Could someone explain the licensing like I'm 5 that doesn't end up with "because publishers want more money"

When I buy a book I am buying the story. The authors words on paper (well on digital ink). When I buy an audio book I am buying the authors words performed by a voice actor.

I get why the audiobook is more. I'm paying for the author's story + actor. But I never understood why once I have the audiobook I could not be allowed the text version for free.

How are they doing anything but trying to charge me twice for the same story?

Well no, because publishers (and authors) do want more money.

When I buy a book I am buying the story.

This is where your analysis goes wrong.

But I never understood why once I have the audiobook I could not be allowed the text version for free.

For the same reason you're not entitled to a free (or even discount) paperback if you buy the hardback, or why I can't return the paperback I dropped in the bathtub for new copy. You're not licensing the story, you're buying a specific implementation of that story. Furthermore different companies have bought the rights to different versions of that story in different mediums, and they all want to earn their money back.

I think I need something better than that. You are talking about 2 physical items, I don't think anyone would argue just because you buy 1 product off the shelf doesn't mean you are entitled to the rest of the shelf

An audio book requires author, editor, publisher, actor, distributor

A digital book requires the same minus actor

Where as a physical book requires the same + printer, binder, possibly artists, typesetter, shipper

In this case the audiobook and digital book have same distributor, amazon

Say I want to whispersync The Way of Kings. I already own the audiobook but now I have to buy the Kindle version for full price, which is $9.99 (I double checked in a private tab to make sure amazon wasnt giving any discounts)

> Furthermore different companies have bought the rights to different versions of that story in different mediums, and they all want to earn their money back.

How am I not paying for the same thing twice? If there is some other entity involved I wouldn't mind paying a small difference for their service to bring it to that medium.

Why is it people get paid twice for contributing nothing to bring it to another medium?

What you can't do is declare that your payment means that you actually bought the rights to the text of the book or anything else you unilaterally and post facto include in the purchase for philosophical reasons, when those things weren't included in the offer that you accepted. What was offered, and you accepted, was a license to unlimited personal listening to an audio performance. If you want something else, make an offer or wait for them to make one.

It's not quite as simple as this because governments often want to get into the deal and declare that certain offers aren't what the offerer offers but what the lawmakers think he should have offered, but absent this intercession, you were offered an audio performance for a certain price, and if you accepted, that's what you got.

Being able to take the legally purchased text and running it through a computerized reader might end up being legal... we need the courts to decide. It isn't enough that the person selling it doesn't want to allow it.

It doesn't apply to licenses for creative works. When you buy an ebook, you are buying a license to create copies of the creative work as necessary to make use of the license purchased (i.e., a copy in your phone or on your kindle).

FSD doesn't let you sell either of those transient copies, and it doesn't apply to the license itself because the license isn't a copy within the meaning of the statute underlying the FSD. In the rare occasion where the license and the copy are one and the same for a digital good, the FSD would apply.

Right, because that's literally how the FSD was written. The sale of a DVD of "X" necessarily includes a license to view "X". The license is subordinate to the DVD (beause it is necessary for use of the purchased good), and because the FSD applies to the DVD, it also necessarily applies to the subordinate license.

if you buy 1M old DVD with license to not resell that you and vendor have physically signed, your FSD is less likely to hold up than if you bought one DVD at Walmart with an EULA.

No, the FSD still applies. You've just described Redbox, which actually fought and won this battle (about reselling used DVDs). It would be different if you bought a DVD that included a license for an online streaming version of the movie that couldn't be resold. Redbox fought and lost this battle against Disney. In a nutshell: Redbox can do whatever it wants with the physical copy of the movie, but the FSD doesn't apply to the streaming license that was bundled with the DVD and so they couldn't sell those separately from the DVDs themselves.

There are limits to what they can restrict.

Yes, the limitations can't violate criminal or contract law (or other laws in general), and can't completely prevent the licensee from making use of the licensing rights they paid for. Otherwise, any restrictions are fair game.

If the putative licensee doesn't want to accept the restrictions in the license, their remedy is to not enter into a licensing agreement in the first place; they have no right to simply ignore the license restrictions.

Amazon is not fundamentally offering a textual copy of the work. If they were doing so they would be violating the license. Instead they are offering a technology to transform the (possibly flawed performance of) the work using a tool (that is itself possibly flawed). That is not the same thing as distributing a textual copy of the original work.

A computer is a device on which the work is licensed to be distributed, and it happens to have the capability to transform the work, just as a volume knob or a graphic equalizer would.

If Amazon can't offer this technology because they are licensed to distribute the work, what is to stop someone selling such technology who is not licensed to distribute the work? The effect is the same, precisely because it is a general purpose computing device with sophisticated functionality. And yet I don't see publishers getting far with a claim that all manufacturers of speech-to-text devices (think of deaf people) should pay them because the device could be used to transform their works.

Of course I know this is not how any publisher sees their work. E.g. you need a different license to perform a work than you need to modify or make a video production of a work (mechanical vs sync license, etc.)

The well-established doctrine of patent/trademark/copyright exhaustion [1] prevents this undesirable result.

1: https://en.wikipedia.org/wiki/Exhaustion_doctrine_under_U.S....

This includes viewing it on a computer monitor. It's irrelevant that the computer monitor maker does not have a license to the work, since the viewer does.

In the case of the audiobook speech-to-text argument: the licensee has a license to the audiobook but not the prose version of the work, which is a legally distinct creation under many decades of international and domestic IP law.

So Amazon or the customer needs to have the license to the prose work for speech-to-text to be defensible, but neither do.

This is not (if I understand correctly) a PDF document they can download and read, it's a technology that allows the words that are currently being spoken to be seen in written text.

At no time does a legally significant quantity of the written work get conveyed to the customer. Amazon will just ask the publishers to show where this alleged copy of the textual work resides in their system. They won't be able to do so, as it doesn't exist there.

This does seem to be what they are in fact arguing. And I think they will prevail. I would not expect them to prevail if they offered a pdf of the book or if their software actually contained a copy of the original text that was being conveyed to the customer.

That's not a fair comparison, and I'm sure you know that. But I appreciate someone playing the devil's advocate here; such exercises are always useful.

> If Amazon can't offer this technology because they are licensed to distribute the work, what is to stop someone selling such technology who is not licensed to distribute the work?

What's stopping them from selling such technology? Nothing, as far as I'm aware. It's similar to how BitTorrent clients are not themselves illegal, but users of the client can use them towards illegal ends.

I think Amazon will argue that that speech-to-text isn’t “creative enough” nor does it involve humans, and it is a purely mathematical function and this not covered.

Most likely this will not end in Amazon’s favor.

and yes hilariously for a couple weeks it was illegal to run software.

This simplification of the legal case is just as bad or worse than a movie saying they found America's launch code by hacking the dark web (in the movie's defense, they were focused on the fighting and explosions, IT accuracy was not on the top 100 items of their priority list.) It was never illegal for a licensee to run software they had licensed on their own computers.

The issue was whether an unlicensed repair/maintenance company, Peak, could run MAI's OS while repairing customer computers. The 9th Circuit said no. Congress subsequently changed the underlying statute so that repair companies could run computer software as part of their repair/maintenance duties.

I'm not sure how useful any of this is anyway though. I took another look at Amazon Transcribe a couple weeks ago and reached the conclusion that its transcriptions are still pretty bad compared to a human doing it. You get the general gist so long as it's a clean recording but there are lots of errors.

Standalone software is going to be even worse.

Imperfect captioning is better than nothing in some cases. But who wants to read a book--much less use it as a learning tool--when there are errors every other sentence?

ADDED: It really doesn't seem like much of a threat to publishers in its current form. Probably lots of errors and the text is only delivered a few lines at a time as the book is read aloud. However, I can see publishers looking down the road and imagining much improved ML transcription and larger blocks of text as something that's legally indistinguishable.

