Posts on this blog represent my opinion. It may be my considered opinion on the basis of my formal study of law and technology. But it is not legal advice. It must not be treated as, or acted upon as, legal advice and no liability is accepted for doing so.

Sunday, 19 July 2009

NPG v Wikipedia

It’s not often that copyright cases get into the news, although there are notable exceptions, such as Lucasfilm v Ainsworth last year, or the Harry Potter Lexicon case in the USA. But the current spat between the National Portrait Gallery and Wikipedia made the front page of the BBC news website a couple of days ago. At its heart (although there is more to it than this) is a simple copyright question: does a faithful copy of a work – in particular, a work that is now out of copyright – enjoy copyright protection in its own right?

The NPG is complaining that a Wikipedia user, Mr Coetzee, copied high-resolution images of some 3,300 paintings from its website and posted them on Wikipedia. In a letter from its solicitors it asserts that these images are all recent original photographs taken by the NPG and that the NPG therefore holds their copyright, and that Mr Coetzee has infringed this. Mr Coetzee’s position is that the paintings are out of copyright, and since the images he downloaded are quite literally copies rather than original or even transformed or derived works then there cannot be any copyright for him to have infringed.

This is an intriguing issue and for the last few days I have been discussing it with a number of my fellow volunteers for the Open Rights Group. This has been a very interesting experience, especially as several of them are considerably more qualified and experienced in IP law than I am, and I’m grateful for their agreement that I could summarise our discussion in this post. (I should add that ORG is not taking any active role in this dispute, although its legally-qualified volunteers are observing it with keen interest. Furthermore, any legal errors or misconceptions are entirely my own.)

Does the NPG have a case? It is relying on s.1 and s.4 of the Copyright, Designs and Patents Act 1988, which provides that original artistic works, including photographs, enjoy copyright protection. Of course, this leads to the question of what ‘original’ means. Every copyright course I have undertaken has stressed that in English law the originality bar is set very low indeed; in effect, if you create an artistic work such as a photograph, you automatically get copyright in it. In the case of a painting, this is undoubtedly true. It’s even true for a simple drawing; s.4 provides that copyright arises irrespective of artistic quality. It is true for a carefully composed photograph. But if artistic quality is not required, what about skill and effort? The ‘sweat of the brow’ argument, often deployed in considering literary copyright, holds that a certain minimum level of skill and effort must be expended for copyright to arise. What is not clear, particularly in respect of a photograph, is whether that skill and effort applies to all aspects of a work’s creation, or just its originality. This is particularly relevant in the NPG’s claim, as whilst it seems clear that a lot of skill and effort when into photographing its paintings, it was aimed at ensuring that the photographs were as faithful a copy of the paintings as possible – in other words, that no originality was introduced at all.

As early as 1869 it was held in Grave’s Case 4 LRQB 715 that a photograph of an engraving enjoyed copyright under the then-current legislation. However, this view has not invariably been followed; consider, for instance, the comments of Lord Oliver in Interlego v Tyco Industries [1989] AC 217:

“It takes great skill, judgment and labour to produce a good copy by painting or to produce an enlarged photograph from a positive print, but no one would reasonably contend that the copy painting or enlargement was an "original" artistic work in which the copier is entitled to claim copyright. Skill, labour or judgment merely in the process of copying cannot confer originality.”

Now, this was a decision of the Privy Council, and furthermore was strictly obiter. However, it was applied in The Reject Shop Plc v Robert Manners [1995] FSR 870, which was an appeal by way of case stated from a private criminal prosecution under s.107 CDPA 1988. In considering whether copyright subsisted in an enlarged photocopy Leggat LJ held that it did not:

“The process was wholly mechanical and there is nothing to suggest that enlargement was for any purpose of that kind. There was, in short, no evidence before the magistrate of the exercise in the production of what he called the “distorted photocopies” of any relevant skill and labour whatever. It follows that the final images were not original works and so no copyright could subsist in them.”

This is what I term the ‘photocopy argument’ – that if the method of reproduction amounts in nature if not exact technique to photocopying the original artwork, then no new copyright can arise. If the NPG possessed a very large flat-bed scanner and scanned its painting, then it would be hard to argue that the resulting digital image enjoyed independent copyright. If the painting was instead photographed and the image carefully processed to even out any lighting inconsistencies and eliminate distortion, then it would in principle be identical to a scanned image. Is it significant that extra skill and effort went into its creation if the only consequence of that skill and effort was to remove any element that distinguished it from a pure mechanical copy?

