I want to say, a lot of effort has been made recently to allow you to Self-Host Omnivore. I have done a lot to move it over so that all the features are self-hostable, including rewriting the entire PDF stack. I received a lot of support from the devs doing this too.
I know the decisions of the Dev team were disappointing, but it's also worth pointing out that the site was kept up until around last month - despite the warning stating that'd be down in November.
Omnivore could have shut down their code base, and prevented self-hosting entirely. I'm glad they didn't.
What's the contribution model moving forward? I see the repository is still active, but is it not still under the Eleven's control? How will it evolve when they stop accepting pull requests?
It won’t be under Elevens control, part of the deal I believe. They’re allowed to remain opensource. Not folded into ElevenLabs.
As for contribution model, it’s still something I’m trying to figure out. For the moment, it was just trying to get a self host build ready and working.
But I have admin rights to the repo, and am not working for ElevenLabs, nor officially Omnivore. I was just a contributor before.
I’ve listened to a few audiobooks on long drives, and have been surprised how hard it is to find good voices on audible. Often a book that might otherwise be good has a prohibitively annoying tone. So honestly the exciting thing here is the customisation.
That said, even in their cherries the emphasis still isn’t quite right in the Tolkien example.
As a first impression, french sounding names should be read as french sounding, even in english text. The voice per se is ok, but as delivery goes (pausing, title vs content), it could be better.
Been using eleven labs for several years now. I was really impressed with their multilingual model a few years ago.
Since then, they’ve released a few cheaper models, but the quality suffers greatly (they still have the old models though so it’s not an issue). They’ve also been releasing a ton of different products around TTS.
I don’t mean this as a criticism — I just am curious why SOTA TTS has not improved from one model by one company several years ago, and why even said company isn’t able to improve on that model.
The biggest challenge with TTS is high quality voice data. The architectures of closed providers still mostly trace their roots to stuff like Tortoise with a few exceptions.
Which is why it's especially ridiculous ElevenLabs allows professionals to upload their voices, charges users of those voices a minimum of $50 per million characters, likely pays under $1 for the compute... and then passes on a whopping $2 back to the professional.
I think the next disruptive TTS competitor is going to form out of just offering to pay better rates than ElevenLabs to their PVC users.
Finetuning established architectures on cleaner synthetic data is already getting open source models increasingly competitive, so getting top PVC samples from the source would likely put you right about where they are today.
Rev share is up to 20% on default rates (depending on notice period). With custom rates they can make their voice more expensive and earn up to $0.2 for every 1000 characters. So you can do the math.
The math is you're paying a pittance considering the insane margins involved and the fact you're using their voices in a flywheel that's actively obsoleting them.
I see users need to keep paying you a subscription fee in order to even get their payouts... but "up to 20%" isn't saying particularly much without the kind of details that should probably be on that page.
-
Considering how much your company owes to an open source model, it's also impressive how little you've returned to the commons.
But no worries, the top comment under this post is an open source model that was finetuned for a couple of thousand dollars by a single dude soliciting the public for random voice samples.
Why would you pay more than necessary to attract the voice talent you need? There aren’t (m)any businesses that pay multiples of market rates just to be nice.
There are plenty of businesses that assign integrity a non-zero value, because most businesses reflect people.
Maybe you're in a bubble devoid of that kind of thinking, so it seems very foreign or quaint.
Even then it's short-sighted thinking at best: the "market rate" is not some magic self-optimizing number.
Underpaying their creators is just creating the opportunity for someone to take the best of them on better terms.
-
Elevenlabs is also able to raise trivially in this environment: you'd think while they're still floating out here without a moat other than high quality data, they'd overpay if anything and make narrators feel like royalty until they're replaced.
This isn't unlike Uber initially paying drivers massive bonuses and undercharging riders until they were able to leverage their massive network to increase prices past what the taxi providers they had decimated were charging. But in this case the marginal cost of providing the service is so low they don't even have to lose money to run a similar play, just take less of it. (in other words, even ruthless greed is not antithetical to paying these folks better)
I don’t disagree that overpaying can be a good strategy, but I question whether it has any bearing on integrity.
I hire someone to paint a fence. We agree on $200 for the job and I pay them $200. We both know that it’s undifferentiated work and I could find a dozen other people who would do it for $200.
Where’s the lack of integrity? Or does it just appear if I know that I actually could afford to pay them $10,000, but chose not to?
We're in a market where people paint fences for $200.
