Kindle 2 Speech:

There's an interesting little copyright kerfuffle afoot in regard to the release of version 2 of Amazon's Kindle e-book device. The Kindle 2 incorporates a "text-to-speech" function; that is, for books that you download onto your Kindle, a function is provided whereby you press a button and the device translates the text into computer-generated speech.

Cool stuff. The Author's Guild, on behalf of book authors, is not so enthusiastic. It is asserting -- thus far, only in public statements, and not in any court proceedings, though they may come -- that the Kindle 2 functionality infringes the copyright holder's rights to create "derivative works," and, therefore, is not within the license granted by authors to amazon to distribute their works on Kindles. [The National Federation of the Blind, incidentally (and for obvious reasons), disagrees]

It's a knotty copyright issue, actually -- though I'm reasonably certain that Amazon has the better of the argument. Here's how it looks to me. Amazon already has the right to "reproduce" and "distribute" the books it sells in Kindle format -- under the terms of which the copyright holder gets a royalty for each reproduction/distribution. So far so good. The license covers reproduction and distribution only; it does not give Amazon the right to "publicly perform" the copyrighted work, or to "create derivative works" based on the copyrighted work.

A sound recording of the book -- an "audiobook" -- is, clearly, a "derivative work" under copyright law. In the Copyright Act, a "derivative work" is defined to include "sound recordings . . . or any other form in which a work may be recast, transformed, or adapted," and "sound recordings," in turn, are defined as "works that result from the fixation of a series of musical, spoken, or other sounds." Because the audiobook "fixes" sounds (onto a CD, or a computer disk, or some other tangible medium), it's a derivative work. So when Audible.com sells you an audiobook copy of a book, they need a "derivative works" license from the copyright holder.

Pre-Kindle 2, in other words, copyrightholders have two separate sources of licensing revenue: Amazon (for the reproducing and distributing copies their book) and Audible (for making "sound recordings" = "derivative works" based upon their book).

Along comes Kindle 2. There's no "audiobook" involved in the Kindle transaction. The copy that customers receive is just the (marked-up) text, in Kindle format - same as before. The sounds are generated on-the-fly when the user presses the right button -- the sounds aren't "fixed" anywhere, i.e. they're not stored separately from the text itself. Therefore, no sound recording; therefore, no derivative work; therefore, no additional royalty revenue for the copyrightholder.

There may be more to it than this - the Author's Guild may be able to come up with an argument that the generation of the sounds, while not a "sound recording," nonetheless creates a derivative work because it "recasts, adapts, or transforms" the original work into a new medium. There are some messy precedents out there on which they may be able to rely to make this argument. I doubt it, though. They face a very difficult slippery slope -- if Amazon needs a separate license because Kindle 2 is creating a derivative work, then presumably so does everyone who reads a book out loud, even to him/herself. That looks a little harsh to me, and I very much doubt that that's the law.

The Death of Copyright, Item #241

There are many reasons why copyright law as we know it is fundamentally ill-suited for the networked age, and why it will (if we are fortunate, and smart) look very, very different 10 or 20 years from now. I've commented on this many times in the past here on the VC (and will keep doing so, unless and until Eugene tells me to shut up).

Here's a nice recent illustration of one of copyright's fundamental problems.

Over the past few weeks, there has been a proliferation of videos on Youtube made using Microsoft's recently-released "Songsmith" software. Songsmith lets you input music into your computer — by singing and/or playing your guitar into your mic, or feeding in a pre-recorded track — and then the software "analyzes" the music and adds backing tracks matching the "genre" of the music you've fed in. It didn't take long for people to take control of the software's capabilities and to begin feeding in classic songs and re-working them, with sometimes spectacular results:

Sgt. Pepper's Lonely Hearts Club Band

I Heard it Through the Grapevine

Both authored by 'azz10,' who, by the looks of things, seems to be pretty gifted at this sort of thing.

That there is copyright infringement here is almost beyond dispute. Let me put it this way: as a copyright lawyer myself, I would not want to be defending azz10's side in an infringement suit. That, alone, is troubling — copyright is supposed to promote creativity, and here's a veritable explosion of creativity — hundreds of thousands of these videos have been posted to date, some great, some awful, many interesting and expressive — and copyright (and only copyright) is standing squarely in the way.

But my point here is different. One of the things that makes copyright so ill-suited to the networked age is that it doesn't scale. Here's what I mean. If azz10 walked into my office and said: "I'd like to hire you to clear copyright on this work I have created; find out who I have to pay royalties to, and get me the rights," I would not be able to complete that task (and no copyright lawyer worth his/her salt could complete that task) in less than 4 or 5 hours of work, and possibly a good deal more. Take the "I Heard it Through the Grapevine" video. Some of the questions I will have to answer in order to do this work competently:

Who owns the rights to the "audiovisual work" portraying the Marvin Gaye performance?

Who owns the rights to the underlying "sound recording"?

Who owns the rights to the underlying "musical work" (i.e., the song itself)> (under copyright law, this is a separate copyrighted work, and can be owned separately from either of the foregoing)

What kind of royalty do I need to pay, i.e. am I paying a royalty for "reproducing" these works, or for "performing" them? Or "publicly displaying" them? Or "transmitting them by digital audio transmission"? Each of these is plausible, none is certain, and it makes an enormous difference in the amount of the royalty and the identity of the persons to whom the royalty is owed (i.e., if it's a "performance," different copyrights are involved than if it's not).