Really the only actual answer to that question is because a lawyer convinced a judge to say that they can't. If they can get a better lawyer and/or a different judge then maybe all of sudden they can.

It's kind of like how just because you have a pair of scissors it doesn't mean you are entitled to cut people's hair (without a stylist license). Or how just because you own a scalpel you aren't allowed to practice medicine (without a medical license). The government has determined that these behaviors need to be regulated. Perhaps you agree about regulating the surgery but don't agree about regulating haircuts... that's a matter for the political process.

In a similar fashion, Congress has passed laws on copyright that specifically say that if you own a copy of a work you can do anything you want with that copy, but you cannot make a new copy in a different form. This made sense when "a copy" meant a physical book, but that law isn't necessarily worded in a way that makes sense for electronic documents. I am old enough to remember when people tried suing over web servers, because when the browser received the file it made a local "copy" of the document in memory (without a license). The courts eventually settled on an interpretation that roughly said "ephemeral copies made during the process of delivering the work don't count as 'making a copy'". But they also settled on an interpretation that said "changing the medium -- like going between text and audio -- does count as making a copy".

So text-to-speech (or speech-to-text) converters are illegal; or, more precisely, using them on any copyrighted work (which is just about every work) is illegal without the permission of the copyright owner. Maybe not the best law for today's modern world, but it's the law we have and entrenched power groups would resist any attempt to change it.

I strongly suspect that would hold up in court in the US. For the same reason that I can alter the book and story in my own personal context of ownership. I can take the story, the book, and rewrite it for personal consumption, while keeping all of the original content in my new version. What I can't do typically of course is sell that version, distribute it commercially.

I could use a note taking system that pulls pieces from a physical book as I read it, or snap photos of pages I find interesting. Or I could do that manually with the audio version and take notes of the audio book as it progresses. Both approaches - and numerous others like them - would be protected so long as you're not trying to make money from what you're doing.

You could run your own text-so-speech conversion (or vice versa)for your own private use on fair use grounds because it's a textbook case of fair use. It's very different when you're doing it for someone else's use. There are plenty of situations in which even commercial conversion activity would qualify as fair use.

Generally agree with what you wrote. Money isn't the central question though. Something like fanfic distributed for free can still be infringing.

Or just get it from TPB or whatever.

I'm a little confused, but who are the people who are getting "paid twice for contributing nothing to another medium"?

Another way to look at it: if you own a music video, do you have a license for the song that the music video is a music video of? I would argue you do—you have to have that license, to be able to have a license for the MV, just like you have to own licenses to all the tracks on an album, to be able to claim to have a license for listening to the album.

And, as such, it really wouldn’t make sense to attempt to sell licenses to consumption of MVs, that don’t implicitly include licenses to consumption of the audio tracks within said MVs. (Has anyone ever argued that people don’t have a right to use youtube-dl to extract the audio track from MVs publicly-posted by their rights-owners? I assume, if they did, this argument went about as well as arguing you can’t record music off the radio...)

This is the equivalence being offered here: a text book is to an audiobook as the audio track of a music video is to the MV itself: a component. In the case of a text book, it’s not one that’s literally there in bits; but since it can be extracted with 100% fidelity, the information that component is made up of is clearly there, in the thing you bought. It’s no different than if an audio track was embedded in your video but was reversed or something.

Totally disagree. Consider poetry, or poetic elements in literature, where the form of the text conveys meaning. That is lossily conveyed in an audiobook.

They are infringing in the same way as pen and paper companies are infringing when you hand-write the contents of an audiobook on paper for personal consumption. I believe that most people would consider this being an allowable thing to do.

Pen and paper companies sell you tools/materials to create works. They don't know or care what you create with those works. For all they care, you could just eat the pen and paper. Importantly to this example: they don't transcribe the audio--you do. Your transcription is a derivative work and if a fair use exception does not apply, would be copyright infringement if the copyright owner cared to enforce it in these circumstances.

Amazon's Kindle transcribes the audio of an audiobook using software created by its programmers for the sole purpose of transcribing audiobooks. You don't do anything. Amazon does all the work, at your request. And that makes all the difference, because what Amazon is doing has pecuniary value and thus fair use would not apply to their actions.

I believe that most people would consider this being an allowable thing to do.

Most people also believe it's okay to speed. It's still against the law and it's not a defense to a speeding ticket when you get caught.

Back when I listened to audio books while commuting, I learned to be careful when selecting non-fiction history because a lot of things could be hard to follow in the absence of, say, maps in the original printed text.

You do not. You have a license for the music video, which is the consolidated song+video adaptation of the underlying song.

I would argue you do—you have to have that license, to be able to have a license for the MV, just like you have to own licenses to all the tracks on an album, to be able to claim to have a license for listening to the album.

This would be wrong under US, EU, and even just international IP law generally. The lyrics, music arrangement, audio performance, and music video are all separate creative works. If you want to get technical, the choreography in the music video may itself be a separate copyrightable work.

And, as such, it really wouldn’t make sense to attempt to sell licenses to consumption of MVs, that don’t implicitly include licenses to consumption of the audio tracks within said MVs.

Licensing music videos for money is actually a thing that has happened since the 1980s. These licenses are separate from the right to consume a standalone non-video audio performance.

Has anyone ever argued that people don’t have a right to use youtube-dl to extract the audio track from MVs publicly-posted by their rights-owners? I assume, if they did, this argument went about as well as arguing you can’t record music off the radio...

Yes, and it was found to be fair use. Decades ago, the music labels tried to ban tape recorders on the grounds that they could be used to record radio broadcasts. The courts agreed, but said that such use would be fair use if the recordings were just for private use.

This analogy isn't going to clarify things. First of all, I don't think it's necessary to bring opinion and speculation into this, e.g. "I would argue you do", if there is actual law here.

So with music videos and songs, I would be very surprised if owning a music video automatically and simply granted you a license to the song. Ignoring the fact that videos frequently feature tracks specifically cut and rearranged for video (Michael Jackson is an obvious example), I think the main issue is that songs and videos have multiple copyright holders, and it's too complicated to be a relevant analogy.

IANAL, but if I were to record a high-quality video of a live band performance, the video recording is my property. I can take it home, rewatch it, etc. But I think it's obvious that I don't in any way "own" the copyright to the song I filmed. I can't extract the audio, burn it a CD, and sell it. And I'm pretty sure I would have very few commercial rights to distribute the video itself, if the band members never signed modeling contracts.

> I assume, if they did, this argument went about as well as arguing you can’t record music off the radio...

OK, I don't think this is terribly important to debate, but are you implying that settled law says people have the right to record off the radio? That's definitely not my (amateur) understanding.

> In the case of a text book, it’s not one that’s literally there in bits; but since it can be extracted with 100% fidelity, the information that component is made up of is clearly there, in the thing you bought.

This will sound like a quibble but I think it's fundamentally important to point out: text from an audiobook cannot be "extracted with 100% fidelity". Whether it's 100% fidelity or much worse isn't even the point. What you assert simply isn't so. But I'll be happy to admit I'm wrong if you find an audiobook from which you've extracted all the text content, perfectly, as well as supplementary text like footnotes, indices, headers, illustrations, and captions, and of course, page numbers. It goes without saying that visual data like layout and typography are also components of book with value, but I think you get my point.

I think that is a minor quibble. You're arguing against a point which is not really the important point. That is, that owning a license for the audio book means you also own a license for the printed book including all the extra creative work that went into it, like formatting, pictures, etc.

A lot of other people in this thread are arguing against that point too but I don't think they need to because it's not relevant to the Amazon case where the generated text version isn't taken from the official text version, so it doesn't copy any of the formatting/etc.