You're building fence painting robot and need someone to teach it how to paint fences by example.
You decide you won't pay the fence painters to teach the robot upfront.
Instead, painters will pay you $20 to even visit the factory.
Then, if a particular painter's fence painting is especially highly rated, you'll pay them a small royalty.
So you send your fence painting robot to compete with fence painters for $20 a fence, passing on a tiny slice of the $20 to the ones who helped teach it.
-
We can consider the creation of fence robot and the competition with the existing market just another piece of the steady march of progress, but there's still obvious room to act with more integrity in this situation.
There was no established rate for what fair wages to teach the robot to paint were, and you can't pay $200 because you don't charge $200... but it's also probably not "-$22/month + 1% royalties".
So much of my time for "reading" is in a context where I can't physically read, so audiobooks are incredibly useful. But being limited to the set of books that gets recorded by the publisher is a real shame.
Haven't tried it yet but AI TTV seems basically perfect now so I'm very optimistic this will work great.
I'm interested for this reason too, even listened to AI TTS books before, but the issue is that they are very monotonous. The tone almost never changes, nor the pacing, it's all delivered with almost no variation which makes listening dull and easy to lose focus
This raises an interesting question around the rights of the author/publisher and who they sold their ebook rights to. If in 3 years we have a perfect AI voice that can read any book as good or better than mid-level narrators, why would you ever buy an audiobook when you could just buy the ebook and pick your voice(s). What a time to be alive
The rise of streaming has made CDs and other offline media obsolete and publishing rights for them largely irrelevant. Audiobooks are likely to face a similar demise. One by one, all the frictions, I mean the colours of life, are fading away, sacrificed for the sake of convenience.
Edit: I think the effect of the invention of vinyl on live performers is more akin to how the commoditisation of HQ TTS will be detrimental to audiobook narrators.
I guess it's the same with other jobs: AI will replace the mid/low quality workers, but the good ones will keep delivering something AI can't.
Two audiobooks that come to my mind:
- The Lord of the Ring series read by Andy Serkis; not only he perfectly switches between each characters voice, but also the feeling of listening "Gollum" for ours is something else altogether
- David Goggins' books; the audiobook version is completely different than the book, since he's not just reading the book, and overall it makes the content easier to digest
It seems that this is using one of the less refined models. In English it sounds like a 4th grader reading in front of a class. Kinda stilted word by word voicing with static pauses between words and no variation in intonation. Tried with two voices and both are the same.
How does it hold up on long stuff? I use Elevenlabs Studio daily and once things start to get into the chapters long, the voice can really start to go off the rails. It'd say they've solved a lot of this over the past 2/3 months, but it does still happen on long stuff.
In autoregressive models error accumulates over time. He likely means the voice starts to make odd sounds/gets lower quality. It would be really interesting if OP could share a clip of this phenomenon!
Various different things can happen, it would take me quite some time to dig up examples but at least with elevenlabs you don't get the clicks and pops you get like on notebook LM for example. 11labs instability comes in the forms of intonation, pitch, accent, garbled words or even once language. I've only seen it happen in the 3k+ words gen's I've done, usually actually around the 75% point of the narration of whatever I've converted, and on average lasting a couple of seconds top.
Yeah - I've experienced this with eleven reader (I don't think you can gen text this long anymore using the reader app, lol) but switching voices fixed it for me
I can go back and try to repro and get a recording....
I think there's an issue somewhere in Kokoro though which means it doesn't actually take advantage of MPS, I did get a modified version up and running, but it was no faster than CPU, even though it passed all the internal tests using mps.
> Yes. The app is completely free to download and use today. Listening to content on the app will not consume credits from your monthly web plan. We do plan to eventually launch some premium version of the app, but even then we will maintain a generous free plan.
For sure, I saw where it was skipping and I wouldn't have been surprised if it were intentional, but good to disprove. Thanks for checking, have a good day
The video shows scenarios of people listening to pdfs of pretty dense material (e.g., computer science, bio mechanics).
Does anyone here actually have positive results doing this? It seems to me listening to anything that's even remotely complex with the intent of learning it just isn't something that's feasible.
For me they are actually best for non-fiction, but it has to be books. Papers are too information dense.
I get easily distracted and lose attention while listening to an audiobook. This is usually problematic with fiction, because suddenly I don't know who this new character is or what's happening. And rewinding to the precise position where I stopped paying attention is of course much more difficult than in written text.
