Follow Slashdot stories on Twitter

 



Forgot your password?
typodupeerror
×
AI

40,000 AI-Narrated Audiobooks Flood Audible (techspot.com) 93

A new breed of audiobook is taking over digital bookshelves -- ones narrated not by professional voice actors, but by artificial intelligence voices. It's an AI audiobook revolution that has been turbo-charged by Amazon. From a report: Since announcing a beta tool last year allowing self-published authors to generate AI "virtual voice" narrations for their ebooks, over 40,000 AI-narrated titles have flooded onto Audible, Amazon's audiobook platform. The eye-popping stat, revealed in a recent Bloomberg report, has many authors celebrating but is also raising red flags for human narrators.

For indie writers wanting to crack the lucrative audiobook market without paying hefty professional voiceover fees, Amazon's free virtual narration tool is a game-changer. One blogger cited in the report claimed converting an ebook to audio using the AI narration took just 52 minutes, bypassing the expensive studio recording route. Others have mixed reactions. Last month, an author named George Steffanos launched an audiobook version of his existing book, posting that while he prefers human-generated works to those generated by AI, "the modest sales of my work were never going to support paying anyone for all those hours of narration."

40,000 AI-Narrated Audiobooks Flood Audible

Comments Filter:
  • by Talon0ne ( 10115958 ) on Monday May 06, 2024 @01:24PM (#64451788)

    "I can read books out loud"... might not be a great career path.

    • by nealric ( 3647765 ) on Monday May 06, 2024 @01:32PM (#64451802)

      Narration is basically a form of acting, but one that is a bit easier for AI since it only needs an audio component. That's why the SAG went on strike for so long recently- they were justifiably worried that actors would do one or two gigs and then AI would simply replicate their performance going forward. Existing AI tools mean that you could have any famous person you think of narrate your book (setting aside legal rules). Want your book read by Brad Pitt? Taylor Swift? Barack Obama? Donald Trump? AI can have your book read by anybody for whom there's a decent amount of public audio available.

      The one thing AI wouldn't necessarily get right is the dramatic tension. Old text to speech just read in a wooden monotone. AI can anticipate inflection to a degree, but it's not going to always recognize a particularly dramatic note or appreciate a sarcastic statement. So I can imagine a lot of the AI-read texts will sound a bit "off" for now. If the authors themselves are commissioning the AI reading, there could probably be a tool to tag passages as ironic or climactic so the AI gets it right.

      • by DarkOx ( 621550 )

        That remains the challenge here. These tools are going to force a redefinition of 'what is a create work'

        The cat is now out of the bag society and industry are just going to have wrap their heads around it. Nobody is going to pay a livable salary to someone taking weeks with special use studio equipment to do what a machine can in a couple hours. At least not unless it gets way more expensive to build to operate the machine - but I don't see that NVIDA chips and data center cooling might be cheap and the p

      • Agreed. Narrating a book is not like reading scriptures in church with a monotonous and halting tone. It's a skill to do this well, and it's actual work not just a cold reading. Anyone who's read a book to a 5 year old going to bed knows that you have to put effort into it, do some voices, show the emotion.

        Consider music - trivial for a computer to to take a score as input and output the tones. But it's not the same as having real musicians play the music. Different musicians result in very different in

        • Voices are the hard part. Unless you're lucky enough to be doing an ensemble narration, you need to do multiple voices in multiple genders....sometimes species. Skippy The Magnificent doesn't sound anything like Joe Bishop, but they're both the same actor.
          • by Shaitan ( 22585 )

            It isn't just that you have to do them, you have to do the SAME voice for them consistently even though you made up their voice 200 pages ago. You also have to spot the transitions so you adjust properly back to narration and identify the character correctly despite the identity of the speaker typically coming after what you are reading and sometimes not being clearly identified at all.

            I personally have trouble with these things and doing it makes it hard to get the inflection right, I'm expending all of th

            • In spite of all that, they sometimes make mistakes that no one catches. One book made a reference to "Stargate SGI...."
              • by Shaitan ( 22585 )

                You are suggesting a human audiobook narrator injected a reference to SG1 that wasn't in the book?

