Techdirt. Stories filed under "text-to-speech"

DailyDirt: Computers To Talk For Us

Michael Ho — Mon, 19 Nov 2012 17:00:00 PST

Computers help people communicate all the time, but they can be really helpful for people who have no voice at all (eg. Stephen Hawking). Synthetic speech technologies are getting better -- with better algorithms to generate more human-like speech and cloud-based systems that allow processor-intensive software to run on handheld devices. Here are just a few examples of some computer-created voices.

Most people usually think of synthetic speech software sounding like a HAL9000 or Gene Roddenberry's wife, so there's been a notable dearth of synthetic kid's voices. Until now. Meet Harry and Rosie -- a couple text-to-speech voices that sound like British children. [url]
The Blizzard Challenge is an annual competition for taking a limited speech database and building a synthetic voice from it in order to learn more about how algorithms can be improved for speech synthesis. Someday, these results will actually make some Mission Impossible tricks seem plausible. [url]
A new smartphone app can translate spoken words in real time from Japanese to English (or Mandarin or Korean) and vice versa. This app doesn't do much for grammar, but it can get across some meaning even though it's not perfect. [url]

If you'd like to read more awesome and interesting stuff, check out this unrelated (but not entirely random!) Techdirt post.

Permalink | Comments | Email This Story

Amazon Gives In To Ridiculous Authors Guild Claim: Allows Authors To Block Text-To-Speech

Mike Masnick — Fri, 27 Feb 2009 18:09:30 PST

Well here's an unfortunate surprise. Following the ridiculous claims of the Authors Guild that automated reading aloud of ebooks using text-to-speech software is a violation of copyright, Amazon has agreed to back down and make the TTS feature optional on a per-book basis. The company issued a statement explaining why it believes that there is no copyright violation at all, but is still making the feature optional:

Kindle 2's experimental text-to-speech feature is legal: no copy is made, no derivative work is created, and no performance is being given. Furthermore, we ourselves are a major participant in the professionally narrated audiobooks business through our subsidiaries Audible and Brilliance. We believe text-to-speech will introduce new customers to the convenience of listening to books and thereby grow the professionally narrated audiobooks business.

Nevertheless, we strongly believe many rights-holders will be more comfortable with the text-to-speech feature if they are in the driver's seat.

Therefore, we are modifying our systems so that rightsholders can decide on a title by title basis whether they want text-to-speech enabled or disabled for any particular title.

While this does, effectively, allow the copyright holders to shoot themselves in the foot yet again, it's disappointing that the company wasn't at least willing to stand up for its right to offer such a feature without needing permission.

Meanwhile, if you don't mind the temptation to bang your head against the wall repeatedly, you can read an interview the Authors Guild's Paul Aiken conducted with Engadget about this whole thing. What's amazing is his inability to even understand how having a computer read aloud a book is no different than a person reading aloud the book or the Kindle reading aloud the book. He seems to think each is a different case that deserves different rights (and, of course, different licenses).

Permalink | Comments | Email This Story

Authors Guild Continues To Falsely Claim That Reading Aloud On A Kindle Violates Audio Rights

Mike Masnick — Wed, 25 Feb 2009 14:18:00 PST

Roy Blount Jr. is a funny man. He appears regularly on the radio show Wait, Wait, Don't Tell Me..., and is one of their funnier panelists. But, unfortunately, I don't think he was joking when he wrote a silly Op-Ed for the NY Times defending the Authors Guild's stance that having a Kindle use text-to-speech to read aloud is somehow a violation under copyright law. Blount, who's the President of the Guild, seem to be arguing not from a legal perspective, but from a "but this is how we want it to be" perspective. That's because he has almost no legal argument at all. Using text-to-speech to read text you legally own aloud is not copyright infringement. It's not a "fixed" version of the material, and it's not a public performance. So, rather than come up with a legal justification for such a ridiculous claim, Blount resorts to this: "audio books are a billion-dollar market."

This is the "but the old way of doing business made us so much money, so any innovation must be illegal" argument. You know what else was a big market at one time? Horse & buggies. And Gramophones. And 8-tracks. But technology made them all obsolete at some point or another, and that wasn't illegal. You don't get to cling to an old business model just because it made you lots of money. Sometimes new technologies come along, and they provide a better experience for people, and the market changes. Already it's a stretch to say that a TTS read-aloud of a book really "competes" with an audio book, but even if it did one day, that doesn't necessarily make it illegal just because Blount and the Authors Guild wishes it were so.

Finally, Blount responds to criticism that the Authors Guild's stance goes against readers for the blind or the ability to read aloud to your kid at night. He basically just says "no, that's not true." But he doesn't explain why. That's because there is no good explanation, based on what the Authors Guild has said concerning "audio rights." Basically, the Authors Guild is trying to pretend copyright says what it wants it to say, rather than what it actually says. And, when people have pointed out examples of how the Authors Guild's interpretation of copyright is bogus, their response is "well, we didn't mean it for that." They are, of course, missing the point. Those examples aren't to show the full impact of what the Authors Guild is claiming. They're to show that the Authors Guild is wrong in what it thinks copyright grants them. For a group of "Authors" they seem to have trouble with basic reading comprehension.

Update: John Paczkowski has a great post on this, comparing Blount's statements to those of John Philip Sousa a century ago, complaining about the introduction of "playing and talking machines" and how they would destroy music.