I found that non-fiction books work great for me, because even if you ignore a page or two it makes no difference, the author keeps repeating their point and propping it up with many arguments anyway.
I used to have papers read to me via TTS when I had a long commute. This was before the current crop of neural TTS, mind you, so the quality and naturalness wasn’t as good, but it was good enough to tolerate and to get the gist of a paper. It failed terribly on equations, of course, but that’s often not too important on the first reading.
It depends a lot on the paper. I've been using a TTS app to read papers for years. Papers that are really equation dense, convey they key ideas in figures or get too detailed aren't listenable. But sometimes review articles or papers with one clear message hit that sweet spot and are very listenable. There's one topic where everything I know about it I learned by listening to a review article on a long run. It was actually quite pleasant!
Severe dyslexia here, but ask me about any conversation or audio book or class I've listened to. Gimme anything audio and gimme it at 1.5x plz! I spend so much money gen'ing audio these days but it's soooo nice to be able to learn so quickly now.
Oh that's totally fair, and I'm happy to hear that audio books have been such a huge help for you!
But just to clarify, do you listen to dense material on commutes or while on walks or are you listening and taking notes at a table or something?
That's my main point. As someone that doesn't have dyslexia, sitting at a table and taking notes while reading dense material is already quite difficult to do when I'm trying to actually learn the material. I couldn't imagine being effective in my learning by just listening while on an early morning commute or something like the promotional video shows.
to your point, if it needs to sink in I'm sitting in silence in my thinking room looking at my white wall or eyes closed, listening, pausing, thinking, listening, rewinding. notes etc are pretty useless for me, this probably sounds insane to you, but my memory can't process written words that I'm aware of (reading is just me saying things back to myself in a inaudible whisper) - It's slightly uncomfortable, but can listen to multiple things at the same time and think about them separately (the news and an audio book), not as good at that was I was in my 20s and 30s, but still there. I find it impossible to imagine how people who can read and can take notes etc think, I have no clue, so I presume my thinking style might be hard for someone like that to grasp also?
This is definitely the future, I'm worried about the electric slip and slide world we're heading into though, where everything is completely spoonfed and consumptive. I can't help but think we're heading back into animalism.
Technology is pretty quickly and apparently not only coming for our critical thinking, but our agency
With llms, "knowing things" is already starting to feel like a thing of a past, not to me, but to a lot of others, there's no longer an incentive to "switch on".
Why should a kid learn anything if a robot is instantly better at everything? Maths got replaced by calculators, deep critical thinking will get replaced by llms a lot of the time, which are word calculators, which is the closest thing we have to a logic calculator.
This is more passive autopilot software, which further promotes learning as something you 'consume' rather than something you seek.
The public consciousness has absolutely taken a semptember 11 tier nosedive since social media, we're approaching what I term cultual schizophrenia, which I posted about on my blog which I deleted, but I've readded it if you're interested [https://substack.com/home/post/p-156983317]. There's no contextualisers in the media to give the right emphasis to the right things.
This is just my perspective, from what I've seen from other younger people of my age. We are heading into extremely interesting times, everything profoundly destabalising thing we've speculated about is happening at the exact same time. We desperately need visionaries in politics.
Some good observations there, but I'm still unclear on why you used the term "animalism" - none of that seems to me at all similar to how other species engage with the world.
I’m sorry you’re stressed, but please at least consider that you may be falling into the generational “kids these days” trap. I’m old, so I have lived through the world being on the brink of disasters caused by AI, social media, gay marriage, violent video games, the internet in general, cell phones, pagers, nuclear weapons, television. Probably a bunch more world-ending crises I forgot.
The world is changing, but then again it always has been. IMO some things will get better, some will get worse, but ghe overall arc of human health and prosperity will continue upwards. There is less poverty, less starvation, more opportunity today than ever… even though some aspects of the world are bad and getting worse. That’s the way it’s always been.
Last time I tried Elevenlabs for German text, it got a lot of numbers and dates wrong.
E. g. saying "1963" when the actual year in the text was 1967. Yeah, the voices sound very realistic. But I'm not sure how useful that is if you can't trust the spoken words.
Does anyone know if it got better in the last weeks?
The ad shows someone listening to an article or a story while driving a large vehicle - this is unsafe (depending on the individual). It's not like listening to music.
That's a hypothesis but not evidence. I can present a counter-hypothesis: I fall asleep while listening to music (or staying in silence). Listening to spoken word keeps me awake or at least helps me notice that I'm getting tired.