                • No, but you just demonstrated what happened. They pronounced it "es-gee-eye."
                  • by Shaitan ( 22585 )

                    lol That makes much more sense. And I agree, I've heard mistakes in human narrated content. The wheel of time is one of the highest revenue generators out there in the fantasy world but all the books are not narrated by the same voice actors, sometimes the narrators will be back for a subsequent book after one they didn't narrate. They bring the same mispronunciations of character names and cities with them when they do.

                    Of course that isn't limited to voice actors, the Amazon WoT series mispronounces severa

        • This is an area that has advanced significantly in the last few years. The last AI narrated book I listened to had 9.8/10 pronunciation, subtle changes in tone that made it "feel" more real, and voices for a few of the characters. The only things that stood out to me was the constant tempo of the reading, and a few passages where a human narrator would have read things differently, e.g. the model number of a space rifle would have been "A seven-hundred" vs. A-seven-zero-zero" or places where the source te

      • by Moridineas ( 213502 ) on Monday May 06, 2024 @02:12PM (#64451882) Journal

        I've played around with ePub text to speech using some different AI voice generators.

        You are absolutely correct that it doesn't always get the inflection and the tone 100% correct, but some of the voices do a very good. Pacing, intontation, pitch--these are all in play, and the tech is getting better, rapidly.

        Someone released a mod for World of Warcraft that voices all of the old unvoiced parts from the last 20 years. It's pretty darn good! They could use some more diversity in voices, but I'm very excited to see the continued innovation in text to speech over the next few years.

        • by dfm3 ( 830843 )
          What about character voices?

          Good book readers do this, reading lines from different characters in different voices and then being consistent with things like changes in pitch, tone, or timbre to match each character's personality. And the best readers are able to read ahead, to make subtle changes in the voices that sound natural in dialog when there's a lot of quick "back" "and forth" "and back" "and forth" between multiple characters without the author identifying the speaker of each line. Can an AI get
      • by mysidia ( 191772 )

        but it's not going to always recognize a particularly dramatic note or appreciate a sarcastic statement.

        I'm sure authors will eventually be able to include tags to identify the rising tension; there are already markup languages such as SSML for this.

        If not it will probably still end up being cheaper to pay someone to write markup tags for your speech rather than hire an exclusive voice.

      • by Shaitan ( 22585 )

        The thing AI is going to have a lot of trouble with is distinguishing narration from dialogue and correctly identifying the speaker of the dialog. Also TTS still gets many words and sounds wrong. Especially with made up words.

    • "I can read books out loud"... might not be a great career path.

      wanting to crack the lucrative audiobook market without paying hefty professional voiceover fees

      Sounds like hefty incomes in that industry, lead to this. Perhaps it was a great career path.

      When Hollywood voices cartoons, you should see what they earn. Especially if they negotiate a percentage of profits.

    • by Luckyo ( 1726890 )

      Some people have a genuinely good voice to listen to. Some don't. Some people have a good singing voice. Some don't. Etc.

      But then, some people were really, really good at being hunters. Best hunters in the tribe. Not so much use for that skill any more now that tools advanced and made that skill far less useful. And when it comes to singing, Hatsune Miku has been a thing for a while.

    • It's not merely reading a book out loud, it is about correct pronunciation, voice inflection, tone, and other things we as humans take for granted.

      There's a reason cartoons and video games get to be so popular. The voices behind the characters are what make the characters come alive. They can relate to the character and their voice conveys the emotion. For as good as AI is getting at rendering voices, it still hasn't mastered emotional input.

      As an aside, have AI try to narrate one of Tolkein's books. How

    • Yes anyone can read books out loud. The real question is will people pay money to listen to your voice for hours?

