Author's Guild Executive Director Interview

[legaleagll]

Engadget has an excellent interview with Paul Aiken, the executive director of the Author's Guild regarding their recent well publicized objections to the text-to-speech capabilities found in the Kindle 2, which can be found here http://www.engadget.com/2009/02/27/t...tor-of-the-au/

I say excellent, not because Paul Aiken did a good job of justifying the Author's Guild position, but rather, just the opposite. Engadget asks tough questions for which Paul was only able to offer clearly bogus answers. At one point, he tries to distinguish between the sale of an e-book that could utilize the text to speech capabilities of a laptop and one that would utilize the text to speech capabilities of the Kindle. The exchange is as follows:

Q: This leads right into the next question -- you've got ebook sales happening right now on computers that can do text-to-speech. So are you going to go after Apple, after Microsoft?

A: Of course not. There's a fundamental difference between a text-to-speech on a general use machine, such as a Mac or a PC, and a dedicated device that is intended primarily for consuming books.

Q: What's the difference?

A: The difference is that there are audio rights involved in the books -- the Kindle converts every book that's sold into something other than just an ebook.

His argument is that because the Kindle can't do as much as a computer it makes is different. So, how many more features does the Kindle need before it is OK to have text to speech? His logic is faulty, his arguments are circular and his reasoning is suspect.

[Ervserver]

Hiring a guy to read a book and selling it compared to building a technology to read a book and selling it. I think they have an argument

[Moejoe]

Originally Posted by Ervserver

Hiring a guy to read a book and selling it compared to building a technology to read a book and selling it. I think they have an argument

What argument is there? Machine-read audio is unlikely to reach the levels of intonation and emotion that a trained actor can achieve. Nobody in their right mind would purchase an audiobook of a computer reading the text aloud if they are able to read the text in the first place, or purchase the audiobook. The text-to-speech function is primarily a boon for those unable to read the on-screen text. It's not a choice, it's an option.

An option, it seems, he wants to milk for all the revenue he can. He should be thoroughly ashamed of himself.

[DesiLinguist]

If I understand his position correctly, I think his argument can be condensed to two interlinked propositions:

(1) If text-to-speech were as good a human reading the words, no one would buy audio books.
(2) Amazon can keep improving the text-to-speech indefinitely and some day make it as good as a human reading the words. Speech synthesis technology has been improving markedly over the last decade due to the research efforts at companies such as AT&T and Nuance.

The reason there is fierce opposition to his position is because (1) is clearly false. However, what happens if (2) were true? What happens if (2) becomes true? So, he is trying to negotiate with Amazon for something that could happen in the future now when text-to-speech is not as popular and the Author's Guild have some leverage.

I do have respect for the fact that he is trying to look out for his members: the authors who create the content that all of us appreciate and love. However, personally, I think this fight is not worth it. Amazon possesses enough business sense not to completely ignore the effects of this technology on the audiobook market (a big chunk of which they own as Audible). I think the text-to-speech technology is just an accessibility feature provided for people whose vision is not perfect. Personally, I don't think Amazon will upgrade the text-to-speech that much. Instead, I see them creating hybrid products where they will create optional eBook+audiobook bundles (for more money than the plain ebook version) and then you can listen or read.

Whatever happens, I would have preferred the Author's Guild to be honest and clearly outline the details of the argument rather than making sensationalist claims that the current text-to-speech feature could, in any way, take money away from the authors.

Just my $0.02.

[DaleDe]

I think the hybrid product approach is advancing rapidly. The new Kindle has support for the latest AAX file format from Audible which is exactly that. It is basically an audio with embedded images, maps, etc. that show up automatically in the window on the K2. This is also supported on any of the iPODs that have a video option as it is basically a video format. There is a bit more on this format in the wiki.

Dale

[rhadin]

Originally Posted by legaleagll

His argument is that because the Kindle can't do as much as a computer it makes is different. So, how many more features does the Kindle need before it is OK to have text to speech? His logic is faulty, his arguments are circular and his reasoning is suspect.

That is not what he is saying. There is a difference. The difference is as simple as this: Most ebooks cannot be read by your PC using text-to-speech because the DRM (or something else) TURNS OFF that capability. You have to purchase a specific product with audio rights.

In contrast, he says, Amazon TURNS ON the audio capability in every ebook it sells for the Kindle, but Amazon has not paid for or negotiated the right to do so. It simply has done so as part of its converting ebooks for the Kindle.