1.2 million people die in road accidents, and most of them are children and young people. Even more are seriously injured.
The best ive heard but still too monotone over time compared too real productions. Feel blown away at first but listen a chapter or two gets difficult. Just a matter of time most likely until it becomes as good or better then the real thing.
I've been looking for a good and convenient way to read papers that are published in PDF for a while.
Ideally, I'd be able to strip out the text content and send it to my kindle in readable form. Since apparently that's science fiction, this looks like a really good plan B! Will definitely give it a go.
This is excellent. I just tested the Finnish voices on my simple news archive [1], and the pronunciation was quite good and clear.
It's unfortunate that I can't export audio clips locally; otherwise I would immediately look into using this for generating my Finnish flashcard decks from the same material [2]. I've thought about doing the same with the audio and video feeds included with this news broadcast, but getting Whisper to sync up properly with what's written down and cutting up the raw audio in that way still seems like more effort than I'm willing to invest right now.
> It's unfortunate that I can't export audio clips locally; otherwise I would immediately look into using this for generating my Finnish flashcard decks from the same material [2].
elevenlabs has an API which seemed quite reasonable when I looked into it. A bit of python should get you what you want pretty quickly.
If you are looking to convert very short texts or words into speach, I had best result with eleven_multilingual_v2 with the following text for tts "Hän sanoo rauhallisesti ja hitaasti: <break time=\"1.0s\" /> '${text}'" An then i use a postprocessing to split at the silence.
This was nessesary as you cannot set the language explicitly and it is detected from the input.
With eleven_turbo_v2_5 you can set the language, but the results are not as good.
This is a cool repo. Interesting approach using uralicNLP for morphology, that's not one I've seen before. This repo's README.md is excellent and thorough too - I'll probably come back to this in March and give it a spin for myself, just to see what you're up to in a little more detail.
I know I'm growing old but this is the kind of tech application that I don't like. Arts should be the last thing to be 100% fully done by a program. Enhancing capabilities in artists? Hell yeah. Replacing completely voice actors? No, thanks.
I understand, It can do things that weren't previously possible, but it will also replace things that were done by humans, by artists before.
Overall, in my opinion, is still a loss.
"replace things that were done by humans" isn't a loss by itself, if it frees up human labour to do other things.
If human replaced by AI can't find better things to do, such that it makes them poorer, or anti-social its a loss but not necessarily AI's fault.
Doesn't apply to all situations, but "replace things that were done by humans" in arts can absolutely be a loss by itself. Making graphics/speech/video a commodity doesn't replace designers, voice actors, or directors, but we've definitely see it can directly harm them and the people that enjoy their work.
> can't find better things to do, such that it makes them poorer, or anti-social its a loss
I feel like this misses the point a bit - lost income/sustainability for artists is obviously a big issue we'll be facing, but looking for a performance indicator in an artistic endeavour doesn't really get you anywhere. There's more ways to value a painting than "what the market would pay" and "potential heat output as firewood", right?
How do you feel about replacing general labor, period, and doing so for a class that no longer maintains a semblance of a social safety net? Do you think there's a difference between displacing one profession and displacing most professions at once?
Do you people ever step out of the abstract and think about the actual context you're living in?
I will gladly pay taxes directed for retraining artists, but I will not pay to listen to Wil Wheaton narrate another book badly when my computer can do it better.
I mentioned typists, you abstracted it to “most professions at once”, and you give me a hard time for being too abstract?
I agree with your criticism, just not sure you understand who you were criticizing. But I hope you can think about actual context and see if that tempers what seems like a pretty emotional take on AI.
I think calling this art is a stretch, as they usually aren’t the author.
By automating it, it lowers the barrier to access this type of audio content for the masses. If you want to choose to pay someone you read something for you, the market allows that. This feels like a net gain.
If the AI content is good enough, nobody will use it, or at least not in the numbers that Audible et similia had before. It will just be a tiny minority following their principles.
We lived this already with social networks. Initially us tech enthusiasts were all like "it will democratize access to news, it democratize producing the news! curated work will still be there, it's a net gain". And we all saw how it actually developed. As someone on the Internet said, I want AI to do my laundry and repeating task so I can do art or other more interesting things, I don't want AI to do arts and force me to do laundry by hand because due to AI taking my job now I don't have money to pay for a washing machine.
> I think calling this art is a stretch, as they usually aren’t the author.