If you buy a music video, you can obviously legally listen to it while facing backwards and not seeing the images. So you do have a license to the song in the video like the GP said (if I understand him correctly). Of course you don't necessarily have a license to some other version of that song that somebody else has put additional creative work put into, however minor. That's surely obvious to everyone here and people seem to be talking past each other trying to point that out.

No, but what this is about is saying "am I allowed to listen to just the music without also watching the video? Can I play it on my phone with the screen off? Or do I have to pay extra to do that?"

You can get the same effect without turning off your screen if you tell e.g. VLC to stream only the audio from a network video stream. All VLC will do, in that case, is to request only the audio segments of the stream, and not request any of the video segments.

Consider this another way: is a web browser infringing a website's copyright when it makes a "derivative work" of the page by blocking the ads on it?

Or, hopefully even clearer: is a CD player infringing the CD's copyright when it applies an EQ preset to the audio as it plays it?

There's a very fine point I'm trying to gesture at here—the point where something goes from being "the creation of an infringing derivative work" to "the way that a device which reproduces the work for the purpose of consumption, chooses to do that reproduction."

Consider this another way: is a web browser infringing a website's copyright when it makes a "derivative work" of the page by blocking the ads on it?

Yes, under copyright law, since the article and the page are separate copyrightable works. The question is does a fair use exception apply? (For something like reader mode, the answer is likely yes, but it gets more complicated for an ad-blocker that just removes ads while retaining other elements of the page.)

Or, hopefully even clearer: is a CD player infringing the CD's copyright when it applies an EQ preset to the audio as it plays it?

No, the CD player isn't a person and isn't infringing anything. The user of the CD player is creating a derivative work by applying an EQ. An analogous case (several actually) discusses this point with respect to running computer software. Courts ruled that the copy was necessary for a user to "enjoy" (make use of) their license. And with respect to the EQ, a fair use exception would apply. And even if it didn't, then a compulsory license would apply (in the US at least), and you would owe the song creator a fraction of a fraction of a penny for every performance.

There's a very fine point I'm trying to gesture at here—the point where something goes from being "the creation of an infringing derivative work" to "the way that a device which reproduces the work for the purpose of consumption, chooses to do that reproduction."

No, it's actually a very bright line. Either you're working with the original, or you're working with a derivative work. The question is whether your use of the derivative work you've created is covered by the license, or by fair use.

Even for personal use, where you aren't sharing the extracted audio? I am pretty sure that would be allowed.

It's generally not worth the copyright owner's resources to pursue something like that unless it's happening on a much larger scale.

If I sell a plot of land I inherited from N generations ago, there was zero of my "contribution" factored in into the price (and perhaps zero of my ancestors who first got it contribution. Such plots might have just been up for grabs at the time). The price can still be millions for said plot of land -- it's all to demand, nothing to do with some compensation for "contribution".

In this case, the publishers (and owners of the story copyright) want to sell the 2 different versions separately.

It's as simple as that.

And you can complain about "useless middlemen, etc" but it could just as well have been the author that wanted to sell his stories as different deals for different mediums, and not e.g. allow automatic conversion from one format to another.

Usually, he who sells something sets the price for what you get (and determines what you get).

You might agree and buy it or not, but you don't get to force them otherwise, because you "really wanted X", or "you thought you got Y too", etc.

You then try to reframe copyright, from "Copyright is forcing consumers to overpay for some things" into (paraphrasing) "Anyone complaining about the current state of copyright law[1] is trying to force artists and publishers to live in poverty."

[1] Which is mostly the result of a century's worth of lobbying by IP holders with little to no effective counter-lobbying by consumers.

Read demand as "people wanting the stuff" as opposed to "people wanting the stuff versus available stuff". I'm not talking about the law of supply and demand, but of, let's say, "willingness to pay to get the thing if no other way is available" (even if it's just for legal reasons that no other way is available).

>You then try to reframe copyright, from "Copyright is forcing consumers to overpay for some things" into (paraphrasing) "Anyone complaining about the current state of copyright law[1] is trying to force artists and publishers to live in poverty."

Nope. In fact I wouldn't care if some artists and publishers lived in poverty.

What I argue (and this is not about personal preference) is that the transaction is:

(a) publisher has something another person wants, sets price

(b) said person can buy it or not

We can imagine a setup that's just based on somebody owning the sole ability to sell a thing and its digital copies, and demanding whatever for it.

Scarcity, marginal cost of copy, etc, don't need to come into play.

People who is only authorized to sell gets to set whatever terms, even if those terms are "I want people to pay twice if they want the text version and the audio version" or even crazier.

That's the reality (and, more or less, the law).

We could make it like "Everybody is authorized to make as many digital copies as they want, for free, redistribute them, even put their names on the work" if we wanted.

It's not something set in stone, it's only set in the interests of publishers and in law. And while we can argue that the interests of creators are not the same, most creators tend to just sell their rights to/go along with the publishers, so the point is moot from a legal perspective.

>Which is mostly the result of a century's worth of lobbying by IP holders with little to no effective counter-lobbying by consumers.

Well, in life people don't get what the want/need/deserve. They get what they've managed to negotiate.

If consumers don't counter-lobby/protest/break some heads they'll get the laws ip-holders want.

If IP holders get too greedy, consumers will just pirate stuff. I mean, consider the situation for music vs films. Music is available at low cost through streaming services, and piracy is no longer a major issue. But piracy remains a serious issue for films, which are still priced way too high.

But I suppose that piracy is a form of protest, so hey.

The only reason a company was willing to pay for the rights to the text copy of the work was because they where being guaranteed some sort of monopoly. If that monopoly is weakened then they have less incentives to pay the author.

If that is a good or bad outcome is left as an exercise to the reader.

The costs associated with printing and distributing a physical book are a much smaller slice of the pie than a lot of people assume. (Back when I looked into this, it was about $2.50 per copy for a hardback as I recall. The author cost for a PoD trade paperback from Amazon is about $5.)

Also, in general, the end-user price for books is generally significantly lower than it used to be. Go back a couple decades and you mostly paid list price for books and paperbacks usually weren't available for a year or so after a book came out in hardcover.

The whole publisher intention behind copyright and license rights is that you shouldn't think of digital products any differently than you think of physical products on a shelf.

In other words, the idea is that just because a digital product costs $0 to copy it doesn't mean it doesn't harm them business wise to make one and not pay them. A physical book/cd doesn't cost as much as its sold either -- it usually costs way way less. So "cost of physical copy" shouldn't be thought as the differentiator...

They don't claim that at all. Digital items are typically licensed, not owned outright (as is the case with a physical book).

That said, the way the market is currently configured, audio books are a separate product from digital text. That's fine and legal. But, we don't have to like it. I have a legally blind* family member and it always feels a bit like a scam when she has to repurchase books that somebody else in the family already purchased.

* She can see, barely. For complex texts (particularly textbooks), she prefers to have an audio copy as well as the printed/digital text.

You own an MP3 in the same way you own a book. If you email an MP3 to others in exchange for payment, the FBI aren't going to bust down your friends' doors and securely erase the files from their drives, it's up to the licensor to sue you for damages. Likewise, if you copy books and start selling them at the flea market, nobody's going to round up and destroy the unlicensed copies, it's up to the publisher to destroy you in court.

In some ways you are. This is the same pain the music industry went through when they went digital. People had purchased records, 8-tracks, tapes, and finally CDs often of the same music and often multiple times because those physical items wore out. Now I can buy a song once and it is effectively mine forever. Book publishers see what happened to the music industry and are fighting every way they can. A lot of times those fights do not make sense, but the industry has decided to fight everything.

Two different things. One is an audio performance, which you bought. It contains elements not in the print/ebook version.