[legaleagll]

Even if 2 becomes true, it still doesn't justify his argument that the Kindle is treated differently than a laptop, desktop, etc... I also don't think that voice synthesis will be anywhere near a competition for a human voice when reading a book at any time in the near or even distant future. In the past 20 years voice synthesis has gone from sounding like a robot to sounding a little less like a robot. There is a difference between what AT&T is doing in trying to develop a human sounding voice when the inputs and responses surround a very limited set of parameters and a voice reading text that could be as varied as science fiction to children's books. The nuances required for human reading, including analyzing the context to deliver a line with sarcasm or humor simply aren't within the realm of technology being developed.

I have listened to a portion of a book on the Kindle2 using text to speech and I could never stand to listen to an entire book. That being said, I could imagine someone reading a book that they didn't want to put down, and going on a short trip somewhere and setting the Kindle to read it to them while driving until they got to their destination and could start reading again. That person is never going to go out and buy the audio book to fill in that gap because every time they start reading they would have to find where the audio left off and vice versa. They would just put the book down and pick it back up when the destination is reached (or just read while driving, which I have seen before).

[legaleagll]

Originally Posted by rhadin

That is not what he is saying. There is a difference. The difference is as simple as this: Most ebooks cannot be read by your PC using text-to-speech because the DRM (or something else) TURNS OFF that capability. You have to purchase a specific product with audio rights.

In contrast, he says, Amazon TURNS ON the audio capability in every ebook it sells for the Kindle, but Amazon has not paid for or negotiated the right to do so. It simply has done so as part of its converting ebooks for the Kindle.

Actually, I don't agree that this is what he is saying. For what it is worth, I just tried a random sampling of about 20 of my e-books (current mainstream titles) purchased from various sources and every one was able to be read by my computer. No DRM shut it off.

[Liviu_5]

Originally Posted by rhadin

That is not what he is saying. There is a difference. The difference is as simple as this: Most ebooks cannot be read by your PC using text-to-speech because the DRM (or something else) TURNS OFF that capability. You have to purchase a specific product with audio rights.

The "most" thing is quite a big error; most ebooks do not have drm; many *commercial* ones do, that's true though even there I am not sure about "most" unless you add "current books from large publishers" as a qualifier, but as with e-music I give almost 100% odds that people have very few (unconverted) drm e-books overall on their devices, pc's...

I think that the crux of his argument is in "convenience" On the Kindle 2 you just touch a button and you get automatic TTS on a dedicated device.

On a Pc/Mac, it's true that you can get a good TTS engine, feed whatever format it wants and listen with a touch of the mouse say, but that's not what people generally want, they like to listen on a dedicated device, maybe an iPod or another mp3 player and here is where you lose convenience.

Again you can save your TTS file to a mp3 but it takes time, work...Usually not worth the trouble considering the quality unless you have no choice
because of your vision

As people pointed out you can make the same arguments about automatic language translation and I expect that if ebook devices sell better, you will get a translation engine at some point too;

[DesiLinguist]

Originally Posted by legaleagll

Even if 2 becomes true, it still doesn't justify his argument that the Kindle is treated differently than a laptop, desktop, etc... I also don't think that voice synthesis will be anywhere near a competition for a human voice when reading a book at any time in the near or even distant future. In the past 20 years voice synthesis has gone from sounding like a robot to sounding a little less like a robot. There is a difference between what AT&T is doing in trying to develop a human sounding voice when the inputs and responses surround a very limited set of parameters and a voice reading text that could be as varied as science fiction to children's books. The nuances required for human reading, including analyzing the context to deliver a line with sarcasm or humor simply aren't within the realm of technology being developed.

If (2) becomes true, then no one would buy audiobooks and the core of his argument becomes a little more valid, his pedantry about the differences between the kindle and the personal computer notwithstanding.

My academic research involves working with speech synthesis systems and while they are very far away from replacing human speech, I wouldn't rule out advances that would make it sound very good indeed. And for some people that would be enough. However, as I said, I don't think Amazon would ever make their text-to-speech feature that good even if such a synthesized voice were available. It will remain, at best, an accessibility feature that will never replace audiobooks. So, I think we agree on that point.

My money is an Amazon to figure this out by making it easy for consumers to consume books in multiple media and, at the same time, turning a profit for themselves.

[Sonist]

Originally Posted by Moejoe

... Machine-read audio is unlikely to reach the levels of intonation and emotion that a trained actor can achieve....

I wouldn't bet on that: it's come a long way in the past decade, and in another decade or two it may be right up there.