I can't even remotely agree.
Narrating a book is absolutely an art. Listen to a book narrated by Stephen Fry, and all other books will sound awful. Considerable care and craft goes into a well-read book.
But this is why I'm actually excited about good TTS tools. Not because I want to displace Stephen Fry, but because there are so many books read by awful narrators and something like ElevenReader would be a huge step up in quality.
I share the parent commenter's concerns about the displacement of artists, but I'm less convinced that TTS tools are a net negative.
I feel conflicted about this. I somewhat agreeing with you, but the other hand not needing voice actors is a big help to people with disabilities that prevent them from reading.
Would've been valid if TTS was, indeed, art, but it's not. Audiobooks won't be able to replace TTS in e-readers just because they need to be produced first. And I don't think my mom would be able to find an audiobook of all the Russian books, or, especially, articles she's reading, and especially synchronise it with the actual book in her reader app.
Of all the criticisms leveled against GenAI, I'd say making the case against "TTS on-demand" would probably be the weakest.
Having natural sounding TTS enhances accessibility for blind users, enables language localizations, etc. It's 100% a win even though there will be (and already is) disruption in the VA community.
Hasn't this been around for ~4 months? Interesting to see this here, since their competitor Zyphra, just released two Apache 2.0 licensed open weights TTS models yesterday[1].
Personally, I can't stand my voice when I hear its recording. I wish there was a way to easily tune it to sound more like what you hear. Maybe even use that adjusted voice during calls.
182 comments
[ 4.3 ms ] story [ 225 ms ] threadNo audiobook exists, drop epub into ElevenReader and have Bert Reynolds read it to you, honestly better than some human narrators.
Companies won't stop pulling this garbage unless we stop supporting them.
Open source models drive proprietary foundation models' margin to zero.
The only reason elevenlabs became a unicorn was their margin. If they became a commodity, they'd find themselves in a deep pit.
I know the decisions of the Dev team were disappointing, but it's also worth pointing out that the site was kept up until around last month - despite the warning stating that'd be down in November.
Omnivore could have shut down their code base, and prevented self-hosting entirely. I'm glad they didn't.
As for contribution model, it’s still something I’m trying to figure out. For the moment, it was just trying to get a self host build ready and working.
But I have admin rights to the repo, and am not working for ElevenLabs, nor officially Omnivore. I was just a contributor before.
That said, even in their cherries the emphasis still isn’t quite right in the Tolkien example.
Since then, they’ve released a few cheaper models, but the quality suffers greatly (they still have the old models though so it’s not an issue). They’ve also been releasing a ton of different products around TTS.
I don’t mean this as a criticism — I just am curious why SOTA TTS has not improved from one model by one company several years ago, and why even said company isn’t able to improve on that model.
Which is why it's especially ridiculous ElevenLabs allows professionals to upload their voices, charges users of those voices a minimum of $50 per million characters, likely pays under $1 for the compute... and then passes on a whopping $2 back to the professional.
I think the next disruptive TTS competitor is going to form out of just offering to pay better rates than ElevenLabs to their PVC users.
Finetuning established architectures on cleaner synthetic data is already getting open source models increasingly competitive, so getting top PVC samples from the source would likely put you right about where they are today.
Edit: And since you're concerned we might not be aware of Elevenlabs' generous terms... why is your documentation so cagey about them? https://elevenlabs.io/docs/product-guides/voices/payouts#thi...
I see users need to keep paying you a subscription fee in order to even get their payouts... but "up to 20%" isn't saying particularly much without the kind of details that should probably be on that page.
-
Considering how much your company owes to an open source model, it's also impressive how little you've returned to the commons.
But no worries, the top comment under this post is an open source model that was finetuned for a couple of thousand dollars by a single dude soliciting the public for random voice samples.
If Google has no moat, you're out to sea.
Maybe you're in a bubble devoid of that kind of thinking, so it seems very foreign or quaint.
Even then it's short-sighted thinking at best: the "market rate" is not some magic self-optimizing number.
Underpaying their creators is just creating the opportunity for someone to take the best of them on better terms.
-
Elevenlabs is also able to raise trivially in this environment: you'd think while they're still floating out here without a moat other than high quality data, they'd overpay if anything and make narrators feel like royalty until they're replaced.