      • Me. I'm the person that will pay to listen to someone's voice for hours. That voice is currently Jared Diamond reading "Collapse: How Societies choose to fail or succeed" and was preceded by his reading of "Guns, Germs, and Steel: The Fates of Human Societies." The Germs book was particularly good.

        So far, the AI-voiced books I've heard have been from the Audible plus collection. I pay them, IIRC, $25/month, and get a credit to purchase any audiobook (purchase/own/perpetual license with DRM) from them pe

    • by Cinder6 ( 894572 ) on Monday May 06, 2024 @02:19PM (#64451904)

      A good narrator can elevate a reading experience from good to great, or great to amazing. To dismiss it as simply "reading books out loud" does those performers a disservice.

      (Obviously, the reverse is also true, and I've had my enjoyment of some books ruined by bad narrators or bad direction.)

      • by Shaitan ( 22585 )

        Then they shouldn't be worried, or need us to pretend people should have some sort of intellectual property ownership over a similar sounding impression of their voice. If humans can do impressions without having to pay the person they are impersonating I don't see why someone doing a voice clone does.

      • by BigZee ( 769371 )
        You're absolutely right. I listen to quite a lot of audio books and can think of a few examples where the reader excels. At the very least, you expect them to read character voices in a way that they are sufficiently distinct from each other. Stephen Fry was particularly good at this in the Harry Potter books.
    • Re: (Score:3, Insightful)

      Maybe not, but AI still sucks at it. Yes, even the new ones. They're better (not human level yet) at homophones and homonyms but can't ascertain the exciting part of the story, have terrible pacing, can't tell who they're quoting, can't understand emotional content, and so on. "I can teach AI to read books better" would be a GREAT career path for the next ten or fifteen years.
    • Maybe voiceover actors can get some government funded retraining. I'm sure H Jon Benjamin could be an excellent programmer.
    • by The Cat ( 19816 )

      The only career path is fat python code monkey, which lasts until age 28.

    • "I can read books out loud"... might not be a great career path.

      I've had to deal with professional voice actors before. You'd be surprised how complicated it is to read something out aloud in a way that is pleasing to others, especially when you're asked to convey a specific audible emotion while doing so.

      What next? You think you're better than paid actors since all they do all day is pretend?

  • ... is good enough for some uses.

    If I were blind or had some other specific need for audio, I'd gladly have "1990s computer voice" over "not available in audio form" any day.

    It's the same reason I tolerate automatic-captions on web-videos when I have to have the sound down: It's better than nothing.

    Now, am I willing to pay a lot more for a high-quality voice-actor? Maybe for some things, but not for most.

    Am I willing to pay a little more for "AI-generated" or even "2024-non-AI-quality" computer-voice vs.

    • by Shaitan ( 22585 )

      I'd pay something for 2024 vs 2020 AI generated voice to be frank. 2020 is Siri, Alexa, that garbage you STILL get with Azure voice service. 2024 is Elevenlabs or even just coqui. It may not pronounce everything correctly (and coqui occasionally sounds like it has something caught in its throat no matter how well tuned) but it sounds like a human speaking.

  • Some of it sounds pretty good.

    A pretty amazing example is voice used on the Cryptic Web Chronicles channel on YouTube.

  • Great news! (Score:4, Funny)

    by GeekWithAKnife ( 2717871 ) on Monday May 06, 2024 @02:08PM (#64451876)

    Previously I wondered if I'm really supporting some poor artist by buying their story telling voice but now that I know it'll support an AI engine I feel compelled to make a purchase!

    I'm so excited for the future! I can't wait for AI to read to my children at night and alleviate the awkward young adulthood conversations I might need to have.

    It also occurred to me that AI can replae a lot of anti-social "data driven" managers. We'll just have AI to tell you that you can do that task in 20% of the time!

    In my next post I'll ask ChatGPT to post for me so I can automate my sarcasm. AI is gonna be great I can feel it!
    • by davidwr ( 791652 )

      Previously I wondered if I'm really supporting some poor artist by buying their story telling voice but now that I know it'll support an AI engine I feel compelled to make a purchase!

