ChatGPT to Voice: AI Voices Are Getting CRAZY Good!!
Вставка
- Опубліковано 1 бер 2023
- ChatGPT API into Voice: AI Voices Are Getting CRAZY Good!
Become a member:
/ allaboutai
Join the newsletter:
www.allabtai.com/newsletter/
OpenAI just released their highly anticipated ChatGPT API, and the possibilities are endless! In this video, we're going to show you how to take advantage of this new tool by combining it with Python and the AI Voice API from Eleven Labs. With this powerful combination, you'll be able to turn all of your text outputs into human-like voices!
Imagine being able to turn those large, tedious PDFs into engaging audiobooks or summaries, or even being able to automatically summarize meetings into concise voice summaries - all with the power of AI!
www.allabtai.com - Наука та технологія
Oh my goodness. WinAmp is still alive???
Thats what I saw!!!
Haha 😅 Winamp ♥️
I still use it, it's the best
Ohh yeah .. i also noticed that 😂😂
Omg! I bet he have mRIC too...
I feel the next phase in this is the ability to have multiple voices for one render. Can really have fun with creating audio stories that can have dialogue between characters. Put some sound effects in there - and you have an audio book with a punch.
Yeah:) the ideas and use cases are just to many to think of atm 😅
This would be amazing! Yes, the stories would be much more engaging and as an English Second Language teacher, I know my students would love imitating the dialogue used by their characters. Please please please
Voice Actors enter chat in 3..2..1..
You could break up your story dialog into an array, assigning each items key as the relevant characters ‘voice id’.
then run through the array submitting each key/value pair to function returning each file and naming it in order of playback.
This would take Some of the manual labour out of the process 😊 but I’m not aware of any SFX apis available yet 🤷♂️ maybe next week there will be at this rate 😂
need some tone and emotion in the voice too
Outstanding. There is a lot to digest here and you've made the framework as easy as can be! Thanks Kris, I'm going to do a deep dive into this later tonight and test it out. I'd love to see a video on the Eleven Labs API.
Thanks Tony :) yes i will do that soon!
It would be great to implement this and use the AI's ability to detect when a different character is speaking and use a different AI voice to match that character's description or, if it's been trained on it, YOUR OWN voice as the character. Or your child's voice as the character in the story.
Yes!! That is a great idea:)
Super possible with the new ChatGPT prompt prefix
What do you find interesting in using your own/your kid voice as the character? I'm really curious!
That was my first thought as well, when i watched the video. Would be great!
@@AllAboutAI Question for you, how or where did you get from the black page with commands to input text and generate audio? thanks
Great video and the Winamp makes me feel all warm and fuzzy.
I LOVE all your videos man. Very informative, I've got the bell turned on! PLEASE have the next video be on the ElevenLabs API!!
Thnx a lot Justin :) Yeah that will happen
Well for jumping into Chat GPT API so quickly!!!! Your video is awesome and use cases you demonstrated just shows your wonderful curiosity, creativity and imagination! Being shown how to change the voices will be awesome and of course how to use the voice providers API. I assume the API has a daily or monthly character cap?
Thanks a lot Johnny :) hmm, i have not checked the cap yet, will do tho!
Thank you for sharing this video about turning AI-generated text into voice! It's incredible to see how far technology has come and how it can make a positive impact on people's lives. This has the potential to open up so many opportunities for individuals who may have difficulty speaking or communicating, and it's heartwarming to see technology being used to make a positive impact in the world. Keep up the great work!
No problem :) yes!! This has so many positive use cases i think!
Wow! I'm really impressed that you were able to generate a summary from a 117-page PDF. I've been hoping to use CHATGPT to read long reports like this. Can you tell me how you did it? I don't know anything about coding, but I was wondering if you could develop this summary function into a standalone software?
Hello :) i cant explain that in comments, but check out my membership if you wanna know how:) and thank you for tuning in!
@@AllAboutAI Where is info about your membership?
@@andrewsnavely join his UA-cam channel membership
AI, new technologies and Winamp ❤
Much respect for using WinAmp...it really kicks the Llamas ass. :-)
Amazing stuff Kris! That Winamp skin is so relatable lol Napster era mate!
Thnx:) haha, that skin brings back a lot of memories yes:)
Excellent video!
There are just 3 youtubers who are making ai videos to this intensity, and you are one of them.