The counter-argument, which was originally presented by Kevin Garnett QC in an article in the European Intellectual Property Review (EIPR 22(5) 229-237) considers a photograph of a scene in which no copyright can subsist, such as a landscape. If the photographer exercises no skill or judgement and has limited capacity even to pick a particular viewpoint, then copyright still arises. This image, taken by me from the London Eye, has not been corrected or cropped, and I did not have the option of where to take it from or even, given that I was on a timed ticket, when I took it. Hundreds of similar pictures must be taken daily. But nonetheless, my picture undoubtedly enjoys copyright. So why shouldn’t another picture of a scene without its own copyright, such as an out-of-copyright painting, not also enjoy copyright, especially if the photographer in that instance employed much more skill and effort than I did.

I think it’s fair to sum up our discussion as concluding that the matter is open. There is no recent binding authority either way, although cases such as Interlego and Reject Shop point against copyright arising. Nonetheless, arguments can be made both for and against; my own feeling is that a straightforward application of the law would support the NPG, but that a reasoned argument could well be made against it.

A complicating factor is that Mr Coetzee is not in the UK; he is, as I understand it, in the USA. A US court has ruled on this very issue, in Bridgeman Art Library v Corel Corp. There, it actually sought to apply UK law and interpreted it such that a faithful photographic copy does not enjoy copyright, a position consistent with the US approach to this issue. But if the NPG does go on to sue Mr Coetzee, it would do so in an English court, which might be influenced by Bridgeman but would be under no obligation to follow it.

The NPG makes other claims, including breach of contract, bypassing of technical protection measures and database right infringement. The first would probably run into problems of consideration (or rather lack of it), whilst the second does not seem tenable – the software used by the NPG to display hi-resolution images by increasing the resolution as you zoom in is hardly in the same league as encryption or digital watermarking. The database rights claim is less clear-cut, though.

The usual case referred to in interpreting the sui generis database right is British Horseracing Board v William Hill [2005] RPC 35. The point generally taken from this case is that significant investment in creating the database is necessary for the database right to arise. On that basis, I assumed that irrespective of the copyright point, the amount of effort put into acquiring the images of the paintings for the NPG’s database would give rise to a database right (which would almost certainly have been infringed by abstracting 3,300 entries from it.) However, a friend who manages a large commercial database reminded me when I was discussing this case that the ECJ decision that BHB is based upon also emphasised that this investment must be in the creation of the database itself. As confirmed by the Court of Appeal, the right arises where the database creator has invested effort in creation of a database of existing available data, rather than in creating or approving the data. Indeed, if the data is such that it is only available to the database creator, then the database right does not arise (see Kon and Heide, E.I.P.R. 2006, 28(1), 60-66).

And this could be a fatal hole in the NPG’s database claim. The whole basis of the NPG’s copyright argument is that it does not permit unauthorised copying of its paintings; the only body with legitimate access to the data that has gone into the NPG’s image database is the NPG itself. Nobody else is in a position to lawfully create such a database, no matter how much effort they put into doing so. On this argument, the NPG’s database does not attract database right protection.

It’s not at all clear at this stage what will happen. Mr Coetzee may well not acknowledge the claim in the hope that the NPG, even if it obtains judgment in default, may well not seek to enforce it in the US courts. But if it does come to court in England, then both the copyright and database right claims may well be a lot more open to argument than one might originally think.

Update: At TechnoLlama, Andres Guadamuz has looked in much more detail at the contract claim. I agree with him that formation is probably not in issue, but as I've noted in my comment to his post I think there are very serious problems with consideration (i.e. the NPG has received no payment or benefit from Mr Coetzee.)


Anonymous said...

This was one of the better blogs I've read on this topic.

Thank you.

Perhaps the one interesting thing that you have neglected to note is that the Electronic Frontier Foundation (EFF) has agreed to represent Mr. Coatzee pro bono.

Simon Bradshaw said...

Thanks - yes, I saw the update to the Wikipedia page on the dispute that noted this, but neglected to mention it. It will be very interesting to see what approach the EFF takes.

David Gerard said...

The NPG has approached the Wikimedia Foundation for talks on the matter. So it's quite likely that the question of copyright on fidelitous copies won't get to court after all.

Frank Wales said...

I started writing a comment, but then got a little carried away.

Physchim62 said...

Thanks also for a new pair of eyes on the deceptively tricky subject. your "legally qualified" colleagues (and yourself quite probably) would be interested in Nottage v. Jackson [1883] 11 Q.B.Div. 627: I've commented on it elsewhere, so I shall refrain here!