      I'm so excited for the future! I can't wait for AI to read to my children at night and alleviate the awkward young adulthood conversations I might need to have.

      It also occurred to me that AI can replae a lot of anti-social "data driven" managers. We'll just have AI to tell you that you can do that task in 20% of the time!

      In my next post I'll ask ChatGPT to post for me so I can automate my sarcasm. AI is gonna be great I can feel it!

      ^^^ above post generated by ChatGPT. Or not. The world may never know.

  • AI Trash (Score:4, Interesting)

    by ozzymodus12 ( 8111534 ) on Monday May 06, 2024 @02:14PM (#64451886)
    I've actually bothered to try a few out. It's crap. All of it is crap. AI doesn't really capture the personality or context. It's all sounds very off. You need a human to guide it, otherwise it never sounds correct.
    • I don't get why authors don't read their own books. If they are worried about how they sound, maybe they could do a read through and then AI could map it with all the expressions to a better sounding voice?

      • Creating a decent sound studio has been a major stumbling block for most until really the end of pandemic when home recording took such giant leaps forward. And then there's the voice acting that is, frankly, quite hard. The number of authors reading their own books has increased significantly in the last two years, but nothing compared to the massive increase that AI brings to the table.

        • by Shaitan ( 22585 )

          Any $100 headset combined audacity is more than sufficient for recording an audiobook.

          • A) It's not just the headset. It's the ability to record very large files, back up to correct mistakes, compile the whole thing into a coherent sound file, tag it for chapters, and be comfortable enough to wear for hours of reading. There's a notable expense in having a good enough computer and processing software, which is generally more than the ancient laptops and hand-me-down desktops needed to write text. B) No, most cheap headsets weren't actually good enough until 2021-ish. C) $100 is a substantial e

            • by Shaitan ( 22585 )

              "A) It's not just the headset. It's the ability to record very large files, back up to correct mistakes, compile the whole thing into a coherent sound file, tag it for chapters, and be comfortable enough to wear for hours of reading. There's a notable expense in having a good enough computer and processing software, which is generally more than the ancient laptops and hand-me-down desktops needed to write text. B) No, most cheap headsets weren't actually good enough until 2021-ish."

              As it happens I did this

      • You wouldn't want me to read any books I've written unless you like the stopped up sinuses and deviated septum sound.

      • by taustin ( 171655 )

        1) It's not "reading a book," it's voice acting, and that's a very different skill from writing a book.

        2) Setting up a decent home studio is not cheap. In fact, it would probably cost more than most independent authors will make off of several books. And without a proper home recording studio, the quality will be so crap it will tank sales on any future book they release.

        3) It takes quite a lot of time they'd rather spend writing their next book.

        4) Not everyone has a voice suitable for performance in the fi

        • 2) Setting up a decent home studio is not cheap. In fact, it would probably cost more than most independent authors will make off of several books.

          For an *audiobook*? It's pretty cheap, actually. You can buy a good microphone with a built-in DAC for about $100 ($200 tops), plug it straight into your laptop, and use any number of free programs for editing (and for whatever minimal post-processing you might want to do). The only other requirement is a quiet place to record.

          • by tlhIngan ( 30335 )

            For an *audiobook*? It's pretty cheap, actually. You can buy a good microphone with a built-in DAC for about $100 ($200 tops), plug it straight into your laptop, and use any number of free programs for editing (and for whatever minimal post-processing you might want to do). The only other requirement is a quiet place to record.

            You haven't browsed the videos made by new YouTubers, have you? They do that and it still sounds terrible. It's not equipment, but skill and effort in obtaining a good mix. Skills tha

        • by Shaitan ( 22585 )

          "2) Setting up a decent home studio is not cheap. In fact, it would probably cost more than most independent authors will make off of several books. And without a proper home recording studio, the quality will be so crap it will tank sales on any future book they release."