I can see the point of Paul Aiken, but the reasoning for differentiating between a computer and the K2 is weak. Computers get smaller - the next gen iPhone will likely have enough power to do respectable text-to-speech.

The reality is that, just like the power-loom rendered weavers obsolete during the Industrial Revolution, text-to-speech will sooner or later displace human readers.

Sound-reproduction technology created a revenue-stream for publishers and authors, where there was none before, text-to-speech may well take it away.

The publishers need to adjust to the new reality, or go the way of the weavers. Their e-book costs are different, as are the methods of distribution.

Trying to "kill" new technology which threatens their current model, by taking congressmen to expensive lunches, may work in the short run. But in the long run, it would be as effective, as the weavers breaking the power-looms.

Text-to-speech is here to stay, and the Kindle 2 is only the first drop before the deluge.

[starrigger]

Originally Posted by rhadin

That is not what he is saying. There is a difference. The difference is as simple as this: Most ebooks cannot be read by your PC using text-to-speech because the DRM (or something else) TURNS OFF that capability. You have to purchase a specific product with audio rights.

In contrast, he says, Amazon TURNS ON the audio capability in every ebook it sells for the Kindle, but Amazon has not paid for or negotiated the right to do so. It simply has done so as part of its converting ebooks for the Kindle.

I don't think he said it as clearly in that interview, but I had a long phone conversation with him the other day, and that is indeed one of the concerns he communicated to me.

Here is one of the key points in the interview, in terms of what he told me in our conversation:

So you're saying that the difference is not actually the device, it's the size of the market?

The difference is the channel. For example, many publishers do not have the audio rights or the multimedia rights when they sell electronic books. If they sell files that go into a Kindle and they know its going to use a audio capability that is exclusively licensed to somewhere else, the publisher has a problem, because the author has licensed it somewhere else.

And that, I think, is a legitimate point. There are thousands of contracts that spell out in legal terms what constitutes audio rights. It might not be what you or I would call "audio books." Amazon may not have the right to ignore those existing contracts. The Guild wants the Kindle books to have a software switch to enable the read-aloud if those rights are granted as part of the ebook rights, or disable it if not.

I blogged on my talk with Aiken, which you can read here.

He didn't entirely convince me of the Guild's position, but he brought me a lot closer. Right now, I'm interested in seeing how this plays out.

[Barcey]

He's absolutely correct that when computerized text to speech becomes good enough it will kill the market for audio books. Who cares? How many people are buying the same book as both an ebook and an audio book? If they are do the authors really think they should be collecting royalties for both sales? I'm paying for the content/story not for the technology method that is used to present it to me.

He should be embracing this new technology. Anything to help sell books. Authors are competing for entertainment dollars and should make it as attractive as possible. Amazon is doing them a favour by making it easier for the customer to enjoy the book. The Author's Guild should be working with the TTS developer's to make the experience better. Maybe they can add markup to the text to indicate inflection or indicate if it's a male or female speaking.

[starrigger]

Originally Posted by Barcey

He's absolutely correct that when computerized text to speech becomes good enough it will kill the market for audio books. Who cares? How many people are buying the same book as both an ebook and an audio book? If they are do the authors really think they should be collecting royalties for both sales? I'm paying for the content/story not for the technology method that is used to present it to me.

He should be embracing this new technology. Anything to help sell books. Authors are competing for entertainment dollars and should make it as attractive as possible. Amazon is doing them a favour by making it easier for the customer to enjoy the book. The Author's Guild should be working with the TTS developer's to make the experience better. Maybe they can add markup to the text to indicate inflection or indicate if it's a male or female speaking.

As an ebook enthusiast, I instinctively share the sentiment about making ebooks more attractive. But...

You can't pretend that a thousand, or a million, legally binding contracts don't exist. And if the Guild did what you suggest, they would be aiding in an increasingly unambiguous violation of those contracts.

I don't think the answer to this question is simple at all.

(Purely as an aside, I'll note that most writers make little to nothing on ebook sales right now, and most don't even get audiobook contracts. So when people start talking about "greedy" authors, as some have at times during this discussion, I can only shake my head. The vast majority of writers can't earn anything even remotely approaching a living wage from their books, so there is an ongoing tension between trying to use new technologies to promote sales of their work and remaining vigilant to avoid being screwed out of their share of the money that does get earned.)

[JSWolf]

Are we going to see publishers and/or authors pulling eBook from Amazon until the TTS issue is solved or new contracts drawn up that allow the TTS in the K2 to be used?