The other is a print version of the same work. The purchase includes the formatting, editing, etc., that went into producing the print version which are not found in the audiobook version.

Why is it people get paid twice for contributing nothing to bring it to another medium?

Why is it that programmers get paid six figures to create just another database application? Aren't they all the same thing? No!? Well, it turns out that creating an audiobook and creating an ebook are two very separate things that each take a fair amount of work and that each deserve to be paid for.

Let's say you are an iPhone user and you have bought a bunch of media from Apple. Maybe you really hate the next iPhone announcement and want to move to Android. Would you expect Google to give you a free copy of all the media you originally bought from Apple?

Also you have to get away from the idea that the price of the content is dictated by the costs to produce that content. That production costs is only the minimum price that would allow production to continue. Prices for digital goods are often just dictated by what the seller thinks demand is (since supply is practically infinite assuming there are no contractual restrictions).

...and someone to buy the microphones and recording equipment, and someone to push the buttons, and someone to edit the audio.

I'm asking the reverse, who is getting left out of collecting money when I purchase the audiobook and then want a digital copy for whispersync?

The digital book also requires formatting work. Depending on the type of book and the publisher's workflow, this can be simple (an export from InDesign) or very complex (a custom process for converting a non-technical author's Word documents for a cookbook with an inconsistent layout into an EPUB that renders nicely across devices).

You want a normal copy of the ebook? Pay $10.

Typeset in Arial? Another $10.

Typeset in Comic Sans? Another $10.

Voiced by a free TTS engine? $20 because our market research indicates people pay more for audio books.

The logical way to approach this is that if a product can be automatically converted from one format to another, getting both formats should cost the same as getting the source format, or at most the set should include a marginal fee for the use of the conversion technology development costs + marginal compute costs).

But of course publishers aren't going to do that. That reality, however, is not a defense of the status quo; rather, it's an argument for copyright reform.

Copyright monopoly forces people to pay again and again for "multiple bites at the apple" for purely-digital products that have no secondary market. They're even more arbitrarily priced than normal IP goods like paper books and CDs and DVDs which at least had secondary markets.

But I am. It would take some work, but I can undo the binding of the hardback and turn it into a softback. I can also scan the pages and print them into a new softback. I can even just retype the story and print it into a new softback.

Now, selling, trading, or lending those copies will be an issue, but I don't think anyone here is arguing that someone should be able to use speech to text and then sell the text generated, so the comparison stills stands.

And I can't expect to be given an official softback book, just like I don't have to expect to be given an official text version. But being able to turn an audio book into text seems well within my rights.

Maybe it is. The law isn't the best place to glean any useful moral principles from. I think if you're willing to put in the effort, nobody's going to stop you from making your own text version from the audiobook. But once you're in a position where you're distributing that unauthorized version, or you're already a distributor of the audiobook version and add a feature that automatically creates a text version (without appropriate licensing from the copyright holder), it's something else entirely.

I should clarify I was talking about the existing legal rights owning property entitles me to. Moral rights as well, but that discussion isn't nearly as relevant or useful in this particular case. Sorry for not clarifying.

>and add a feature that automatically creates a text version

This is where I'm not so certain about. If they add a keyboard and notepad that lets me take audio myself, then where is the significant difference in adding software that makes it easier for me to do the transcription? Is there a certain level of ease that anything easier than that becomes illegal?

I don't think there's any uncertainty here. They're creating a text version and distributing it without appropriate licensing. If they give you a notepad, and you happen to write a text version, it's not them creating or distributing it, but it's still unlicensed. If it makes its way on their platform in some way, they're going to need permission, again.

Further, I don't understand why anybody here is giving Amazon the benefit of the doubt. If they want to do this, they can negotiate the required licensing from the publishers. The fact that they've done it without doing this or even discussing it with the publishers shows their clear intent to do what they want, regardless of how much it runs afoul of any laws, simply because they're big enough to fight any lawsuits for ages.

If they create it on their machines and distribute it, then I agree this is an issue. It doesn't matter if it is an AI or if they paid someone to transcribe it, they are distributing a work they likely don't have permission to do so. But if they provide you an application to create it on your machine, then it is you who does the work to make a copy and distributing the application is not an issue (there is an underlying issue of distributing a neural network that was trained by private work, but I'm not certain that is the case here and I think that is a separate issue to discuss that is irrelevant to the use of that neural network).

It is my understanding that the user's device does the translation, but if that isn't the case then my entire argument doesn't apply to this particular case.

>Further, I don't understand why anybody here is giving Amazon the benefit of the doubt.

Personally I don't care about Amazon. But I do care about the precedent set when it comes to distributing applications that make doing something legal easier (at near zero cost). I personally don't even prefer the argument they use and don't even think it works that great as an argument (because it doesn't work if you replace the AI with a person hired to do the same job).

I'd say if this were an open source project created by some volunteers because they were genuinely of the opinion that this provided value to people, and it would in the same way that DeDRM is, there'd be no problem.

> It is my understanding that the user's device does the translation

It's Amazon doing this for financial gain, integrating it into their app with which people consume audiobooks bought through their platform, based on licenses they legally acquired that presumably don't allow this use. I don't see any context in which this is okay. We can't excuse anti-competitive behavior just because if you do the right amount of twisting it seems okay and they found a neat way to skirt contractual and financial obligations. I'd say most significant issue is the scale. They're essentially providing automation of creating unlicensed copies for everybody on their platform and presumably will have a significant impact on the original copyright holders. Which simply wouldn't be the case for an open source project (the amount of people who use DeDRM is miniscule).

Yes. The level of ease where a technology impacts other businesses/versions of material (e.g. the audio book industry here) it is made illegal.

If it's unrealistic/inconvenient as to not have impact, it's ignored.

I don't see why you would be. You aren't permitted to manually copy a book to notepaper, either. The whole point of copyrights are, after all, to manage the rights to copy.

[0] https://answers.justia.com/question/2017/09/03/is-it-legal-t...

[1] https://www.quora.com/Are-audio-transcriptions-legally-allow...

IANAL but, at least as a practical matter, anything along these lines that's done for personal/internal use seems likely fair game. Someone is seriously going to make a case that I can't transcribe a copy of a speech or other recording for my own use?

It's external distribution that gets stickier--especially as you get beyond short excerpts to a full and complete work.

ADDED: And, yes, anyone can sue anyone for anything. But arguments that something could technically be construed as a legal violation even though no one has ever been called on it in the history of the legal system aren't terribly convincing.

How about students photocopying books?

Universities seem pretty convinced that's not permitted.

They note however that The concept is not defined in the legislation, so they seem to be going by feel really.

https://www.soas.ac.uk/infocomp/copyright/library/photocopyi...

Audible on the other hand is not allowed to sell you a text version of an audio book that they licensed even if they are using a clever way of producing the text version of the book.

1) https://en.wikipedia.org/wiki/Sony_Corp._of_America_v._Unive....

Do you have any case law dealing with personal use copies only? In my mind it is like copying an mp3 file. I can make as many copies all over my harddrives as I care. It is only when I share those copies with others it becomes an issue.

>The whole point of copyrights are, after all, to manage the rights to copy.

While it is named copyright, generally they are applied to distributing copies, not the actual creation of copies. For example, anytime I make a backup of my computer I make copies of all data on it, including many I have no copyrights to. It only would be an issue when I started sharing/selling those copies.

I'm afraid not (IANAL), but university text-copying policies seem to assume that personal copies are very much forbidden, for what that's worth.

> While it is named copyright, generally they are applied to distributing copies, not the actual creation of copies

It also covers creation of copies. Unauthorised recording is copyright infringement, despite that it's an act of creation, not of duplication.

> anytime I make a backup of my computer I make copies of all data on it, including many I have no copyrights to. It only would be an issue when I started sharing/selling those copies.