This isn't unlike Uber initially paying drivers massive bonuses and undercharging riders until they were able to leverage their massive network to increase prices past what the taxi providers they had decimated were charging. But in this case the marginal cost of providing the service is so low they don't even have to lose money to run a similar play, just take less of it. (in other words, even ruthless greed is not antithetical to paying these folks better)
I hire someone to paint a fence. We agree on $200 for the job and I pay them $200. We both know that it’s undifferentiated work and I could find a dozen other people who would do it for $200.
Where’s the lack of integrity? Or does it just appear if I know that I actually could afford to pay them $10,000, but chose not to?
You're building fence painting robot and need someone to teach it how to paint fences by example.
You decide you won't pay the fence painters to teach the robot upfront.
Instead, painters will pay you $20 to even visit the factory.
Then, if a particular painter's fence painting is especially highly rated, you'll pay them a small royalty.
So you send your fence painting robot to compete with fence painters for $20 a fence, passing on a tiny slice of the $20 to the ones who helped teach it.
-
We can consider the creation of fence robot and the competition with the existing market just another piece of the steady march of progress, but there's still obvious room to act with more integrity in this situation.
There was no established rate for what fair wages to teach the robot to paint were, and you can't pay $200 because you don't charge $200... but it's also probably not "-$22/month + 1% royalties".
So much of my time for "reading" is in a context where I can't physically read, so audiobooks are incredibly useful. But being limited to the set of books that gets recorded by the publisher is a real shame.
Haven't tried it yet but AI TTV seems basically perfect now so I'm very optimistic this will work great.
Edit: I think the effect of the invention of vinyl on live performers is more akin to how the commoditisation of HQ TTS will be detrimental to audiobook narrators.
Two audiobooks that come to my mind:
- The Lord of the Ring series read by Andy Serkis; not only he perfectly switches between each characters voice, but also the feeling of listening "Gollum" for ours is something else altogether
- David Goggins' books; the audiobook version is completely different than the book, since he's not just reading the book, and overall it makes the content easier to digest
https://chasingperfection.co.uk/post/2013/01/14/text-to-spee...
https://claudio.uk/posts/audiblez-v4.html
I can go back and try to repro and get a recording....
Umm, it does.
> We don't currently support Apple Silicon, as there is not yet a Kokoro implementation in MLX. As soon as it will be available, we will support it.
I thought that meant that it didn't support Apple Silicon in general, but they were just talking about GPU support.
I might try using F5-TTS-MLX instead actually (https://github.com/lucasnewman/f5-tts-mlx) and see how that does.
Interesting, because the hero image is a Mac App screenshot.
> Is the app free?
> Yes. The app is completely free to download and use today. Listening to content on the app will not consume credits from your monthly web plan. We do plan to eventually launch some premium version of the app, but even then we will maintain a generous free plan.
Can anyone else confirm?
[1] - https://unicornriot.ninja/2024/sextortion-coms-inside-a-vile...
I use it to listen to PDFs. It works, but has plenty of hiccups with headers, footers and colons.
Does anyone here actually have positive results doing this? It seems to me listening to anything that's even remotely complex with the intent of learning it just isn't something that's feasible.
I get easily distracted and lose attention while listening to an audiobook. This is usually problematic with fiction, because suddenly I don't know who this new character is or what's happening. And rewinding to the precise position where I stopped paying attention is of course much more difficult than in written text.
I found that non-fiction books work great for me, because even if you ignore a page or two it makes no difference, the author keeps repeating their point and propping it up with many arguments anyway.
But just to clarify, do you listen to dense material on commutes or while on walks or are you listening and taking notes at a table or something?
That's my main point. As someone that doesn't have dyslexia, sitting at a table and taking notes while reading dense material is already quite difficult to do when I'm trying to actually learn the material. I couldn't imagine being effective in my learning by just listening while on an early morning commute or something like the promotional video shows.
you could build your local TTS using kokoro browser though — https://huggingface.co/spaces/webml-community/kokoro-webgpu
Could you expand upon this? Any milestones towards that which we should be mindful of?
With llms, "knowing things" is already starting to feel like a thing of a past, not to me, but to a lot of others, there's no longer an incentive to "switch on".
Why should a kid learn anything if a robot is instantly better at everything? Maths got replaced by calculators, deep critical thinking will get replaced by llms a lot of the time, which are word calculators, which is the closest thing we have to a logic calculator.
This is more passive autopilot software, which further promotes learning as something you 'consume' rather than something you seek.