          Nonsense. You can make a very decent voice recording sample at home with nothing but audacity and $100 headset. Just look at instruction for making voice samples for digital voice projects or a podcast. The next $10k you spend makes absolu

      • by The Cat ( 19816 )

        For the same reason catchers don't play shortstop.

      • by dvice ( 6309704 )

        Anyone can do a brain surgery. Some of us are just slightly better at it than others.

      • Some authors do.
        There is an abridged Neuromancer out there, read by William Gibson.
        Christopher Buehlman narrated his book The Black Tongue Thief, which was a fun, irreverent fantasy novel .
      • Some authors do read their own books.

        Sean M. Carroll, for example, does an amazing job at it. He choose to read himself because "no layman reader can explain the formula's in my books to a listener. It takes some specialization to be able to do that."
        https://www.audible.com/author... [audible.com]

        Then there's Stephen Hawking, where one could argue whether he read or did not read his own audio books ;p but hey, they are in Stephens voice that makes it stand out and which everyone knows.

        Carl Sagan started to read Pale Blue

    • It's crap. All of it is crap.

      It might be low quality but if, for example, you are blind or partially sighted I suspect you'll might really appreciate having access to books you otherwise might not. It would be at least interesting to hear their take on it.

      • by Shaitan ( 22585 )

        It's all relative. Before kindles some of us were taking palm pilots/clie's and reading ocr scans of books.

        They won't error in exactly the same places but have similar sorts of artifacts at a similar ratio to what you'll have with modern TTS here, maybe more with a model tuned to this purpose. Would I have paid the $20 cover price for books in this format? Nope. So I hope they aren't trying to charge the outrageous prices currently charged for audiobooks to get this content but was it a perfectly enjoyable

    • I'll buy audio books just because they're narrated by a certain VO artist. RC Bray is a big one.

  • by MeNeXT ( 200840 ) on Monday May 06, 2024 @02:15PM (#64451894)

    Is this considered derivative work? This is AI generated so it is not copyright-able? If I buy the book and generate an AI version is that copyright infringement?

    I have so many more questions that can have unlimited opinions. As labour to produce items decreases I see that cost is either increasing or just staying where it is and profits being exported to the jurisdiction with the lowest tax burden.

    • by mysidia ( 191772 )

      Is this considered derivative work?

      If A new work contains enough copying to Implicate the other work's production work, then the new work is a derivative work, Yes. That is how you can establish that something is a derivative work in court (By showing similar parts of the original work contained in the new work).

      This is AI generated so it is not copyright-able?

      Copyright does not extend to parts of a work that are not of human authorship, But it would be incorrect to assert that the entire work is non-C

      • by mysidia ( 191772 )

        If A new work contains enough copying to Implicate the other work's production work, then the new work is a derivative work
        friggin' auto-correct:

        If A new work contains enough copying to Implicate the other work's REPRODUCTION RIGHT, then the new work is a derivative work

      • If A new work contains enough copying to Implicate the other work's production work, then the new work is a derivative work, Yes

        I don't think so. I think it's just a mechanical reproduction of the original work, not really any different than a photocopy -- just a different tangible medium. In order to be a derived work, it would have to be a new work, i.e. some minimal amount of creativity would have to have been added, and I don't think an AI can legally add creativity. Perhaps the configuration choices of the person who set the AI up could be considered "minimal creativity". But if not, it's just a copy in a different medium.

        • by Shaitan ( 22585 )

          "some minimal amount of creativity would have to have been added, and I don't think an AI can legally add creativity"

          There are limitless variations of possible outcome from the process and how they collapse depends on how the user configured the AI software used for the purpose. If two people do this, it wouldn't have the same result. Either way, copy or derivative if the author did this the result would still be protected by their original copyright and not somehow jailbroken, as the asker implied, from pr

          • by mysidia ( 191772 )

            There are limitless variations of possible outcome from the process and how they collapse depends on how the user configured the AI software used for the purpose. If two people do this, it wouldn't have the same result.

