Question

I have a user with a collection of JPEG files, rest on a Windows 7 NTFS file system.

The user has annotated these images using the Windows 7 “Details” dialog. For each file they have added a title, a comment (which maybe different from the title) and a tags field.

It appears that that the title is stored inside each JPEG file as part of the EXIF data.

However the Comments and Tags fields are not.

I’ve looked for them in Alternate Data Streams; there are none present. I can’t see whether they’re NTFS Extended Attributes because I can’t find a tool which lists those specifically.

My question is: where is this metadata stored, and how can I access it from within Ruby?

Answer 1

Eric’s answer comes close (thanks Eric).

The answer is that Windows uses XMP embedded in the file to store some this information. Some other parts are stored in EXIF.

XMP is an RDF format first developed by Adobe.

However, the support for extracting and manipulating XMP in Ruby is poor — there are some libraries on github, but nothing very well integrated.

• xmpleton doesn’t show all the metadata;
• xmp crashes on Microsoft’s metadata;
• fasttrack is not integrated into the rubygem.org infrastructure at large.

Of the three fields I’m interested in:

• “Comments” is stored in an EXIF field “XPComment”.
• “Title” is stored in EXIF under the field “XPTitle” and in XMP under the DC namespace in the field “Description”
• “Tags” are stored in EXIF under the field “XPKeyword” and in XMP in the DC name space as a list of “Subject” fields.

The approach I’m using is mini_exiftool, which wraps the Perl-based EXIFTool program.