As for the breach of contract claim, it would be an entire test case on its own given the practices of many websites. Fortunately, this one would have to pass the hurdle of s. 171(2) CDPA before it could be successful: "Subject to those savings, no copyright or right in the nature of copyright shall subsist otherwise than by virtue of this Part or some other enactment in that behalf." To put it in layman's language, you can't get copyright, or anything that smells like copyright, except by Act of Parliament.

Gregory Maxwell said...

I feel compelled to repeated expressed by the others above: this is a great piece of coverage.

You made a comment I found interesting: "If the NPG possessed a very large flat-bed scanner and scanned its painting, then it would be hard to argue that the resulting digital image enjoyed independent copyright."

I'm not sure if you've seen the NPG's digitization operations. I haven't. But I've seen the setup used elsewhere and the actual operation is very much like a "very large flat-bed scanner":

A large format camera with a betterlight scanback is rigidly attached to a display platform where the art is placed, along with the lighting. All the camera movements, focus, etc, are fixed except a rail allows the camera to be adjusted to match the measured depth of the frame. Images of a calibration target are taken, then the art. Basically scanner turned inside out and scaled up to a large size.

After the initial setup everything proceeds almost mechanically— there is very little per-piece configuration (though some lighting changes need to be made for works of different types and sizes, but this is all pre-formulated and by rote) and certainly no artistic input. Not only does this conserve time, but it's actually good stewardship of the information: If some future researcher is trying to analyse or improve the images it important the the images be as consistently created as possible.

If the NPG isn't using a very objective and effectively mechanical capture process like the above then they are wasting taxpayer money and making things difficult for future researches who may not have the luxury of being able to refer to the original work.

Simon Bradshaw said...


Thanks for that - very interesting, and if this is how the NPG has captured its images it adds weight to the 'mechanical copying' argument.

Putting my Practical Legal Advisor hat (or rather wig) on for a moment, if this looked like it might be going to court I would be seeking pre-action disclosure by the NPG of, amongst other things, the exact process it used to create the images in its database.

Simon Bisson said...

No need to get the NPG to disclose - it's the HP art photography system in action.

From talking to HP Research folk in the past, there's actually quite a lot of skill required in using the system. Each image require individual calibration, and an appropriate lighting direction needs to be chosen. You wouldn't light a Turner in the same way as a Van Gogh. So the NPG can argue that skill has been required.

With regards to Zoomify, it's a tool for producing an idexed tile map of the image (much like the way you zoom into Google Maps), and it requires the person setting up a zoom to determine the zoom resolution, the number of steps, and the type of level blend. In the same way then, there's definitely an argument for a) database protection applying and b) circumvention.

An interesting one to watch. I suspect, like the previous Wikimedia cases, it will be settled with the NPG donating medium resolution images in return for a takedown of the high resolution.

David Gerard said...

@Simon - er, what previous Wikimedia cases? This is, as I understand, the first time a museum or gallery has sent a lawyer's letter.

Simon Bisson said...


According to the thread on Slashdot and on the Talk page on Wikipedia there have been previous cases - including one with a major German museum, all of which were settled by the museums offering medium resolution imagery to Wikimedia (which to be honest, is all they actually need).

David Gerard said...

@Simon - ah, got it :-) There have been numerous somewhat heated disputes, but none going nearly as far as this one.

Ideally WMF doesn't want this to go to court, and I think I can safely say Derrick Coetzee doesn't either. I'd hope the NPG realise they may well lose horribly as well, which I expect is why they contacted WMF to start talks.

It's actually causing problems in WMF's relations with museums that aren't quite as (hm) assertive as NPG. Very annoyingly timed.

Physchim62 said...

very deliberately timed, I think.

overton said...

David Gerard It's actually causing problems in WMF's relations with museums that aren't quite as (hm) assertive as NPG. Very annoyingly timed.

Gosh you don't say, who'd have thought it?

As I've said on other blogs copying technology has reached a stage now where exact replicas of works of art can be created.

The resulting digital files that drive the 3D printers, or CNC machines to reproduce the work may be the culmination of months of work and the expenditure of £10000s.

What appears to be being argued here is that if a disgruntled employee where to upload copies to website in the US, or a data stick, or laptop is lost or stolen that contains the files, then the company that created those files has no legal redress. That because the underlaying work is PD then someone can directly use those use files to drive their own 3D printers, or CNC machines.

Physchim62 said...

It's an interesting question, but it's not the one raised here.