Is that a good or a bad thing?:)
@@AllAboutAI damn bro, that's an amazing thing for us ai enthusiasts
Great video... I didn't find links to google collabs you presented. I liked PDF summary in 2 minutes
Thnx :) I have tutorials and step-by-step videos on this on my Membership page if you are interested :) ua-cam.com/users/AllAboutAIjoin
As alwayes Kris nailed the how to use those tools in such a good way .. i wanna see video about whisper too and use cases for that .. thanx
Thnx mate:) for sure! Got something planned already!
Great as always:)
Thnx as always micbab:) appreciate your support:)
Cool program you built. And btw, the movie Frost is actually called Frozen in English :)
tnx :) haha damn
Damn that sounded really good
Amazing work. I was looking at the repo and dont see all the code you are working with in the video. Where might I find that? I was only seeing the first example you gave.
Thnx Kyle :) i have not had time to upload all to the GitHub repo yet. Soon tho :)
One quick question. Videos like this, adventure and sci fi stories and backstories, using generated AI images and voiceovers are monetized ?
Great tutorial, thank you for sharing
I got puzzled by the instruction to ignore previous instructions. I believe that in the API there are never previous instructions. You either include the whole thing in the request or you don’t.
Anyway, incredible work you are doing here!
Thnx Roberto :) I guess it is just an old habit haha, i think you are right tho
@@AllAboutAI it’s an irrelevant detail but might increase the use of tokens if integrating on a large scale.
P.S.: You are sure on the way up and this channel will blow anytime now, so probably my last chances of an interaction XD
Thanks for inspiration! Did you share Google colabs shown in the video somewhere? They may be quite helpful.
Thnx for tuning in :) They are a part of my membership, check out link below:)
Do you have any references for how to set this up initially? Anything would be appreciated!
Love the video. Requesting video on how to use "Eleven" API as well. thanks!
Noted :)
Liked and Subscribed. Wish shared growth.
As usual amazing. Constantly learning something new and getting smarter day by day. By the by some videos can't access. How to access, is it paid?
Cool :) You can find the membership in the links below if you are interested in learning more!
5:50 is like someone from BBC who talk about animal...😄
Amazing content mate! It would be great to have a video covering how to use the Elevenlabs API! Do you think it is possible to run chatgpt in real time with Elevenlabs, generating voice on the spot?
Incredibly interesting, and so many use cases ;-)
Yes!! Thanks for tuning in :)
Thanks for the great video! Question for you. Do you think Eleven Labs API (or any other company) is currently able to generate either text to speech or speech to speech that is good enough for comedic performances? For example could it generate usable voices/performances on a show like the Simpsons or Family Guy?
Question: Many very good books are available in PDF format, from the early 1900's. Normally I just use a TTS app to listen while at work, but it read everything in a monotone manor. Would it be possible to input an entire book and get chatGPT to remove all the extra spaces, dots and all, correct the grammar and then convert that into a easiy to listen AI voiced MP3 file? This would mean that books from 1800-2000's suddenly can be listened to, would love to hear your input on this. Thanks for the great content.
Hello :) Yes that is possible for sure. But i will be a bit expensive tho. I would test it out with a small sample first :) If you contact me on mail i might help you
This is the same problem I'm having. For years, I've been using Balabolka with Zira voice option to convert old books into audiobooks. But, I'm looking for something new, with Natural Language Processing. Is there any open-source project/library that I can use? I'm proficient with coding too, but I just can't find anything that wouldn't cost me a thousand dollars just to listen to a simple audiobook.
Love this video
Thanks
I have just discovered your channel! congratulations for your work! I would like to have my youtube videos with the voice in different languages. I've got the subtitle file with Whisper and I translated the srt file the Google's API. I would ask you an advice. Is there any Text (srt file) to speech free tool that you recommend? Thank you very much in advance!
Excellent video!
Questions, I've being use the new Bing to summarize PDFs, specially Edge side bar. I wonder if there is any advantages on using the Chatgpt API versus Bing.
This API calls you made are linked to your account and have an associated cost correct?
Also since Bing uses a new version of GPT I believe the answers are better, although limited in characters. Does the API already have parameters to return more characters or to have a large memory to store more context?
Thanks!
Hello! Since Bing is so limited to ppl yet i have not botherd testing it. Yes the API is 0.002$ per 750 words. Kinda cheap tho. With a Python script you can work around the character limit, so that is the big advantage:) tutorials on my membership if you wanna know more
@@AllAboutAI thanks for the quick answer. A follow up question about memory, I notice that chatgpt plus have a memory limit of about 4000 characters before starting to loose context. Your demonstration of summarizing Ex Machina script goes way beyond that limit, would you say that the API have a larger memory compared to the Chat?
