vs. supplier.
XMP-dc:Subject,XMP-iptcExt:Scene,Potential match: Subject vs. scene representation.
XMP-plus:PropertyReleaseID,XMP-plus:ModelReleaseID,Potential match: Property vs. model release identifiers.
XMP-iptcCore:CreatorContactInfo,XMP-photoshop:AuthorsPosition,Potential match: Creator contact vs. author role.
XMP-xmp:Rating,,Defines the rating or score of the resource.
XMP-dc:Description,XMP-photoshop:CaptionWriter,Potential match: General description vs. caption writer.

[StarGeek]

I don't have time at the moment to get into details, but I can instantly see a bunch of inaccuracies in it. For example Title and Headline are very different.

The IPTC standard doesn't match XMP tags to other XMP tags. There can be some relation, such as Credit and Creator, but they are different tags.

Another example is City vs Location. An example of that would be New York City would be in the City tag, and Location is a more specific place in the city, such as "Madison Square Garden" or "Statue of Liberty". A neighborhood of the city would also be appropriate, such as a specific borough in NYC.

As for a list, use the IPTC standard tags. Except for the GPS tags, the EXIF tags can be left alone unless they need editing due to a mistake, such as forgetting to shift the time between Summer/Fall.