            This is true, but because of the way AI systems work -- the person who uses AI to create something essentially has a much higher burden of proof to show that each of the elements of the work they wish to claim protection on Are actually elements that Come directly and deliberately from t

            • by Shaitan ( 22585 )

              'That's analogous to the "monkey pressing the shutter button"'

              A monkey pressing the shutter button gets the same result as the human doing so. The creative element is the human having them press the button. If I roll glue on a paper and toss up some confetti and let air currents drop it randomly it is still my creative work and eligible for copyright by me even though ACTUAL random processes placed the visual elements. The creativity came in the form of deciding to do such a thing and call it creative. Do I

              • by mysidia ( 191772 )

                A monkey pressing the shutter button gets the same result as the human doing so. The creative element is the human having them press the button.

                It seems that you failed to realize the reference. The Monkey pressing the button to Take her own Selfie is a specific case already ruled by the US government as Non-Copyrightable. Having the animal press the button is Not a human creative element eligible for copyright protection.

                The human having them press the button is the very argument for copyright-ability

                • by Shaitan ( 22585 )

                  "The decision to do something is Not a copyrightable nor protectable element for a work."

                  You've just declared that literally no creative element is copyrightable.

                  "And started curating at their own discretion which Pages of your document they would print using different methods, and what pages they would exclude from the final book."

                  This is completely false. This is called arrangement and/or production and the arrangement carries its own copyright, it doesn't automatically convey rights to the underlying mat

        • by mysidia ( 191772 )

          just a different tangible medium.

          If you used an AI to create a new book, then it's the same medium. It is Not a copy in the literal sense, and it's quite a bit different from photocopying - suppose the original is in English, and you ran it through the AI page by page to create a French version.

          This would be infringing, because it's just an adaptation -- even if you used an AI.

          In order to be a derived work, it would have to be a new work

          No.. In order for the new author to claim their own copyright on the

      • by Shaitan ( 22585 )

        This is actually quite an accurate and comprehensive summary of these points. I'd also note, that if the author does this the resulting work is still a derivative protected by their original copyright. This would not "jailbreak" the book from copyright as the poster might have been implying.

    • If I buy the book and generate an AI version is that copyright infringement?

      You can't distribute the performance because it is a derivative of the book. But generating AI audio for your own use seems like it falls firmly into fair use. (IANAL, nor do I hand out free legal advice. but I get dragged into software IP issues a lot and am happy to casually discuss the matter)

      But currently the performance can't be copyrighted. So if you own the book, you probably can pirate any AI generated audiobook if you don'

      • by Shaitan ( 22585 )

        "But currently the performance can't be copyrighted. So if you own the book, you probably can pirate any AI generated audiobook if you don't mind fighting a legal battle that you will theoretically win ... eventually."

        Technically untrue but true enough in practice. I've certainly never heard of successful prosecution/suit against anyone who downloaded an MP3 of a recording they had lawfully purchased on physical media.

        • It's a gray area. Probably would fly in the right court, and maybe in SCOTUS if they would hear it (they won't). But major commercial copyright holders are a huge pain in the ass and it's best not to give them any excuse.

          For what's it's worth I have authorize a copy being made for me. I had a book printed from a PDF I purchased. And printer didn't have a problem with it when I gave them proof that I had the authority to make copies (this particular book's license). And to them it seemed like a normal way th

          • by Shaitan ( 22585 )

            "P.S. when you use an AI speech system you aren't running it locally. You send your data off to a data center."

            Unless you are running it locally. I run an LLM system locally with STT (speech to text) and this TTS (text-to-speech) https://huggingface.co/coqui/X... [huggingface.co], I pair it with a wireless omnidirectional BT microphone/speaker device. It lets me walk around and converse with the LLM naturally and hear the responses spoken like conversation. Does a great job and sounds authentic enough. Using a clone of my o

    • Is this considered derivative work

      It would just be a mechanical reproduction of the original work, so only the original work's copyright would apply.

      In the case of a human narrating an audiobook, the resulting work is a derived work, and both the narrator and the author have rights to it. To copy and redistribute it you need the authorization of both.