If you spent months of work and tens of thousands of pounds creating an exact copy of the Venus de Milo, would that give you copyright over the Venus de Milo? Of course not! Yet that is what the NPG is trying to assert…

Frank Wales said...

@overton Well, they might have redress for things like breach of employment contract by Mr Disgruntled, or they might have insurance for lost equipment.

But, if some Hypothetical Inc were to expend a fortune diligently creating something that the law did not provide protection for, such as unprotectable copies of PD works, then they ought to have no redress for any "loss". That they spent a fortune on this was their decision, and their risk.

After all, unless they were incompetent at backups, they still ought to have copies of what they created, so they still have what they spent the money on.

If they saw it as an investment, rather than as an expense, more fool them; it's not up to the public or the law to compensate anyone for misunderstanding the difference between investments and expenses.

overton said...

So nice to see so many wikipedia administrators and list-server members in this thread trying to dissemble.

Of course you wouldn't have a copyright on the Venus de Milo, that is a essence of stupidity and dissembling that the wiki apologists are trying promote.

However, you ought to have rights over the actual digital files you have created that are capable to drive a CNC robot to cut a replica from a block of marble.

Physchim62 said...

I'm neither a Wikipedia administrator nor a listserv member, but I'm sure the NPG are gutted that people are daring to interrupt its nauseous stream of proxy propaganda.

If you were looking for a way to protect the CNC instruction files under copyright law, I would suggest that you don't ask Farrer & Co., as they don't seem too hot on such matters. Personally, I don't feel inclined to tell you the answer.

Anonymous said...

The HP system involves custom calibration and lighting to produce a version of the image that gives a remarkably accurate 3D image of the paint surface and texture; it takes a lot of time (a decade for the permanent collection at the NG I think), costs a lot of money, needs training of the operators and is used as the basis of the print on demand system in the National Gallery shop.

After all that investment and work, it doesn't feel intuitively that these should be treated as out of copyright snaps when they haven't been put in the public domain - and maintaining that a gallery should have to hand over something it has invested so much in and is being 'assertive' when it doesn't feels rather disingenuous. Were any of the galleries who have resolved similar appropriation of images asked to hand over the medium resolution images before the appropriation? Was the NPG? It could feel like some species of coercion by WMF.

adam said...

Regarding Zoomify and the (possible) circumvention of NPG's databases - there are open-source tools such as unzoomify that make it very easy to download and recompile the tiled Zoomify images.

overton said...

there are open-source tools such as ...

The trick here is that you have to use a tool to get around the normal operations of the technology. As such the action of downloading the protected data is a deliberate act of circumventing the controls.

Basically its equivalent to forced entry.

That someone found it relatively easy to do is not an issue. If it were then the law is meaningless as logically by virtue of the fact that the someone was able to access the data the data was not "effecively secured".

Physchim62 said...

Well, you don't have to use unzoomify, you can do the same thing by hand, it just takes longer.

But that's not the point. Technological measures are only protected if they are used to protect a valid copyright. Otherwise, I could simply use a "technological measure" (eg, a spray-can of paint) to place a copyright claim on Big Ben, then go an retire on the royalties from postcard manufacturers! (unlike the NPG, I would graciously allow passing tourists to take their own photos)

Frank Wales said...

Basically its equivalent to forced entry.

Actually, I think it's like solving a pretty easy jigsaw puzzle; I don't know of any effective locks that are based on that principle.

Claiming that zoomify is a security measure is like claiming that "arrange these words into a well-known phrase or saying" is a form of data encryption.

In any case, this is beside the point. If you erect a fence around public property, how effective it is doesn't matter; the public are still entitled to climb over it.

overton said...

Actually, I think it's like solving a pretty easy jigsaw puzzle;

If you leave your front door ajar and someone pushes it open to gain entry, that is forced entry. It doesn't matters how little force was actually needed, the issue is whether any force whatsoever was used.

Similarly it doesn't matter how easy it was to bypass the measures put in place to deter someone from downloading the data, the issue is simple whether they knowingly bypass those measures.

David Gerard said...

@overton - and if you put a fence around something you don't own, you can't claim it's theft if someone goes in and makes a copy of the thing you don't own.

The NPG puttin a fence around something they don't own doesn't make it theirs. Not even if it's a very expensive fence that was a lot of work and has lots of fiddly bits.

Simon Bradshaw said...

The NPG's argument that Zoomify is a copy protection measure (let alone an 'effective' one as required by s.296 etc CDPA 1988) is heavily undermined by Zoomify's own FAQ:

In particular, it says "For this reason, we provide Zoomify as a viewing solution and not an image security system."

