TYPO3 Core

TYPO3 can be configured that pdf files are "image files", so that a thumbnail is generated for them
using imagemagick. But during meta data extraction width and height of pdf files are not extracted
because the mime type of pdf files is not image/* but the indexer checks for this type.

Instead the isImage method on the file should be used to trigger extraction of width and height.
imagemagick can do this with the identify command.

While the exact pixel values do not matter much for a PDF, the ratio of width and height (langscape or protrait or even square) can be very useful to "know".