I started writing this blog post after following up on Dreamstime’s excellent new DMCA-violation reporting tool they added to their system. I went and read the DMCA and followed up by reading various interpretations of the clauses within it, and subsequently drafted this post and sent out to every microstock agency I have a contact with. Honestly, it has been quite a journey, and I think I’ve significantly annoyed various agency ‘policymakers’ along the way. That said, it was entirely worth it to get ‘some’ clarity on this issue since every time I’ve met with photographers we’ve ended up discussing this very matter.
Although I would have liked this post to have been the ‘last word’ on this topic, that doesn’t appear to exist, and until someone really tests these matters in a court of law, I don’t think we’ll ever have the last word. In the meantime, here’s my journey
This is quite a long and detailed post before I get to the conclusion, and most of it is very important to photographers, so before we get into the nitty-gritties; here’s the two-cent tour:
Before I get into this any deeper, and because I REALLY do not want anyone to try and sue me for misrepresentation, slander/defamation, or anything else (as has been suggested they might)… a VERY important disclaimer:
The basis in law to indicate infringement:
There are two main legal documents we are looking at here;
Firstly, the InfoSoc Directive (European Directive 2001/29/EC)
It has been known for some time this law prohibit the removal of any Copyright metadata (including both watermarks and embedded IPTC/Exif/Xmp/other metadata) from any digital media. This has however been largely ignored because in Europe, member nations are required to pass their own national laws to uphold and penalise for the violation of European directives. The EC is prosecuting some nations for not implementing this directive, but the dust has not settled yet.
Despite this lack of penalties in many member states, the directive stands as european law. The relevant clause (Article 7) reads:
You can read the full text here. It is of course dry legal speak, but it’s very clear that the removal of ANY data provided by the rightholders to identify the work is illegal. Since many European nations have no penalties for violation of this act, the agencies have (as far as I can tell) universally ignored this directive.
Secondly, the DMCA (Digital Millenium Copyright Act):
From wikipedia: [the DMCA] criminalizes production and dissemination of technology, devices, or services intended to circumvent measures (commonly known as digital rights management or DRM) that control access to copyrighted works. It also criminalizes the act of circumventing an access control, whether or not there is actual infringement of copyright itself.
The DMCA is almost universally recognised online, and many agencies are familiar with DMCA takedown requests (it was the addition of an automated platform for handling them released by Dreamstime that put me onto research these laws further).
The relevant clause in the DMCA (Section 1202(b)) reads:
Read the full text (pdf) here. In the DMCA it is made even more clear than in InfoSoc, that to intentionally remove or alter any copyright management information, or to use/distribute any digital work knowing copyright management information has been removed is breaking the law.
What constitutes copyright metadata information (CMI)?
This is a little unclear in it’s specifics in both InfoSoc and the DMCA. The bare minimum covered includes the specific copyright fields in all of Exif, IPTC AND Xmp formats. Also protected is certainly the artist, author and by-line fields of all 3 formats. More questionable is whether title, description and keywords fields are required to be retained. Title is commonly regarded as required with extended description and keywords fields (in any format) being more flexible. The makernotes and other ‘satellite’ data such as GPS or editing software flags are probably disregardable.
How might microstock agencies be violating these laws?
I’ve run tests on downloaded images from several microstock agencies and found none to contain ANY copyright metadata. Some retain title/description and keyword metadata in either Exif, IPTC or Xmp/rdf formats, and almost all still contain makernotes (the metadata embedded by the camera such as focal depth, shutter speed, etc).
Given my limited samples (at most 50 images from each agency due to limited funds/access) I wouldn’t normally expect that it be so unusual to have no copyright metadata fields filled, as it’s commonly believed that many photographers do not write copyright metadata to their images.
So I checked picWorkflow import stats and found that around 60% of images (by about 35% of uploaders) have copyright metadata filled at time of import (those which do not are automatically populated with their owner’s most commonly-used copyright metadata so 100% of images exported by picWorkflow have copyright metadata). This is enough for me to infer that somewhere between a quarter and a half of randomly selected and tested images should contain metadata.
The type of metadata remaining in some files also makes it clear that data was ‘intentionally’ stripped at worst (by software or code-libraries intended to manipulate image metadata) and ‘knowingly’ stripped at best (through poor-quality resize software or code-libraries).