Indeed. That seems to count as fair-use provided you don't distribute to others.

(More generally, there are already lots of products that “covert” a book to a similar but more useful item: bookmarks, stands that hold it open, etc.)

Publishers may not like that product, and may even embargo dealers that sell them, but it’s not clear that they would have any kind of legal right to stop them, even if they were only used for this one purpose.

You're allowed to physically do whatever you want to the copy of a book you've purchased. You can fold it, you can tear it, you can burn it. Hell, you can even eat it if you really want.

I can also scan the pages and print them into a new softback. I can even just retype the story and print it into a new softback.

(The quote is from the parent's comment, not the one I directly replied to but my reply fits here better.) This is not permissible under copyright law, and never has been. That would be creating a new copy, which is exactly what copyright law was created to prevent.

My understanding is that this is entirely legal as long as you don't sell the copy. Borrowing/giving it away would likely be illegal but not worth prosecuting. But a copy solely as a personal backup is allowed. We do that with data all the time.

I mentioned in another post, my understanding is the AI is running on the users machine (given neural networks are cheap to run and expensive to train, it seems plausible the phone is capable of running it). If I'm wrong then my comments don't apply and the issue is more convoluted (I would almost agree it isn't legal in such a case, but I would first have to see about some case law to what I think of as equivalent services, such as online backups/storage/vcr to dvd services).

[0] https://answers.justia.com/question/2017/09/03/is-it-legal-t...

[1] https://www.quora.com/Are-audio-transcriptions-legally-allow...

I bought many blu-rays that came with the DVD version and an UltraViolet code.

It made the choice of buying the blu-ray copy a much simpler choice at a time I didn't have a blu-ray player yet, and it still meant I could watch the DVD version noe, and watch the improved HD copy later.

At least some book publishers offers you the ebook versiin when you purchase a hardcopy, but it's still not universal.

If they want to minimize piracy, their best option is to make their content available in many ways as possible.

I'm not saying the movie industry got it nailed (they still have a long way to go), but they definitely are better than the book industry now.

And people wonder why pirates exist...

Paper is a way to make information into something you can sell as potatoes. But information is different, it's more like that 'single fish he feed dozen people with'.

For decades already we're thinking about what to do with that low price to copy.

Instead of ripping it would be ever so nice to scan the barcode to gain access to digital versions but I understand the issue, I paid for physical copies and for the production and distribution of those, I did not pay for someone to host and provide for me the means to download the copies.

Now some movies do come with a digital version which you can download separately and never use the physical media.

So I have a dresser drawer filled with CDs I had not used for five if not more years. However I do have to say, I never will lose access to that music nor the physical books I own which is more than I can say for stuff I purchase digitally through Amazon or Apple

When you purchased a book it didn't give you ownership to all other implementations of the story.

You still must pay for a recording of a reading, a translated version, the story in a film, and so on.

The right to make implementations is agreed by contract... Licensed... And restricted to the form of an implementation and usually a legal/geographic domain.

If you purchased the right to make implementations in clay tablets, you would not have the inherent right to produce a shadow puppet version too. For that you would need another license from the author (typically via their publisher).

What in the world does this have to do with Amazon providing a feature on their audiobook distribution platform, for which certain licenses were acquired to be able to distribute these audiobooks, that automatically creates unlicensed copies? If you want to create a clay-tablet version from your own copy for your own use, go for it. But the moment Amazon decides they're bored with audiobooks and start selling clay-tablet versions on amazon.com without appropriate licensing, that's entirely different.

Of course I agree that distributing a book as text is basically nothing compared to producing and distributing audio. Making an audiobook (for the most part, I assume) requires the the book text exists in the first place, so the book text is a sunk cost in terms of production.

But it doesn't make economic sense, because a book in e-text format has actual value, and continues to have value and utility even if you end up buying the audio version too. As long as an e-text format has value, how could the market justify reducing its value to 0 when it's part of an audiobook purchase? Wouldn't the price for all audiobooks (that now come bundled with text for free) end up being increased across the board?

People still value the experience of text, enough to be satisfied with paying for text and text alone. Would the market be better if e-text prices stayed the same, and all audiobooks were bundled with text for "free", at a higher price? Sure, if you're the type to buy both audio and text. But not if you frequently buy just audiobooks and are now forced to pay a premium for the bundled text version.

Thinking of real-world factors, I'm assuming that the pay structure and system for audiobooks is different than it is for text. If a writer trades audio rights/royalties for better royalties from print/ebooks, giving away text with audio will hurt their bottom line.

Close to this, an other angle is to see the text version as source, and the audiobook as a representation.

The source having more information, or allowing for more representation has a real value. The same way raw files/original film in photography have a different value from a print.

To go on with the analogy, historically we pay for photographic prints and rarely for the raw file (or the raw file would cost more than individual prints). And there is the same debate, when a photographer shoots a wedding and the couple didn't negotiate the raw files, they can be surprised they don't get them by default, even when paying for the print.

I think in general the source/original material should be included in the sale of the representation, and the whole priced accordingly, but I get that it's not how the publishers see it, or how some customer would bargain to get only the cheapest part, and that behaviour becoming the least common denominator.

Once you write a piece of software, it is also a sunk cost. Therefore I guess you shouldn’t be able to charge for it based on the number of people who are using it?

I gave a short talk on the economics of information and started with a very similar example which was this; Given that 3 ounces of ink and 200 pages of nice 20lb paper, bound with sewing and combines with a cardboard cover into a book costs roughly $3 in reasonable quantities, costs $20 when the ink is used to print J.K. Rowlings story of Harry Potter and the Deathly Hallows?

Clearly the same package with my scribbling about wizards and magic would likely not even be valued at enough to cover the $3 printing cost :-). So it is clear that the information has an intrinsic value, but what is less clear is how it develops that value and how much of the value can be captured.

The book publishers have gone down this road before[1]. 10 years ago the first 'text to speech' features came out in Kindle and Amazon backed down. Ten years later the Kindle is in a stronger place, so Amazon is perhaps inclined to not back down this time.

What I find particularly amusing, is that publishers, when asked back in '09 if it was "Okay" for parents to read their books aloud to their children, they agreed but felt reading aloud to a group of children (as a teacher might do), or to a group of people unrelated to you, was a violation of their commercial rights which unfortunately they couldn't do anything about. Not so with the Kindle, which is standing in here as the reader, that gives them someone to sue.

As technology gives people the ability to put information 'goods' in a real goods market framework (hard to reproduce, Etc.) the owner of those goods exploit those capabilities and litigate anything that would damage their capability to do so.

Personally, I think it is a matter of time before the economics of information become more mainstream in their understanding much like the changes that allow people to reason about "service work" in discussions of GDP. But it is going to take a while. I look forward to reading papers that talk about economies working with goods, services, and information as their drivers of economic activity.

[1] https://www.techdirt.com/articles/20090227/1759173928.shtml

It's like software licencing. You buy a licence for the text, or the audio.

It's why buying an MP3 doesn't entitle you to the Sheet Music.

> I get why the audiobook is more. I'm paying for the author's story + actor. But I never understood why once I have the audiobook I could not be allowed the text version for free.

The audiobook isn't just "more", it's technically a separate work altogether. The audiobook producer will acquire a license to use the book in its own work. Presumably the consideration for the license is either straight up cash or some cut of the profits from the sales of the audiobook, or some combination of the two.

But copyright is not just one right, it's a bundle of rights. And most likely, when the publisher licenses use of the book for creation of the audiobook, they are (probably explicitly) not also licensing rights of distribution of the text of the book. So disseminating the text of the original work would be a violation of the publisher's rights.

Write a book. Your average book that reaches shelves takes 4 to 8 months to write. About 2 months for the first draft and roughly 4 to 10 rewrites depending on the author's quality.