It is not an 'encryption, scrambling or transformation mechanism under s.296ZF(2)(a) and by Zoomify's own admission is not a copy protection mechanism under s.296ZF(2)(b).

Physchim62 said...

If you leave your front door ajar and someone pushes it open to gain entry, that is forced entry.

And if you leave the front door to a town hall, or a church, or even a museum open, it will be assumed that people will push it; because they have a right to push it.

Gillian Davies said...

if anyone knows more about technologies like Zoomify...ones which actually purport to be - even if are not- security measures, I'd love to hear.

on Bridgeman- as you prob know, the UK Museums Copyright Group represented by Jonathan Rayner James said a pretty stark 'this is not UK law' and so Bridegman images (so ergo NPG images) WOULD BE COPYRIGHT-PROTECTED at UK law.

David Gerard said...

@Gillian Davies - 1. Zoomify doesn't purport to be a security technology either. Their FAQ states this expressly. NPG can claim it is, but the manufacturer says it just isn't. If the NPG claimed a piece of cheese was a lock, that wouldn't make it a lock.

2. The UK Museums Copyright Group is headed by a guy from the NPG, so their hopeful assertions, as well as hardly constituting a legal opinion, are somewhat less than independent.

Simon Bradshaw said...

I think that Bridgeman is best described not so much as 'persuasive' as 'suggestive'. An English court might well look at it as one possible interpretation of CDPA 1988, but would bear in mind that it was interpreted through the lens of US copyright law, which has some important differences from the English approach.

pbhj said...

marypcb >>> "maintaining that a gallery should have to hand over something it has invested so much in and is being 'assertive' when it doesn't feels rather disingenuous"

If the gallery hadn't invested public money your comment would be valid. The UK public bought or were donated the images originally, the public paid the wages and bought the equipment to make the facsimiles. The UK public are now paying the lawyers and the NPG people to prevent them from getting to see images of the artwork that we own.

If only the NPG were about getting art to the people and not some ghetto for snobby artists then we'd have a way that any UK subject could get a copy of the NPG works from which they could order a copy to hang on the wall _only_ for the cost of printing of the image.

As anyone in the UK could then just take a disc of the images and send it abroad, quite legally, there's no reason to restrict such access.

Old masters want to be free. Indeed the NPG claims that one of it's aims is "the provision of access to the national collection of portraits for all sections of the population and visitors from abroad and of the facilities necessary for visitors’ enjoyment of the Gallery".

Also the FOI stuff is interesting: if they succeed in claiming database rights then the public by rights should be able to request the portraits under FOI legislation. That would be hilarious!

pbhj said...

Just read this on the NPG site: "We also exert strict controls on all photography in the Gallery, which is allowed only on the understanding that copyright rests with us and that any further reproduction deriving from resulting photographic materials is subject to our written permission."

Absolute poppycock. If the paintings are out of copyright then they're wrong in this assertion. This is an attempt by a public body to apply an ongoing copyright against that legislated by parliament.

pbhj said...

Not trying to spam you, but Green Amps is quite relevant, a case concerning IPR for maps produced by Ordnance Survey.

Anonymous said...

For those saying that this unzoomify-tools is comparable to breaking in a house with the front door open, you're wrong.

Just like you can download an image from Wikipedia, or Google images, you can downlaod an image from sites such as the Royal Collection.

There's only 1 difference.
On Wikipedia's servers, they serve the images in a [img]-tag. Which means most browsers will allow drag-and-drop to your desktop.

Down side is that you have to download the entire image, you can't zoom in or anything.

The royal collection serves them in a Flash-player that allows zooming on several levels and only downloads the piece of the image in view at that moment.

Flash does not allow drag-and-drop to desktop. however I'm 99% sure that Flash is not used as a measure to protect the image. Instead, it's a made for user experience.

Most people on the Royal Collection site dont need to download, they want to be served in a nice Flashy component.

Now, if you use a different browser (not Firefox, or Internet Explorer) but unzoomify, then you download the image that their Flash-player uses directly.

For clarification, images on the royal collection are owned by the Crown, and such are subject to Crown copyright.

In contrary to normal copyright (which expires 70 years after the authors dead), a Crown copyright expires 50 years after publishing.

So, whether you agree with me on or not that this tool is not 'breaking' anything. The images in question are public domain anyway. So it's only to blame for the Royal Collection that they make it so hard to download them.

As a matter of fact, they are actually illigal, in stating "(c) 2010 UK", because the Royal Collection does no longer own the copyright !

Thanks for reading,