This is also backed up by a complete lack of metadata on image thumbnails and previews at agency sites. There is however technical reasoning for this, and some provisions are granted for this in the DMCA and InfoSoc as these are agency-watermarked and often use a unique filename or display an image ID which can be used to trace the file to it’s owner. Providing they are not transmitted to third-parties for any use other than caching for performance/traffic reasons. This suggests for the most-part that the same process used to resize their photographer’s images is knowingly stripping metadata (The violation being in the for-client-download output).
I’ve also read through the terms and conditions and contributor contracts (where they could be obtained) of the agencies mentioned below and despite occasionally references to granting them the right to store and distribute photographers images, or to resize them and convert into alternate formats. There is no mention I could find of stock photographers granting agencies the right to remove copyright metadata. The response from Nancy Wolff (below) suggests the contracts may cover such removals but I’m not sure I agree that sweeping ‘modification’ statements would be applicable with regard to metadata for which specific conditions exist in law.
What about traditional agencies, are they safe?
I don’t have hundreds of dollars to purchase and test traditional agencies sold-files. If you have purchased any, please open them in photoshop or other software capable of viewing image metadata (I used Opanda’s IExif which reads Exif, IPTC and Xmp for my tests) and review their contents. I’d like to find out more here.
Could Microstock agencies be fined for such violations?
I first found the mention of fines for copyright metadata removal here, and since most legal penalties are per-infringment, I took a look into the details written in the DMCA and found (Section 1203(c)) does indeed state:
Which seems to indicate that any photographer who found their work to have been distributed to image-buyers without the copyright metadata which was applied to their image when it was uploaded to the agency, would be entirely within their rights to demand payment (of how much I am not qualified to state, but the upper limit is $25,000).
I also can’t say whether each sale (distribution without metadata), or each image (import/storage without metadata) would take precedence, I guess that would be upto the lawyers and judges involved to decide, but there have been many cases successfully prosecuted for both examples under the DMCA. In the ‘best’ case (for the agencies), paying even modest damages to a large number of photographers would indeed be a huge burden for any microstock agency to bear.
It’s been highly unlikely to happen so far solely because of most microstock photographer’s lack of knowledge on copyright (due to the immature marketplace) and the lack of interest for those top-level photographers (who are knowledgeable) to harm their agencies, but there ‘might’ be grounds for photographers to raise cases for damages in future.
Which microstock agencies might be violating these laws?
Every agency I’ve tested so far appears to some extent to strip (or fail to preserve) metadata.
To be clear, my investigations are limited to less than 15 puchases for each agency, and a bunch of google-image searches to find web-use images online other people have purchased (about 10-25 per agency), so these observations are not accusations, or proof, but do warrant extra investigation, and I’ve asked the agencies mentioned to provide their responses to these observations:
I’ve tried to test a sample of agencies across the microstock ‘spectrum’ to get some idea of the overrall picture (I also tested Cutcaster images, but with a very small sample-set so dropped them from above).
If I missed something, or if there’s an agency you’d like me to specifically check out. Leave a comment at the end and I’ll add them too.
Why wouldn’t microstock agencies include copyright metadata in the sold images?
There are two main reasons as I see it. One is technical, until a couple of years ago managing metadata on image files was very difficult on the backend-server. With IPTC, Exif AND Xmp metadata all doing the rounds, and requiring different libraries, common image resizing libraries in many languages and platforms, both free/open-source and commercial were very limited in what they could do, and extra libraries to specifically manage metadata were spotty-at-best.
Recently this has improved slightly with various backoffice suites for image-management, and libraries released by php developers, adobe and various other groups to make handling image metadata more reliable and consistent. The technical reason no-longer stands.
The second reason (though purely speculative) is simple business. If your customers can cut you out the middle and go direct to your photographers, then what is the agency there for? This simple reasoning may have meant that for new/growing agencies it was initially risky to share copyright/ownership with clients who might go direct to the source.
I’m not sure this is as much a risk anymore either, since in the maturing marketplace image-buyers are comfortable with their agencies and the remaining risk is in buyers searching for photographer’s image on other ‘cheaper’ agencies.
Why is this important NOW more than ever?
As a photographer myself, and as a business-owner with an interest in representing photographer’s interests, I am seeing more and more violations of our copyrights. Just whilst researching and writing this post I identified around 100 images stolen from agencies and used either outside of the license terms, or without any valid license at all (people using watermarked images is much more common than I realised).