Then it goes to a publisher that hires an editor to comb the book a few times over. Discussing aspects of the book and a decent amount of rewriting from here as well.

Then theres cover art.

Oh and typesetting/formatting for actual publication.

About a good year's worth of work along with many people for one book.

So, go ahead, do that all for free and dont feel salty when some other company takes your story, exactly word for word, makes easy money off of it and gives you nothing.

Just saying. Rent and food isn't paid with like buttons, reddit karma, hn upvotes and hugs/kisses.

For that, look to the past before copyright was a thing. Sure, there was the immediate result of when you put out a book others would copy the text (sometimes poorly). But another huge issues was _translations_. If you put out your book in, say, English, that meant you had a fairly small market. Those with the money for books and the background knowledge and interest in your topic were probably among the smaller well-off classes in several countries. Translations of a hot title were big money - and they'd be done, probably by others, and probably not as you (the author) would like.

Thus, while the first copyrights were "don't copy the text", derivative works like translations were soon added. And because other languages mean other countries, there was an incentive to create international agreements, and this led to modern copyright.

Arguably, all of the above are pretty decent reasons for the copyright protections we have. Consumers benefit because creators can expect to be paid for their work, and because creators are more likely to focus on quality than just getting it out first (particularly on translations).

In that light, the argument that a speech-to-text translation requires permission makes some sense, at least legally. One could even argue that the copyright-holders (publishers, in this case) have a reason to not want a poor speech-to-text translation to ruin the reception of their works.

Of course, this falls down when the publishers are preventing purchasers from making use of such services even when those purchasers are aware of the source of the translation, and any motive other than greed falls away when the publisher wants to be paid a second time. But the publishers aren't coming at this from nowhere, it's an existing, long-present concept.

Arguably laws should change. Alternatively, consumers could demand changes in behavior (I have low expectations of that, but that's the theory).

The opposite is also true, that Amazon removed and no longer allows text-to-speech for Kindle books, except for certain devices that advertised the feature like one of the Kindles and the Fire Phone. I use my wife's old Fire Phone to do this, and it works rather well.

IMO both are completely separate products. It should be up to the publisher on whether or not they want to bundle the text version for free with their audio book (some do).

I bought $medium, and I want to consume it however I want. I didn't "request a license" to use it how they want me to.

For Example, take a Stephen King book. If I bought a Stephen King hardcover book; I can use it as a paper-weight, bug-killer, timeless tome to read under the covers, or Other. If I bought the DVD movie, I can play it for personal enjoyment where-ever I want, however I want. I can also just listen to the audio output, or mute it and just have the video playing. I can consume it in anyway that I want. If I am buying an e-book, the options and methods of consumption change are different but should not be limited. If I ask my partner to read that hardcover book to me at night the Publisher doesn't get to charge me addition money for it.

There have been text-to-speech computer programs going back DECADES. If I had Sound-Blaster 'SAY.exe' [1] read to me the ebook that I bought NOBODY would be crying over "you are stealing our monies!"

The modern text-to-speech is nothing more then SAYv2019.

If I don't want text-to-speech, I'll but the Audio-book and hear it performed by a real-person.

[1]https://ibm-pc.org/manuals/other/creative/SoundBlaster15.PDF - page 25

My spidey sense tells me it's probably in violation but it's one of those things that are so close to impossible to enforce that no one would ever get caught, plus if they somehow did they caught (their evil genius 5 year old planted a camera in the room, recorded it and sent it to the publisher) nothing bad would come of it because any publisher who sued a parent into ruins over reading a book to their kid would probably be outcast from society if the media got a hold of that story.

Now, if you recorded that reading alone and put it on Youtube, I imagine that would for sure be a clear cut violation even if you didn't monetize the video. The same goes for reading it out loud in a public place. The only real difference there is you're no longer in the privacy of your own home and there's concrete proof of the violation in a much different context than reading to your kid. Now you're distributing copyright material.

I had something similar come up with one of my video courses. Someone wanted to buy a single license of my course and then teach the course live to a bunch of high school students in Africa. They asked me if they could do this with only 1 license even though a bunch of students would be consuming the content. I said sure no problem and wished them the best and to ask me any questions if there were any issues. I guess that's one benefit of publishing things on your own. You can treat these things on a case by case basis.

That's a very good one.

My personal opinion is Youtube should not be able to add captions without your consent because suddenly this changes everything.

What if one day Youtube decides to add a transcript feature where they can turn a video into a decently formatted blog post. Now imagine if you had your own website where you used to make a blog post about your video manually and then uploaded your video to Youtube with a link back to your site.

A move like that could remove a lot of traffic to your site because chances are Youtube would outrank your personal site by a lot in search results and would also give people less of a reason to ever click back to your site.

But this is really similar to audio books vs text. They are both independent mediums and having access to 1 shouldn't guarantee free access to the other. It should be up to the discretion of the content creator (or publisher if they are not self published).

Auto-generated captions (something I deal with on a regular basis as I do create video courses) tend to be horrendously bad in quality even on major platforms. Especially with tech content (even if you talk crystal clear).

Auto-generated captions that came out bad leaves both the reader wondering what "jean kids" means when you really said "jenkins" and it also de-values the content creator's brand because now you think "how did such crappy captions make their way into this video?".

Now with that said, I'm all for providing human created / vetted captions that are accurate, but that should be up to the content creator to decide.

The platform should never step in and create new content based off your copyright content in a format that you didn't explicitly release.

You can find lots of these on the App Store; in the days of yore they were also distributed on CD-ROM (as a separate purchase from the corresponding book and cassette book).

One of my favorites is the app for "Moo Baa La La La" which lets you tap on individual animals in addition to the "read-along" mode.

https://apps.apple.com/us/app/moo-baa-la-la-la-sandra-boynto...

So, you could say that this ML product is putting certain software developers out of business.

It used to be you'd purchase a print book, you'd own it, and you could do whatever you wanted with it (e.g., resell it, burn it). That's not the case anymore. You can't resell it, not because of the format, but because you don't own it in the first place.

I'm not a legal expert, and I know it might sound like I'm parsing words, but I feel there's some validity to my 50k feet assessment.

That's essentially what is happening, but that's not the onus for the copyright law.

Standard IANAL caveat.

The law is set to protect and encourage content creators by clearly defining what is copyrighted work, and how that copyrighted work can be used.

> why once I have the audiobook I could not be allowed the text version for free.

Speculation, but it could be that publishers haven't figured out the simple economics of rolling the cost into the price of an audio book, or that publishing deals with authors prevent that type of bundling.

If you display the color image on a black-and-white monitor, you will be seeing something virtually identical to the grayscale version.

By the logic I'm seeing here against Amazon, it would therefore be considered copyright infringement to sell black-and-white monitors?

So that makes it more like a derivative work Amazon is distributing than a client-side format shift.

So then, let me tweak your analogy a bit. Let's say a studio sells copies of a movie to theaters. They sell the color version and the B/W version, and different rooms/times might have one or the other. The theater sells out all the tickets for the B/W version, but can't any sell tickets for the color version.

Thinking quick, the theater management decides to sell tickets to a color showing as a B/W showing, and when patrons come in, they put some layer over the screen so it looks B/W.

In that case, I could kind of see how that would violate the copyright/license agreement of the studio -- you're selling showings of the B/W version you didn't purchase. [2]

Your scenario would be more like a customer coming in with filtering glasses that make it look like the B/W version to that customer only, in which case, yes, it would be harder to make the case that it was copyright infringement.

[1] https://news.ycombinator.com/item?id=20800777

[2] And yes, in practice, the studio wouldn't turn down the money, but let's say they're stubborn about it for some reason.

Don't know. But is it similar to DeCSS technology and the MPAA.

