Open main menu
Archive This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.

I can't copy EXIF data to a WebM video.

I've taken some MP4 videos using my phone and my camera. They include some useful-seeming information, as seen by running `exiftool MVI_0156.MP4`, for example:

   Make                            : Canon
   Camera Model Name               : Canon EOS 750D
   ...
   Drive Mode                      : Continuous Shooting
   Image Size                      : 1920x1080
   Lens                            : 70.0 - 300.0 mm
   Lens ID                         : Canon EF 70-300mm f/4-5.6 IS USM
   Rotation                        : 0
   Scale Factor To 35 mm Equivalent: 1.6
   Shooting Mode                   : Program AE
   Shutter Speed                   : 0
   Create Date                     : 2016:03:09 10:19:45.76
   Date/Time Original              : 2016:03:09 10:19:45.76
   Modify Date                     : 2016:03:09 10:19:45.76
   Circle Of Confusion             : 0.019 mm
   Field Of View                   : 6.8 deg
   Focal Length                    : 190.0 mm (35 mm equivalent: 301.8 mm)
   Hyperfocal Distance             : inf m
   Lens                            : 70.0 - 300.0 mm (35 mm equivalent: 111.2 - 476.5 mm)

However, when I convert to WebM, this information is stripped. So I tried to copy it over, and got this:

   $ exiftool Corvid\ in\ tree\ calling.gk.webm -tagsFromFile MVI_0156.MP4 
   Error: Writing of WEBM files is not yet supported - Corvid in tree calling.gk.webm

Is there a common textual format to stash EXIF data in for videos while ExifTool gets the power to do this? Is this just a shortcoming of WebM, and it'll never work? ("not yet supported" seems to imply that's not the case.) What's the right thing to do here? grendel|khan 20:05, 9 March 2016 (UTC)

PDF thumbnails are incomplete.

I've recently uploaded a couple of PDF posters; I think they're actually vector images--extracting the raster images with `pdfimages` didn't yield anything, and the SVG conversion tools I tried out didn't work so well either--so this seemed like the right format to use. Some of them thumbnailed just fine (for example, File:DOE-LPO Illustrated-Posters 01 CSP Full.pdf), but some of them (File:DOE-LPO Illustrated-Posters 02 ATVM Full.pdf, File:DOE-LPO Illustrated-Posters 03 ECON Full.pdf, File:DOE-LPO Illustrated-Posters 05 WIND Full.pdf, File:DOE-LPO Illustrated-Posters 06 GEO Full.pdf, and File:DOE-LPO Illustrated-Posters 08 NCLR Full.pdf) didn't. The uploaded files are the same size and `sha1sum` equivalently to the originals Is this a limitation that there's an open bug about? Should I file one? grendel|khan 19:23, 9 March 2016 (UTC)

These files are very large with lots of details. I suspect the rendering server for JPG breaks the job. I think contacting server admins about rethinking limits is advicible here hence. I think the bug above matches to your issue. Rastering with GIMP works well so you might provide prerendered PNG as second source if limits stay the same. -- Menner (talk) 20:30, 10 March 2016 (UTC)