      But in this case, it's just a copy. If whoever ran the AI on it added some of their own creative choices, for example, inserting, deleting or modifying text, then it would again be a deri

  • Oh look another form of DDOS, it's definitely an accident and a minor bug in the business model and totally not a feature. It's such a weird accident that these tools keep accidentally finding themselves into the hands of scammers and fraudsters.+end of sarcasm+

    The tools being marketed as "AI" are just tools for folks to create, build and monetize giant swarms of fake accounts. They do not actually work in a productive, legal fashion. The most profitable use-cases of AI, much like crypto, are probably n
    • Most of human society is ill-equipped to deal with something human-esque that runs at thousands to millions of times our speed. That's why we have had such fundamental breakdowns every time we get new tech that automates something only humans could do before. This one is poised to be bigger than the previous ones, if it keeps going as it has been going.

  • by Baron_Yam ( 643147 ) on Monday May 06, 2024 @03:16PM (#64452022)

    The blind just got a lot more accessible as an audience. That's good for them and for authors.

    • The blind just got a lot more accessible as an audience. That's good for them and for authors.

      Also, potentially a lot more books became available to those of us who prefer to listen rather than read. If the AI "performance" isn't too grating.

      For most fiction and some types of non-fiction, I prefer audio books over the printed word, because they're more time-efficient. I read far faster than narrators read, but I mostly can't do anything else while reading. Being able to drive, mow the lawn, work on refitting my boat, etc., while consuming audio books has significantly increased the quantity of "r

  • by Megahard ( 1053072 ) on Monday May 06, 2024 @04:05PM (#64452104)

    Maybe we will get the audio version of A Million Digits of Pi [amazon.com]

  • There is so much potential in AI, especially for derivative works, open source and fan projects. E.g there are abandoned EA games which will never have sequels/remakes/updates from EA. Where actors voiced/acted parts in later games, but the same characters they played don't have speech in earlier games. AI could nicely fill the gap and allow fan projects to dub the parts those actors played. Similarly with artwork, trying to replicate a style in a fan project can be an absolute nightmare, especially with 2d
  • I'll ignore any book without a paid voice actor because I'm not paying for an AI performance. I'll hold off until I have time to sit and read a legit print copy.

  • I remember when one of the e-book readers was going to have a text-to-speech option and all the professionals got their dicks in a vice about it.
  • Tens of thousands of books now accessible in audio form thanks to ai lowering the production cost
  • by Tom ( 822 )

    I'm writing as a hobby. Nothing big yet, but 2 published books to my name. I'm writing one right now that I publish chapter-by-chapter on my Patreon. I also make computer games as a second hobby and tried AI voice acting for the dialogs on the one I have on Steam currently.

    I wish there were any reasonable AI voice stuff. I'd love to make it an audio book, but my voice acting talents are minimal and of course I'd want different voices for different characters. But all the AI voices I've tried so far are very

  • The logical step is to distribute the text, maybe with some metadata about emphasis and so on, and then let the reader's device read it aloud.

  • AI applied to audiobooks is simply the next step in the same progression of AI and skilled labor, as applied to audiobooks, that we see everywhere else in our society.

    I have been listening to TTS books since before 2008, when I used eSpeak to make an audiobook of the Shon Harris 4th Edition Complete CISSP Study Guide. I have since used eSpeak for gigabytes of fiction. However, now I listen to the Google TTS on Android, using AlReader.

    eSpeak sounded like a "drunken robot", and I hid some of those errors by u

  • Everyone's favorite incompetent grifters have been flooding Facebook with ads for the "new method", because this allows them to cut yet another step out of their bogus revenue stream, and probably because the last grift is hopefully finally dried up.

    They're advocating using LLMs to "write" the books too, so you don't have to pay anyone at "all" (except them to "teach" you how to run this garbage scam) and to more effectively just absolutely flood the zone with shit.

Never trust a computer you can't repair yourself.

Working...