B&W Monitors could be argued to have multiple purposes, a side-effect is showing copyrighted images.

I'm not fully informed as to what Amazon is doing, but if their technology is being applied specifically to create digital text versions of Audio Books, then it could be argued in a similar vein as DeCSS, which was argued to be necessary to provide "backups" of physical media.

It could also be argued that Amazon is used legal maneuvering for a future showdown in which they can argue the effort to create digital text books is minuscule and thus drive the market price down

Again, IANAL.

The product isn't the story by itself.

Copyright = right to make a copy

These are actually separable. If you remove the cover from a book, you are destroying the license token, without necessarily destroying the [now unlicensed] copy. This is something various entities in the book industry may do, to refund unsold copies from the distributor without incurring return shipping, warehousing, or handling charges. Instead of sending it back, the book is rendered unsellable by removing the cover with the ISBN on it, the paper is tossed in the trash, and nobody gets paid for the unsold copy. You can theoretically fish the pages out of the trash and read the book, but that's now technically piracy, because the copy is unlicensed.

If you rip the cover off of a book you bought yourself, no big deal. You still have the license, because you bought it, but then you can't resell your copy, because the license is no longer embodied in a physical token attached to the copy. You can't prove you have the license, and any potential buyer couldn't be certain that you transferred it to them.

For an audio-book on physical media, when you buy it, you buy the license implicitly attached to the copy. But since copyright is a mess, a recording of audio is a different copyright from the material used to generate the audio. An audio-book of Jeremy Irons reading a story has a different copyright as an audio-book of Bobcat Goldthwait reading the same story. You're not licensing the story itself, but a specific recording of the story. It's the same reason why a piece of music can be in the public domain, while a recording of it made by the FooBar Ensemble in 1991 can still be copyrighted. That theory of copyright is also how museums and collections try to claim copyright on digitizations of visual art that has long been public domain. There are a lot of tests about "sufficiently transformative" and "slavish copy" and whatnot. There's also a lot of money in ensuring that cross-media copyrights remain separated.

From the publisher perspective, someone has a license for an audio recording. They probably also sell licenses for print copies, and licenses for e-books. And these are all sold individually, not as a bundle, because they are allowed by law to treat all these as separate works.

From the consumer perspective, buying the audio recording is implicitly licensing both the specific recording and the work on which it was based, because otherwise, how could one possibly legally enjoy the content of that recording? So the consumer considers it perfectly acceptable to transform the format of what they believe they have licensed. I paid; this copy is mine now; I can do with it as I please. Parent post reinforces this. "I'm paying for the author's story + actor." But you're not. You licensed a copy of a recording, period. It's a derivative work, and you're piggybacking on the recording-owner's licensing of the original story, that gives them the right to make an audio-book.

This confusion is not the consumer's fault. It arises from the non-intuitive cross-media parts of copyright law. It arises from the publishers and distributors of copyrighted works exploiting the ambiguity between buying a license and buying a copy. They would dearly like to have their cake, and eat it too. When people buy cakes, and then eat them (as one does with cakes) they get upset, because they sincerely believe that selling a cake and selling the right to eat a cake are completely different things.

If they choose the non-transferable individual licensing model, they can no longer sell the same work to the same person multiple times. If I lose my file, I still have the license, and I can either demand a fresh download from the distributor or acquire one from another source without being considered a pirate. If they choose the license-attached-to-copy model, they can no longer capture the resale market. I can read or listen to my book and then resell it to someone else. What they want, but should not have, is a revocable non-transferable license attached to a single copy, so that if I buy a copy, and lose it, I still have to buy another copy, but if I then locate my original copy again, I can't resell it or get a refund for it. That model is worth more money to them, in their estimation.

In my opinion, the publisher argument is bullshit. But it is founded in law. The law happens to be an ass. So they will likely win, until a federal appellate judge decides that they want their public service career to be destroyed by Disney et al.

When I was a teen (well, younger teen), I learned to understand and comprehend western accents by watching a lot of videos by North American or European creators and using close captions on YouTube. At this point, I don't need them anymore, but I certainly did when I first got started.

Clearly this feature on audible can serve a similar purpose, and can definitely help students like myself in developing countries who are taught how to read and write English, but not exactly speak it - at least not in the same way as the western world.

The music publishers accepted an offer from Apple to allow any song from anywhere to be counted as a “purchase” for $25 a year through “iTunes Match”. There is some number between $0 and $full_wholesale_price that the publishers would accept.

Kindles actually used to have headphones jacks and a feature to automatically transcribe books that you owned into auto-generated speech.

It was a fantastic accessibility feature, but the publishers complained, and the feature and the hardware supporting it were both removed. Although like people are saying, I guess the headphone jack also probably wasn't a very popular feature.

Headphone jack was probably removed to save money... I think the high end kindles support Bluetooth.

Translation: It's optional for the publishers to enable DRM restricting it or not.

If I buy a paperback book, there is nothing stopping me from building a machine I can insert the book into that scans the page, OCRs it to text, and then converts the text to audio.

DRM'd ebooks are a travesty, which is why I immediately remove all DRM as soon as I buy one so that I have full freedom to do whatever I want with it.

Nothing except the extreme hassle of doing so.

Digital is different, because it’s so much easier and cheaper to perform such transformative operations.

eBooks are extremely restrictive by comparison. They are so locked down that you can't do any of the above things I mentioned. You only have one basic freedom: to read it with approved software on an approved device with the publisher in complete control.

* Manufacturers of reading glasses don't require a license to transform works a user consumes into a format more accessible to them

* A built-in magnifier app does not require a license to transform works a user consumes into a format more accessible to them

* A built-in audio transcriber - and this seems to be the sticking point somehow? - should not require a license to transform works a user consumes into a format more accessible to them

A lot of people are getting caught up on this whole "whether it creates a derivative work" thing. And my point is: yes it does, of course it does, any method of consumption effectively does, and these should be explicitly allowed in the name of doing what's right for users and accessibility.

Whether it's built in to the app or built in to the OS, the same transformation is happening. And it's such an essential freedom to be able to consume media that you purchased in an accessible form that I could really give a rat's ass if the publisher thinks that you're only "licensed" to consume that media when it's emitted as sound waves from their approved audio source.

Those publishers probably don't have much leverage against third parties who make tools for personal use. Making copies for personal use is legal, as is manufacturing devices which can make copies. I figure the Android functionality is probably legal given this case:

https://en.wikipedia.org/wiki/Sony_Corp._of_America_v._Unive....

So training set==test set and you don't care because the only purpose of the model was to circumvent certain potential copyright issues.

I feel this same issue exists in many other areas too. A model gets trained on a bunch of copyrighted data, and then can generate similar data... It seems like a kind of gray area. On one hand you might say the learning process is just "inspired" by the copyrighted material... someone else might say it's just derived from it (and should require licensing).

This often seems to get overlooked, particularly in image generation applications. Yea, of course you can generate interesting looking pictures. The question that often remains unanswered is how novel are these images... what is the closest image in the original dataset.

But on the other hand, we certainly wouldn't apply that same standard to a human who "learned" from copyrighted works, so where do we draw the line?

Are you actually having trouble figuring out the line between commercial ML product and human being?

Let's say, for example, that a human who was not a native-born English speaker taught themselves to speak English solely by listening to hundreds of copyrighted English fiction audio books, and comparing them to translations of those same works of fiction in their own native language. (Maybe not the easiest way to learn English, but entirely possible.) Now suppose that human used their newfound mastery of the English language to teach someone else to speak English. Since their knowledge of English was solely derived from copyrighted works, would their knowledge of English be considered a derivative work and therefore subject to copyright? Obviously not.

Now suppose the same thing occurred with an AI. The AI learned to translate some other language into English by reading hundreds of thousands of translated books in both languages. Is the AI model a derivative work, and therefore subject to copyright?

