Get better sounding AI voice output from Elevenlabs.
Вставка
- Опубліковано 21 лип 2024
- 🚀 Unlock the Full Potential of ElevenLabs Text-to-Speech! 🚀
Dive into the ultimate guide to mastering ElevenLabs' AI voice synthesis.
👉 Try ElevenLabs for Free:
excelerator.tech/getelevenlabs
In this video, we'll cover everything from selecting the right voice and model to fine-tuning settings for emotion, pace, and pauses.
Let's demystify the art of creating lifelike voices with ElevenLabs, ensuring your text transforms into captivating audio that sounds almost indistinguishable from a real person. We'll explore:
🔵Selecting the Right Voice: Understand how matching the voice to your project's style makes a difference.
🔵Choosing the Best Model: Navigate through ElevenLabs' models to find the perfect fit for your needs.
🔵Mastering Settings: Learn how to use the stability and similarity sliders, style exaggeration, and speaker boost to enhance your voice creations.
🔵Smart Prompting Techniques: Discover how to nudge the AI for optimal results, including how to add natural pauses, slow down the speech, and inject emotion.
🔵Practical Examples and Tips: Get hands-on with easy-to-follow examples and insider tips to elevate your voice projects.
No tech wizardry required! This guide is designed for anyone eager to leverage the power of AI for voice synthesis. So, if you're ready to bring your text to life with unparalleled realism, press play and let's embark on this audio journey together.
👉 Try ElevenLabs for Free:
excelerator.tech/getelevenlabs
Start experimenting with these techniques today.
As an affiliate, I may receive a small commission, at no additional cost to you, if you follow my link and end up making a purchase. I sincerely appreciate it!!
#ElevenLabs #TextToSpeech #VoiceSynthesis #AITechnology #AIvoiceover
Chapters:
00:00 Demystifying ElevenLabs text-to-speech
00:45 Selecting the right voice
01:32 Differences in Voice Models
03:52 Settings (Sliders & Speaker Boost)
09:07 Adding a Pause
12:17 Pronunciation
14:36 Emotion
17:10 Pacing
19:42 Go create
I'm new to ElevenLabs, and this is the best tutorial I've seen so far. Thank you.
Glad it was helpful!
I like how thorough you explained things here. Great video 🌟
Thank you so much!! Glad you enjoyed it!
Thanks for the care and time making this.
Glad you found it helpful!!
Love ❤ from India bro, it really helped me to improve
Happy to hear that!
you got a brand new subscriber, congrats. Nice Explaning style, keep making vids like this.
Awe, thank you!!
Absolutely... This API needs to include more hyper text esque instructions... Would be SO MUCH EASIER to give the AI instructions like etc.
Agreed! The very first text-to-speech tool I used 3-4 years ago had this for their premium voices and it was really nice. Not sure why the current TTS platforms don't offer it.
Great job very helpful and meaningful!
Thank you so much!
Very helpful, thanks.
Thank you so much, very helpful!
Happy to hear you liked it. Thanks for watching!
Great Video, Thank You!
Glad you enjoyed it!
I'm now using Speech to Speech for pronouncing specific words correctly, like foreign names and stuff like that, then I stich it to the rest in an editor to get my final result.
That's a creative solution! Very impressive.
Great video! What do you use for video editing
Glad you liked it. I used CapCut to edit this video. I've used Premiere Pro, PowerDirector, Movavi, Camtasia, Descript, and others... but CapCut has become my go-to.
You're good bro. thanks
Glad you liked it.
I'm glad I'm not the only person to experience language switching during generation.
If anything weird can happen, I'm pretty sure it happens to me. I'm lucky that way. 🤣
@@excelerator same
The Break Time Syntax worked better than all of the other pause methods for me.
That makes sense, given it's the one "hard-coded" method that doesn't depend on the robots to think.
what i dont like about their payment model is that, you get charged even if result is unusable. even when you want to add emotion, when you enter shout or yell it will count on your credits. they should set an emotion slider that doesn't cost the user.
True. From their perspective, I guess they incur cost every time we generate so that's how they charge. Sort of like old film cameras - we used up film on a lot of pics that didn't turn out but still had to pay for the whole roll, and pay to get all of them developed.
I'm not arguing for their pricing model, just trying to understand it. But I agree if we had something simple like emotion sliders or tags that would help us get better results on the first generation it would be more efficient... and cheaper!
Does it make sense to let 11L first comvert the whole script/chapter at once to give it maximum context? (And then work on each paragraph one by one). Thanks for the video!
Maybe. Elevenlabs suggests going one paragraph at a time. I don't know that it will look beyond the paragraph for context, but it may be worth trying.
@excelerator Thank you for taking time to answer 🙏
.... is there a space before the /
I liked the way you explained it all .... thanks
Yes, there is a space before the /
Glad it was helpful for you!!
@@exceleratorI tried this but I can’t seem to get it to work. Any advice? Thanks,
I'd suggest just double-checking how you have typed in (make sure no typos or anything.) Maybe try to test the extremes, like changing 1.5s to 3s (something that would really stand out and let you know it understood what you're asking for and then try reducing it.
Also, try a different voice with default settings - just as a test to narrow down why it's not following your instructions.
If the "code" isn't cooperating, you might also try making sure you have a period at the point you want the pause and then hit enter on your keyboard to get the next part on a new line, since it seems to understand a break in paragraphs should have a pause between them.
Yo do know we get charged per generation 😂
Sadly, yes. 😉
@@exceleratorDo you know Murf Ai? We don't get charged every generation there if we change or try other voices to read text (as long as the same text), but we get charged on Elevenlabs every generation, even with the same voice and same text. So, Murf Ai is better in terms of saving generation etc.
Good info. Thanks for adding that. Are you getting good results with Murf? I've tested a bit and wasn't really impressed. But maybe I need to try again.
These companies should only charge if one downloads the file. It's robbery taking our credits when trying to simply get a correct generation.
I've thought that too. But, I guess what costs the company is the generation (more than the download.) And I'm pretty sure if they charged by download they would probably make it 10x the credits they charge to generate to cover a bunch of generations. I'm just speculating though. I'd like everything to be cheaper!!
thanks - changing languages when words are similar is annoying
It sure is!!! 🫤