Thursday, December 03, 2009

Are Social Media Websites Preserving Your Photo Metadata?


The members of the Controlled Vocabulary forum are attempting to answer the question above (do social media websites or other image sharing services you use preserve your embedded photo metadata after upload?) by conducting a survey of the various services.

When Social Media websites or other online services intentionally or accidentally remove metadata embedded in the images you upload, they are violating the first two principles put forward in the Metadata Manifesto namely; 1) Metadata is essential to identify and track digital images; and 2) Ownership metadata must never be removed.

Any automated system for managing or displaying digital files should honor these principles. At minimum, these systems need to preserve ownership metadata by default and discourage removal of other metadata by warning users about the legal implications of removal.

It seems hard to imagine that simply by uploading an image to one of these websites that all the metadata within a file would disappear. Especially since there is no warning or notification given. However, the preliminary results of this Controlled Vocabulary Survey regarding the Preservation of Photo Metadata by Social Media Websites make it look like the people behind these services have little awareness of the importance of preserving embedded photo metadata.

If you don't find your favorite service in the list of preliminary results, there are instructions on the social media photo metadata preservation survey web page about how to participate. Simply download the testbed file seen above, and upload it to the service you want to survey. Instructions on how to use a simple online tool to view the metadata are given on the page above. You should see sections outlining all of the IPTC, XMP, and Exif metadata fields contained in the image.

However it appears that even the Blogger platform suffers from the same issue this survey hopes to point out. If you investigate you will find that even the image that was resized to 276 pixels wide (and then uploaded to accompany this post) has had all of it's XMP metadata removed!

It looks like you have to upload the image to your own website and then reference that image in your blog post if you are serious about preserving all of your embedded photo metadata.

4 Comments:

Blogger stanford said...

Thanks for confirming what I already suspected about blogger... I thought maybe I was missing something. Nope.

8:13 PM  
Blogger Farennikov said...

I found your post while trying to research pros and cons of keeping embedded metadata in images. I'm working on optimizing 3 high traffic sites where images make a large portion of content. And a little test shows that after conversion 40 thumbnail images to JPG's (at 75% quality) results in 700 KB total size with metadata, and 260 KB total size without metadata -- thus metadata taking over 60% of file size. I think it's too much and am considering removing it from thumbnails. What would you say about that?

1:54 PM  
Blogger David Riecks said...

Farennikov:

I'd first ask how you are measuring these differences. If you are looking at a local drive, the true size may be obscured by the sector size of the hard drive, as well as other hidden files (especially if on a mac). If you are looking at the files on a web server, then those hidden files (resource fork data) shouldn't be part of what is being measured.

Second, I'd ask what you are using to embed photo metadata, and what fields you are including. Are you embedding both legacy IPTC (IPTC-IIM), or both IPTC-IIM and XMP?

If you aren't using any XMP specific fields, you might see what you get by only retaining the older binary form of IPTC metadata as it's more compact.

Lastly, if these are "thumbnails" only, then you might consider retaining the key data, such as Creator, Copyright Notice, Source, Credit, and Title (to hold the original image filename). Those fields shouldn't add more than 2 or 3 kilobytes to each image file.

Hope that helps.

David

2:10 PM  
Blogger geetha said...

Thanks for sharing this useful info. Keep updating same way for Adobe day CQ5.

Regards,Siddu Corporate Training

5:50 AM  

Post a Comment

<< Home