That still leaves the question about should ML training sets do/should count as copyright infringement, but that's completely distinct from the idea that a human can do it without causing a copyright infringement.

As part of the ruling that found the book digitisation project was fair use, the judge stated "Google Books is also transformative in the sense that it has transformed book text into data for purposes of substantive research, including data mining and text mining in new areas" [1]

To me it feels like a continuum. On one end of the scale, the model has been informed by copyrighted data. The model has learned something about the structure of the data, which generalizes to new examples easily. On the other end of the spectrum it's just an efficient encoding of the corpus dataset (in this case mappings of sounds to text). It's not learned anything...

How do you distinguish these cases? Can the second type of "machine learning" be used to circumvent copyright in some cases, and is it?

[1] https://en.wikipedia.org/wiki/Fair_use#Fair_use_in_particula...

One particular argument in Authors Guild, Inc. v. Google, Inc. is about search functionality, which may be little similar. If I understand correctly google books lets search the book. But not sure, if they improved or modeled their search engine on copyrighted books.

On the same line, if I remember correctly google used classic and out of copyright material to teach its translation models. Which will resonate more to the current case.

Or am I missing something and they're openly admitting they used it?

You can memorize a whole book, and it's not copyright infringement. If you, create a compressed zip of book you that infringes if you don't own the copyright.

If a trained model really doesn't infringe, then that's it copyright is dead. Because it's not difficult to overfit a model such that it can replicate the original work. And if that model is no longer under the original copyright it can be shared without restriction.

of course you don't - a compresssed zip, or any transformation of the text is a derivative work. But the neural network you use the text to train is not a derivative work unless you intentionally only used this single book to train it for the express purpose of reproducing the original text. In which case, that is no different than doing a format transformation, and that is protected under copyright.

What i m saying is that any book that is used as a source of knowledge cannot claim derivative work on the resulting culmination of knowledge.

Or 3 Or 4... And reproducing how much of the original work exactly, under what conditions. Does the size of the model factor in to the analysis?

It doesn't seem as clear cut as you suggest to me.

Fair Use Within copyright law there is a doctrine of law known as fair use. Although fair use does not actually give permission to make copies of a work or otherwise use a work without consent, the doctrine provides a defense to copyright infringement. The factors considered when applying the doctrine of fair use include: the purpose and character of the use, the nature of the copyrighted work, the amount of the work used, and the effect the use has upon the market for the copyrighted work.

However, Audible has to license content from copyright holders, which gives copyright holders a certain amount of power over Audible to force them to not offer such a feature, independent of the legality.

And if this conversion is done not locally, but via SaaS, that's just an implementation detail. I believe there's precedent for this (which I cannot find at the moment), but if this was not the case, then you could not touch any document you did not have copyright on using e.g. Google Docs.

Edit: Found the precedent where time-shifting (i.e. video recording) was found to be non-infringing, regardless of if it's done locally or remotely [1].

https://en.wikipedia.org/wiki/Cartoon_Network,_LP_v._CSC_Hol....

If I run a local speech-to-text client on my computer for my own use, that's pretty established fair use.

If I run the same software on a Raspberry Pi and transmit the results over my LAN for my own use, I assume that's also fair use.

If I provision an AWS VM and run the same software there, and then transmit the results back to my device for my own use, is that fair use?

If I get my best friend to provision the AWS VM and run the same software there, and then transmit the results back to my device for my own use, is that fair use?

If Amazon provisions the AWS VM for me, and runs the same software there, and then transmits the results back to my device for my own use, is that fair use?

People are drawing a line here that I genuinely don't see. The whole thing feels incredibly arbitrary to me. The agency of action is that the user is asking a computer to perform a task. The means of action is a piece of software running on a computer, computing results on the fly.

Did Amazon do it out of their own volition, or did you instruct them to do it? Also what you're considering fair use, may or may not be fair use. Fair use is a specific exemption made in a specific context. You cannot reapply that exemption in an SaaS context and call it equivalent. You'll have to first demonstrate the legal basis for your claim that the two contexts are equivalent.

Also I do recognize that all of us here are non-experts giving our two cents, so any disagreement we have is TBH inconsequential to what the law actually says and how its actually interpreted! :)

By purchasing the work through Amazon's store, and clicking the "enable speech-to-text", he instructed Amazon to do it. I should mention this looks like an extremely vague distinction, and it's not even mutually exclusive. Amazon provides a service of its own volition, and he took advantage of that service.

>I should mention this looks like an extremely vague distinction, and it's not even mutually exclusive.

I don't agree with your opinion that its vague.

> Amazon provides a service of its own volition, and he took advantage of that service.

If the action that Amazon took could be proven to be "just as easily" done by the end user, then perhaps Amazons acting on behalf of the end user could be considered equivalent to the end user themselves performing that action.

For e.g. If these is a fair use exemption for making a copy of a CD for a personal backup, then if I told my neighbors kid to do it for me, I think it could be considered equivalent as me doing it myself. Copying a CD is at present, ubiquitous enough, that assisting someone is probably on safe legal ground.

I'm not talking about any EULAs - I'm trying to say that clicking the "speech-to-text" button is no different than opening a file in notepad, changing a few lines, then clicking save - you're instructing a program to do something. Would you say that notepad.exe/Microsoft altered the file "of it's own volition"?

>> Amazon provides a service of its own volition, and he took advantage of that service.

>If the action that Amazon took could be proven to be "just as easily" done by the end user, then perhaps Amazons acting on behalf of the end user could be considered equivalent to the end user themselves performing that action.

>For e.g. If these is a fair use exemption for making a copy of a CD for a personal backup, then if I told my neighbors kid to do it for me, I think it could be considered equivalent as me doing it myself. Copying a CD is at present, ubiquitous enough, that assisting someone is probably on safe legal ground.

All with their logos on the screen.

Oh. Can I dream...

I don't think Fair Use has a chance either: https://en.wikipedia.org/wiki/Fair_use#3._Amount_and_substan...

> In general, the less that is used in relation to the whole, the more likely the use will be considered fair.

If a whole book is transcribed, it's certainly not Fair use anymore.

What if the user takes notes, associating them to time-stamps (audio) or written directly into the pages (print)?

In all those cases the user has altered the work, and made a derivative one, if you will. But the publishers want control over how you consume their works - they want control over what happens on your own device, so they can extract more money from you.

It's hard to express how abhorrent I find this.

If you can successfully argue that in court, chances are you'll win.

> In all those cases the user has altered the work, and made a derivative one, if you will.

It's a difference if you do it for your personal notes, or if you then distribute them at scale and make money from it.

Just for the record I don't like how it's currently handled, it's just my best understanding of what the law and the courts say.

I think that's my main argument - this text-to-speech is the same as doing it for personal notes, and that if this is made illegal, then a huge number of other traditionally allowed activities become illegal.

That it happens on someone else's server is irrelevant: 1) It's the same as hiring someone to take notes on a book you own. 2) There's precedent where someone making a recording on your behalf is legally the same as you doing it yourself. 3) If all else fails, the transcription can be done locally.

I don't know if this is actually true? Because this isn't suing individual audiobook holders, it's suing the company that is choosing to sell copyrighted-text-in-speech-to-text as a service.

Yes, implemention details tend to be important when it comes to copyright and fair use.

One thing that may also come into play - what about accessibility and the Americans with Disabilities Act? In this age of "slippery slopes be damned," I wouldn't necessarily be surprised to see someone file suit alleging that these sort of licenses - that limit scope only to audio - as a whole are infringing upon their rights under the ADA.

(Whether it would stand up or not is another question - anyone can file suit for anything, but there's no guarantee of a win. But it seems to me that Amazon might have a good defense here, just based on that idea.)