What about Microsoft? Do I owe them a royalty, inasmuch as I've used some of their "genre" tracks in my video?

Some of these involve chasing down facts that may be hard to uncover (like who owns the various copyrights). Others involve difficult questions of law. I'll need to examine some documents (copyright assignments, possibly; the language of the Songsmith license, certainly; things like that). I've done this sort of work, for paying clients; it's not impossible, but it does take some time. The point: it takes orders of magnitude more time to do the copyright clearances than it does to create the work. Think about it — it could well have taken azz10 15 minutes to synch up these tracks on Songsmith — and easily 10 or 100 times more work would be required to do so in compliance with copyright law.

That is an absurd state of affairs. Multiply the waste involved by 100,000 for each of the Songsmith videos posted on Youtube. And then mulitply that by 1,000,000, for each of the Songsmith videos created and not posted on Youtube.

That's what I mean by a failure to scale. This is, remember, all supposed to be about encouraging creative work. In the old days, with a (much) smaller number of relevant events needing copyright protection, the ratio of creative work to law-compliance work may have been reasonable. But it is reasonable no longer.

How we figure out how to change all this is another story. [I try to tell that one in my book]

Amazon Caves on Kindle:

Amazon has modified its new Kindle 2, the subject of a somewhat over-heated copyright dispute that I commented on earlier, so that its new "text-to-speech" functionality will be enabled/disabled at the option of the publisher -- i.e., that each publisher will get to decide for itself, when licensing books for Kindle sale, whether to allow it to be "spoken" through the S-2-T processor or not. Not good news for the consumer, it seems to me -- although if Amazon makes it clear, when you're purchasing Kindle books, whether the S-2-T functionality works or not on that particular title, it's probably a wash for most of us [though not for the blind or the partially blind -- they lose, in this deal]

[thanks to Tom Hynes for the pointer]

Google Books Controversy Heating Up:

I suspect that we're going to be hearing a lot more about the Google Books settlement over the next several months. There will be several hearings about the fairness of the settlement terms in the Fall; the Supreme Court will be hearing the case of Reed Elsevier v. Muchnick, which raises many of the thorny jurisdictional issues that are embedded in the G.B. settlement; and a number of challenges to the settlement are already being prepared and will likely be filed over the next few months (see below).

It's a very complex set of issues, and I don't have a simple or straightforward position on it myself. To begin with, it is, technically, very complicated; to my eye, the best summaries/discussions of the details come from Fred von Lohmann over at the Electronic Frontier Foundation, and James Grimmelmann at NY Law School; well worth a visit if you're interested in learning more about what's going on.

Personally, I don't have a huge amount of sympathy for those who complain about the "unfairness" to copyright holders in Google's plans [see Robert Kunstadt's rather intemperate remarks in the National Law Journal here].

"Google took from the authors first—and belatedly now seeks to legitimize its misconduct by this settlement. Instead, Google should be punished hard, to deter such schemes. Otherwise, Google will succeed where Napster failed. Respect for IP law—as well as the rule of law in general—will decline. If an enterprising homeless person pitches a tent in Google's corporate parking lot for a 'Thomas Jefferson Used Book & CD Flea Market,' will Google let it stay there (since it only occupies part of the lot)? How about for a profit percentage as sweetener to induce "settlement"?

Google's conduct fits the definition of a public nuisance, and may be enjoined as such. It imposes a small harm on a large number of authors. The harm, copying only a portion of each work, is calculated so as not to make it worthwhile for an author to incur the expense of suing for injunctive relief. Google, despite its cute slogan "Don't be evil," is like a large paper mill releasing noxious gas over a wide area, sufficiently diluted that you smell it but faintly. Like those banks that are now "too big to fail," is Google really "too big to infringe"?

That is pernicious nonsense. The Google Books project has the potential to become one of the great information-gathering activities in human history -- every book (just about), at everyone's fingertips, searchable and instantly accessible from any corner of the globe. And we want to deter that?? Because that will decrease "respect for IP laws"? Talk about putting the cart before the horse!! Because it will inflict some sort of terrible "harm" on copyright holders? I'm not terribly sympathetic. Copyright, as Jefferson stressed so long ago, is a "social right" -- given by society because we feel it serves useful ends (incentivizing authors to produce new creative works). When it ceases to serve those ends, it should be eliminated. The Google Books project is another example of how copyright interests, these days, do little more than obstruct useful innovations. There are 7 million (or more) out of print books that Google would like to place on-line where they can actually be accessed and read. I'm sorry if that infringes someone's copyright, but really -- in what way is society better off, exactly, from recognizing the copyright holder's rights in this circumstance?

But that's not to say there aren't worrisome things about the project. Grimmelmann points to some of them, particularly related to antitrust concerns, in his paper. And separately, I've joined a group of authors, organized by the EFF and Stanford's Center for Internet and Society, that is preparing to challenge the settlement on privacy grounds. As it stands, Google will be able to obtain a staggering amount of information about what you and I (and everyone else in the world) is reading, and I want to be sure that that information is destroyed before it can be misused. More on that side of the issue in a later posting.