Studio is now available to everyone, with new features (walkthrough)
ฝัง
- เผยแพร่เมื่อ 11 ก.พ. 2025
- Try Studio now: elevenlabs.io/...
Studio, our longform text-to-audio editor for creators and storytellers, is now available to everyone.
Bring your audiobooks, voiceovers, articles, and even podcasts to life with multiple characters, fine-grained voice customisation, pacing control, and more.
We first launched Studio in 2023 under its previous name of Projects, and it was only available to users with a paid subscription. Since then, we’ve released dozens of new features and upgrades, and now we want everyone to experience the magic of Studio, regardless of which plan you’re on.
Anyone on our free plan can now create, edit and export up to three projects in Studio. When you upgrade a paid subscription, there are no limits on how many projects you can make.
MB01ED1AXPKEIQN
![[UNCUT] ขยี้ "อลิส" ปม "คลิปลับแตงโม" เปิดชื่อคนพยายามแฮก "มือถือแตงโม" I คนดังนั่งเคลียร์I 10ก.พ.68](http://i.ytimg.com/vi/_RPLEIi5Sz8/mqdefault.jpg)
Try Studio now: elevenlabs.io/studio
Please fix your credit system. I end up wasting a lot of credits before getting the right output
They need to working on giving us full atomic control over each word in terms of voice intonation, prosody, cadence, volume, pitch, etc. It is very difficult to make these voices sound natural and takes a lot of manual work and re-generations/credits to get it just close enough to what you want. If I hear another monotone voice again, I’ll go crazy! You can get OpenAi’s voice to sound extremely natural live when speaking to it. Perhaps eleven labs need to have a feature like that where you talk to the LLM through the voice and actively tell it how to modulate itself, similar to how a voice actor in a studio is directed to modulate their voice by a director for an animated show. That... would be ground breaking.
Yeah...They actually have this now with their 'Agents'. I spun up an actress agent who i'd coach through her lines... at 500 effing credits per minute. It did not work as well as I hoped. Not much better than 'She exclaimed with SURPRISE "Oh no, a monster!'"'. Just a terrible experience that I'm glad I only subbed to for 1 month.
Wonderful, thank you for sharing ❤❤❤
Very good improvement! I hope that soon you would give some options to reduce the speed in some words because the system sometimes accelerate the voice in just 1 word :D
How many credits does it use up getting it right tho? You don't mention the credit usage?
Same ! How can I listen without use all credit
The answer is A TON! Just like the regular text to speech. Although this is better at keeping the same prosody/tone throughout different sentences and paragraphs, even if you generate each paragraph separately. That is....most times, anyway...
You can see he has 3 BILLION credits in his account. At 11 dollars per 100,000 credits that works out to $0.00011 per credit, and at 3 billion credits his account is currently loaded with 330,000 DOLLARS worth of credits. I don't think he cares much about needing a gazillion rerolls.
@ He is working for the company after all. So he can get infinite credit if he needed.
Probably a billion credits
Three BILLION credits? I just picked up the 100K plan a few days ago and just two sentences costs 300 credits. This looks like a neat feature, but it looks like it will cost a fortune... Just going by the Conversational AI cost of 500 creds a MINUTE! Jesus we're back to 900 number pricing. That's a non-starter too. And the inflection in eleven labs is way off. I have to reroll 4 or 5 times to get a sentence to sound right. I spent 10K credits getting a paragraph of audio.
The pricing is so crazy yet the output is nothing compared to notebook llm. I use notebook llm to create the podcast and use a different awesome AI tool to easily edit the audio and add any background music, inflections, etc, with simple prompts.
1:31 The voice should’ve been trailing off….still not ready for professional use. You seem to not understand acting, and how actors behave with lines. I suggest you to include in your team, casting directors, directors, first ADs, and as many good experienced actors as you can get to fully understand how to deliver lines. How to pause, enunciate, trail off, and so many more options you have in the performance world. To me, the demo you showcased, is so unprofessional that I have problems to believe you’re presenting this as an outstanding feature.
Plus one to this... I can add 'Such and such said enthusiastically "Oh boy we're going to the beach!"... and that helps, but I still need to reroll a ton of times. This AI has no acting chops.
+1 they have stagnated for months now. Another player that gives full atomic control over each word’s delivery is soon to creep up, probably within a few months. And I am sure to jump ship. It’s what I’ve been wanting from them for like a year now.
Podcast unfortunately still is not even close to as good as Notebook LM
Yes! The pricing is crazy yet the output is nothing compared to notebook llm. I use notebook llm to create the podcast and use a different awesome AI tool to easily edit the audio and add any background music, inflections, etc, with simple prompts.
True
@@RafaelChibuikewhich tool do you use?
100% AGREE! And the customise settings feature doesn't even work as well as NLM
I get HTTP error when I try to make a podcast...
How many credits and what is the cost?
I got so many issues with my Professional Voice Kurt. Every time I make a TTS my Voice is loud and strong on the first 2 minutes and after that the voice sound very weak and the loudness getting down ,
its not clear how you download, is it piece by piece?
Right, I can't figure that out either. I'm wanting to record an article from my website. It just seems to do a piece at a time. I guess you download paragraphs and join them all together.
Hi, nice features! Cn you help me getting consistent pronounciation for a german text? I struggle getting the A.I. reading words consistent in the course of the document. Names are pronounced differently over an oder again. I tried spelling them differently to emphasize vowels and keep the pronounciation consistant, but it doesn't really work. Am I missing something here? Maybe you have tools or settings on your site, I could use? Please help me out with that. Thanks in advance for any idea.
I fear we may have reached Eleven's upper limits of AI voice capability. 2024 with nothing but UI features and no emotional conversational voice improvements while competitors crept up, and the larger player surpassed.
Agreed. Hoping they can get going, because the monotone voices are so annoying now.
What alternatives do you think are out there that are better?
Wouldn't it be great if the model used the context of the conversation to use tonality. Like it said "politely" and it was not polite, this surely is the next step
Nice. I am fighting to make latvian language voice clone. Resulted so so lat-englished...
Any ideas if customers text is digested by any of your llm models for model building and enhancing? Ie will you use my text?
Start with the costs!
RL@FT?
Hey, been using ElevenLabs and LOVE IT.
Unrelated question, though, what software was used for the screen recording of this walkthrough? The mouse movements are so smooth and fluid...how'd you do that? Is it an OS setting? Post-production software?
screen.studio/
@@elevenlabsio B-E-A-UTIFUL, thank you
lots of trying - each time you pay with credits. the thing is, I know what I want but AI screws around and at the end its a lucky shot ... hmmm. Quality of the speakers still not at the Google Notebook LM level.
You guys should make it so credits are onley used when you actually download the file. Because now you just use 50% of ceridts on tries. Its a waste of money.
Not good yet, nothing intuitive on the matter of controlling de emotion, tune and the voice pitch…
Watch this space!
Nice, now give us MUSIC!