You’re sure to come across metadata at some point while editing images. So here’s a summary of some technical stuff:
Information attached to an image or file, just like we used to write names, places or dates on the back of a photograph.
Types of metadata
EXIF – Usually written by the camera, mostly about the capture conditions like the shutter speed, date taken or camera settings.
IPTC-IIM – Mostly about the contents of an image, ratings, keywords, ownership and rights.
XMP – The latest way to store metadata, XMP was created to overcomes some limitations of the IIM and EXIF data format. The XMP block can include just about anything including EXIF or IPTC fields, it’s often used to record proprietary information, such as the adjustments made to an image using photo editing software like Lightroom.
Where is metadata?
An image file contains several blocks of information: EXIF, IIM, XMP and the IMAGE data itself, and to make things even more confusing, the XMP block can be stored in a separate file which sits next to the image, rather than in the image itself. These sidecar files have the extension .xmp and if you open them with a text editor you’ll see they usually contain a list of settings and other information.
Previously, the IPTC data like keywords would be found in the IPTC-IIM block, but that’s now considered out dated, because the 90’s were just so long ago :). IPTC data should now be stored in the XMP block, or preferably both for maximum compatibility, hopefully your software knows this.
Since some software doesn’t doesn’t read every block, or may prioritise one block over the other, sometimes metadata can appear to change or disappear because its been written to a different block than the one your software is displaying.
One example, apply a star rating to an image in Bridge, and then change the rating using Capture One. Bridge won’t recognise the Capture One rating because it’s been applied in a sidecar file which Bridge won’t read. What you see can also depend on the preferences used in either application. This sort of thing happens across a lot of different applications, but doesn’t mean they incompliant with the standards, they just use them differently, and perhaps in ways that aren’t the most suitable.
The main point here is to be aware that metadata may not be as straight forward as you think, so when strange things happen you’ll need to dig deeper.
Hopefully my explanations are technically correct enough! During the course of writing this article I’ve had some correspondence with Carl Siebert, a real expert in all things metadata, find him here if you want to dig deeper. https://www.carlseibert.com.