The public consciousness has absolutely taken a semptember 11 tier nosedive since social media, we're approaching what I term cultual schizophrenia, which I posted about on my blog which I deleted, but I've readded it if you're interested [https://substack.com/home/post/p-156983317]. There's no contextualisers in the media to give the right emphasis to the right things.
This is just my perspective, from what I've seen from other younger people of my age. We are heading into extremely interesting times, everything profoundly destabalising thing we've speculated about is happening at the exact same time. We desperately need visionaries in politics.
Basically I'm not doing too hot
The world is changing, but then again it always has been. IMO some things will get better, some will get worse, but ghe overall arc of human health and prosperity will continue upwards. There is less poverty, less starvation, more opportunity today than ever… even though some aspects of the world are bad and getting worse. That’s the way it’s always been.
E. g. saying "1963" when the actual year in the text was 1967. Yeah, the voices sound very realistic. But I'm not sure how useful that is if you can't trust the spoken words.
Does anyone know if it got better in the last weeks?
1.2 million people die in road accidents, and most of them are children and young people. Even more are seriously injured.
If that's the case, maybe a driver's license isn't your thing?
1.2 million people die in road accidents, and most of them are children and young people. Even more are seriously injured.
Ideally, I'd be able to strip out the text content and send it to my kindle in readable form. Since apparently that's science fiction, this looks like a really good plan B! Will definitely give it a go.
[1] https://kindlemodding.org/jailbreaking/WinterBreak/
[2] https://koreader.rocks/
It's unfortunate that I can't export audio clips locally; otherwise I would immediately look into using this for generating my Finnish flashcard decks from the same material [2]. I've thought about doing the same with the audio and video feeds included with this news broadcast, but getting Whisper to sync up properly with what's written down and cutting up the raw audio in that way still seems like more effort than I'm willing to invest right now.
[1]: https://hiandrewquinn.github.io/selkouutiset-archive/
[2]: https://github.com/Selkouutiset-Archive/selkokortti
elevenlabs has an API which seemed quite reasonable when I looked into it. A bit of python should get you what you want pretty quickly.
If you are looking to convert very short texts or words into speach, I had best result with eleven_multilingual_v2 with the following text for tts "Hän sanoo rauhallisesti ja hitaasti: <break time=\"1.0s\" /> '${text}'" An then i use a postprocessing to split at the silence.
This was nessesary as you cannot set the language explicitly and it is detected from the input.
With eleven_turbo_v2_5 you can set the language, but the results are not as good.
> can't find better things to do, such that it makes them poorer, or anti-social its a loss
I feel like this misses the point a bit - lost income/sustainability for artists is obviously a big issue we'll be facing, but looking for a performance indicator in an artistic endeavour doesn't really get you anywhere. There's more ways to value a painting than "what the market would pay" and "potential heat output as firewood", right?
Do you people ever step out of the abstract and think about the actual context you're living in?
I agree with your criticism, just not sure you understand who you were criticizing. But I hope you can think about actual context and see if that tempers what seems like a pretty emotional take on AI.
By automating it, it lowers the barrier to access this type of audio content for the masses. If you want to choose to pay someone you read something for you, the market allows that. This feels like a net gain.
We lived this already with social networks. Initially us tech enthusiasts were all like "it will democratize access to news, it democratize producing the news! curated work will still be there, it's a net gain". And we all saw how it actually developed. As someone on the Internet said, I want AI to do my laundry and repeating task so I can do art or other more interesting things, I don't want AI to do arts and force me to do laundry by hand because due to AI taking my job now I don't have money to pay for a washing machine.
I can't even remotely agree.
Narrating a book is absolutely an art. Listen to a book narrated by Stephen Fry, and all other books will sound awful. Considerable care and craft goes into a well-read book.
But this is why I'm actually excited about good TTS tools. Not because I want to displace Stephen Fry, but because there are so many books read by awful narrators and something like ElevenReader would be a huge step up in quality.
I share the parent commenter's concerns about the displacement of artists, but I'm less convinced that TTS tools are a net negative.
So I guess in your worldview a concert violinist also doesn’t make art, when they are playing a Mozart composition?
Having natural sounding TTS enhances accessibility for blind users, enables language localizations, etc. It's 100% a win even though there will be (and already is) disruption in the VA community.
[1]: https://news.ycombinator.com/item?id=43004589
https://www.youtube.com/watch?v=ua4rYsMdC4U
I found that it’s my preferred way to use their reader, as it makes the reading more neutral and transparent for my brain.