With the increasing use of tools like PicScout, Tineye, and the amazing functionality of the new Google image search making finding duplicate images easier than ever before, my concern is that image-theft is becoming ever easier.
In addition to theft, with orphaned-work laws in the works all over the world, there are already millions of purchased images in use online with NO ownership metadata. Orphaned works laws will make the use of these works commonplace due to the ability of future-users to say “no copyright metadata, it must be free” and to get away with it scot-free even if photographers make follow-up claims after the use is discovered.
Without a proper and clear definition of photographer’s copyrights being retained for the entire life of the work, the cost to all photographers is significant.
What should agencies do about this problem?
Firstly, agencies MUST retain any copyright metadata in images which are uploaded with it embedded. This metadata must be transferred to any-size version of the image sold. Without doing this, agencies ARE (in my opinion) breaking the law.
So they are not caught in the ‘enabling others to violate’ clauses of the DMCA, agencies should also include in their client’s license terms that it violates the DMCA to knowingly remove copyright metadata from the source image. This is a little foggy when the client is dealing with derivative works. I’ll have to research that more.
Where images are uploaded to the agency without copyright metadata embedded, it is my recommendation that agencies ‘should’ embed at least basic information in the form “© year artistname – agencyname”. This would not only ensure the copyright is correctly attributed but also that image-buyers can find the creator of images they have purchased, if only on that own agency’s website.
What do I want from agencies in response?
Agencies will hopefully respond to this post by declaring their position on metadata-retention under DMCA either to me, or on their own blogs/forums. I will post any links I find or am sent by the agencies onto this post.
For those agencies mentioned above, it would be good to hear a firm committment from them on modifying their systems to retain all relevant metadata on purchased files. They should also mention a timeline for any changes to their systems to be implemented.
Ideally it would be great if agencies also commit to embedding copyright metadata where the artist omits it, but is of course not required by the law, so additional praise will go to those who do so
And what do the agencies have to say about it?
Responses are colour-coded, , .
In addition to comments made above, I received an additional and thorough response from Dreamstime’s CFO. I’ve included it in full because he makes several very important points throughout.
I sent the following questions:
And Nancy kindly responded with the following:
I’m not 100% convinced by her argument, since PACA represents agencies rather than photographers (and I’m somewhat of a cynic), but she makes a compelling case AND is infinitely more experienced than me in such matters, I can’t refute them with much (if any) authority.
Both detailed responses (from Dreamstime and Nancy) seem to make the case that agencies ‘may’ well remove copyright metadata, and by doing so are NOT violating the DMCA (though no direct commentary on violation of InfoSoc, but understandably so).
If you know any lawyers who directly represent photographer’s interests, please pass this post onto them for their input.
So… do agencies violate photographer’s DMCA copyright?
I do not know for sure… but it does not appear so..
When I started writing this post, I was pretty certain that they were. Though with my small sample sets AND Nancy’s input, I’m not so sure anymore.
Some certainly DO strip metadata, though which parts are being stripped legally or illegally has become even murkier than before. This might just have to be one of those things I wait out and see if agency’s competitive natures will lead them to become more ‘moral’ on the matter, without having to rely on the law to enforce such ‘positive’ behaviour.
I’m hugely grateful to those agencies (Dreamstime, Pixmac and Deposit Photos so far) who have already committed to preserving and even expanding photographer’s copyright metadata in response to this post and I’m hopeful to hear from those agencies who have yet to voice their policies.
Maybe by working together we can help prevent millions of photographer’s stock images falling down the back of the sofa and being hoovered up by orphaned-works laws
As a stock photographer, what can you do about it?
1) Always include copyright metadata in your image, illustration AND footage files. Images are very easy to apply metadata to in Photoshop, Lightroom, or pretty-much every other image editing suite in existence. Illustration files (.ai and .eps) also have good provision for metadata, though footage file formats are a little different, .mov and .mp4 wrappers have good support.
2) Submit only to agencies who read image metadata from your files, and get into the habit of checking your metadata before and after import to your agencies. Fotolia for example make this easy to check, as each image a contributor uploads has a details page showing all metadata they could read. Other agencies don’t do this as well, and I’d like to see more of this type of contributor’s tool.
3) Be militant about your rights. Your images are YOUR work and you are granting agencies the right to sell them for their cut of the income. IF an agency violates your rights, be sure to call them out on it. It’s likely they’ve done the same to more people, and together you stand a much better chance of finding a reasonable outcome.
What do you think?