Permalink | Comments | Email This Story

Authors Guild Digs Itself In Deeper Concerning Kindle Text-To-Speech

Mike Masnick — Fri, 13 Feb 2009 12:01:59 PST

We were among many different sites that laughed at the ridiculous assertions by Authors Guild executive director Paul Aiken's attempt to rewrite copyright law concerning the "right" for people with a legally purchased ebook to have a Kindle ebook reader with text-to-speech read the book aloud. Aiken said:

"They don't have the right to read a book out loud. That's an audio right, which is derivative under copyright law."

This is, of course, incorrect. To demonstrate how incorrect it was, we applied it to other scenarios: such as reading to your kids at night. This wasn't to suggest that we actually thought Aiken meant you couldn't read to your kids at night, but to show how wrong the original comment was. Amusingly, the Authors Guild seems to have thought we actually believed this, and has issued an attempted clarification, which only serves to make things worse.

First, the Authors Guild isn't backing down -- but says it's "studying this matter closely." Instead, it recommends that authors who haven't already made ebook agreements avoid doing so. Because, you know, when you're an author the best possible thing is to make it more difficult for people who want to read your book to get it. That doesn't seem very smart. All because someone might listen to their Kindle read it aloud? I would think that the potential of lost readers would be a much bigger concern.

Furthermore, in clarifying that, yes, indeed, reading to your kids is legal, the Guild claims:

The remarks have been interpreted by some as suggesting that the Guild believes that private out-loud reading is protected by copyright. It isn't, unless the reading is being done by a machine. And even out-loud reading by a machine is fine, of course, if it's from an authorized audio copy.

I'm having trouble seeing how that's correct, though, it would be great if the usual assortment of IP lawyers chimed in on this one. Having a machine read to you in the privacy of your own home isn't a performance, so there's no performance rights issues. Then if you go down the whole list of rights covered by copyright, none of them seem to apply. It seems that the Authors Guild is making up a new right (watch out, now the lobbyists will try to get it added): the "you can't do anything innovative unless you pay extra" right.

Permalink | Comments | Email This Story

According To Author's Guild, You Cannot Read Books Out Loud

Mike Masnick — Tue, 10 Feb 2009 10:35:47 PST

As you may know, when player pianos were first introduced, there was a massive fear among musicians that it would be the death of the live music industry. With a player piano, why would anyone ever need to go see a live performance again? In fact, the roots of our rather draconian modern copyright law come, in part, from this dispute. You can see it time and time again, as each new technology comes along, someone freaks out, demands an extra layer of gov't granted rights (and often a compulsory license) and only later discovers that these "threats" were really opportunities all along.

The latest such example comes from an absolutely extraordinary statement from Paul Aitken, the executive director of the Authors Guild, upon hearing that the new Amazon Kindle has an experimental text-to-speech factor. Rather than think about how this feature might expand readership, he immediately insisted that it's illegal to use it:

"They don't have the right to read a book out loud. That's an audio right, which is derivative under copyright law."

By that reasoning pretty much any use of text-to-speech software is illegal, which would make for a fascinating legal case. And, actually, if you take that reasoning further, any reading outloud from a book that is not yours is also a violation of copyright law, according to Aitken. Read to your kids at night? Watch out for the Authors Guild police banging down your door. I would think that in purchasing a book, some of the associated rights that come along with it are the right to either read it aloud or have someone else read it aloud to you, but perhaps we'll soon be hearing about a special new "text-to-speech" right with a special compulsory license that will add a nice little extra charge to every book you buy. In the meantime, who put the Luddites in charge of the Authors Guild?

Permalink | Comments | Email This Story

IBM's I'm-Sorry-Dave-I'm-Afraid-I-Can't-Do-That Patent

Mike Masnick — Tue, 6 Jan 2009 16:29:07 PST

theodp writes "Astronaut Dave Bowman may have found the HAL-9000 more believable had the Heuristically programmed ALgorithmic computer been equipped with the technology described in IBM's new patent for Generating paralinguistic phenomena via markup in text-to-speech synthesis. In the patent, IBM describes how you can dupe others into believing they're dealing with a real, live human being by using markup language to feign sadness, anger, laughter, filled pauses (uh, um), breaths, coughs and hesitations (mmm). Here's an example that shows how to make a more lovable HAL: I'm cough sorry Dave sigh, I'm afraid I can't do that."

Permalink | Comments | Email This Story

IBM Seeks Patent On Typing-To-Speech In A Call Center

Mike Masnick — Mon, 21 Apr 2008 18:48:00 PDT

theodp writes ""Caller: What is my account balance? The call handler responds by typing in the response '250 dollars.'" That's an excerpt from a pending IBM patent for cutting offshore call center costs further by hiring reps whose local accents make them incomprehensible to their U.S. customers without the magic of IBM text-to-speech synthesis, which Big Blue explains converts typed responses into "the native language and accent of the caller so that the outgoing voice sounds familiar to the caller.""

As Theodp noted in sending this in, you would think that Stephen Hawking's computerized speech system might count as a bit of prior art. Of course, while the patent covers more than just that, it's hard to see how the idea of letting someone type responses that are converted into speech deserves monopoly protection.

Permalink | Comments | Email This Story