Link This | Email this | Blog This | Comments (1)

XMP files make me crazy!
May 2, 2008

When you modify an image using Adobe Bridge or Photoshop, those applications create “sidecar” files called XMP files, which carry the information about the image, some of the modifications you have made, and metadata that you have added. One of the most important things that photographers do with images is to embed their copyright information and captioning. Called IPTC data, this information assists database programs and asset management systems in sorting, filing, and creating valuable search criteria for the image.

In version CS2 of Photoshop and Bridge, Adobe created these sidecar files only with Camera Raw files. In Photoshop, Tiff and JPEG images, that version of the software wrote the metadata to the file itself. In version CS3, Adobe creates the XMP files for every image type except PNG, their open digital negative file format (more on that in a future blog).


The IPTC data, which can include such items as caption, copyright, photographer contact information, key words, genre, and more is used by photo agencies (the IPTC is a press photography association). Search engines can then select images based on these data.

The problem with XMP files is that they are often (usually) left behind when photos are transferred from an agency to the client, or from the agency to a publication. When this happens, the copyright and caption information is lost and the photo continues on its way without the intellectual property claims and information needed for subsequent use. Adobe, ironically, claims that the XMP files are a “feature” of Photoshop and Bridge, and on that point I most emphatically disagree.

I understand that all of the IPTC information cannot be recorded in the photo file itself. But, the copyright holder’s name and contact information, caption and key words most certainly can be embedded, and I think it’s a mistake for Adobe not to do that.

Sophisticated asset management programs like Canto Cumulus and Extensis Portfolio, among others, read the XMP files and add the information in those files to their catalogs, but if the XMP is missing, they read nothing and they catalog nothing. This is a mess, and Adobe has made it worse from version CS2 to CS3 by making the XP3 files ubiquitous.

One solution to the problem is to use a different program to embed keywords, copyright claims and captioning. That program is Photo Mechanic, and it’s available from Camera Bits in Oregon. Photo Mechanic will embed most (but not all) of the information needed by photographers into images including Camera Raw. Photoshop could/would never do that.

My work flow with digital images includes bringing photos from Compact Flash cards into the computer (Image Capture), renaming and sorting (Adobe Bridge), embedding copyright and keyword strings (Photo Mechanic), and retouching and editing (Adobe Photoshop). I could streamline the process by using Photo Mechanic to import the images and rename them, but I like the user interface of Bridge better for that.

This XMP file loss problem can be a problem for those in premedia because images can arrive without origin, copyright information or contact information. Sometimes those bits of information are critical to the legal and ethical handling of images for asset management systems and print production.

Posted by Brian Lawler on May 2, 2008 | Comments (1)