My main use case with this tech is summarize multiple scientific papers and then compare the summaries.
@@AllAboutAI reading the documentation for the API looks like there is 8000 token memory limit for context
ahhh the good old winamp which i'm still using too :)
Yes, love it 🤩🤩
I’m curious. How many tokens/words/cost did the 117-page PDF use up to summarize using the new API?
Hello! its was around $0.35
Id just like to say you're a OG for using Winamp.
Thank you very much for this amazing video! May I know where can I get the code you used?
Can the app make phone calls with the summary recorded ?
It'd be awesome about ChatGPT and software development
Are you using the chatgpt api to summarize the whole ex machinima script? Is there no token limit here like on the davinchi model?
Yes :) yeah but i use a trick with python that works around it :)
What did you use for voice-to-lip sync? It looks good :)
The Girl?:) that was from D-ID video
@@AllAboutAI I did not expect such a fast response. Are you sure you haven't plugged in ChatGPT to answer people haha. Just kidding, keep up the good work!
Did you take any specific approach to address the copyright issue, or did you simply use the TechCrunch article as is, without any concerns?
Well it was just an example here. But if i was goona to something commercial i would a credited Techcrunch :)
Greatly explained, could you please do detailed video on Eleven Labs how to use etc.., and could you please share the python code repo to go thru and learn and implement same kind of stuff . Thank you very very much. 🙏🙏
Hello! I have tutorials and step-by-step videos on this on my Membership page if you are interested :) ua-cam.com/users/AllAboutAIjoin
Kris for some reason I am not hearing the 11 Labs voice after running the script, but I clearly see the text output and that the confirmation that the text had been converted to speech but I hear nada?😎
Its is saved in a .mp3 file
do I need a membership to learn this from you or will you upload more videos on how to make this?
Hello :) I do have detailed instructions up now on my membership yes, with access to the script. So for now its a members feature
@@AllAboutAI i really hope you will do a video before someone else does I’m not a fan of subscription based things.
But I have subscribed to you and hope you share it in the near future ❤️
I need to do this, that you so much for showing a peak
interesting so by using the API you can feed more text to chatGPT. what is the maximum number of words by API?
Not exactly, but with the API you can work around the limit:)
how do i set this up in colabs and the coding? If you could leave the code in the description that would have been nice.
Hello! I have tutorials and step-by-step videos on this on my Membership page if you are interested :) ua-cam.com/users/AllAboutAIjoin
@@AllAboutAI will the digital generalist be enough?
how do i access the google colab notebook? do you have a link somewhere?
I am putting all script to the community repo this weekend :)
I am trying to write an API script (code) to have multiple ChatGPT personas interact simultaneously with each other. Basically, have 3 distinct ChatGPT personas (like prompt #1: "You're a 30 year expert in engineering" - Prompt #2: "Acting as a experienced social worker" - Prompt #3: "assume the role of a 18 year old college student") I want these 3 chatGPT personas to have a *verbal conversation between themselves* on a particular subject. This is very tricky, as I need to inject questions based on their responses in real-time, and also need to automate the ChatGPT text to each voice. 11Labs doesn't seem to support this kind of "streaming" input. Any ideas?
Very Interesting. Can you contact me on Discord?
I believe 11Labs just released the streaming input endpoint. Hit me up if you wanna work on it together!
That really whips the llama's ass!! - and the Lamda's ass too :p
haha love the refrence😅 tnx for tuning in:)
Membership is not available in my country, is there a way to get access from your website.
Send me a e-mail and we can talk there:)
How can I make calm and composed voice?
The only way I have ever edited my stories is with audio. How can I try what you have created?
It is a members feature for now :) So if you really wanna try you can join here ua-cam.com/users/AllAboutAIjoin
link to the google share?
Do you have the github code about it?
Can you please provide that code in description
Please make a video about ElevenLab’s API. And you can clone your voice as well as a demo in your next video
Thnx for tuning in :) yeah i will have to do that
You can finally turn Books into Audiobooks if you prefer listening over reading.
Yes:) might be a bit expensive with a whole book, but doable yes:)
@@AllAboutAI why is it expensive?
@@SW-fh7he Eleven Labs voice API is not free
Wow Amazing video. Please do share how to implement the Eleven Labs API
The scripts are not yet on Github? When will be available?
Tomorrow or this weekend :) to busy today with other stuff!
Well what do you say about mine videos using this method
Where can we access this google collab codes?
Hello! I have tutorials and step-by-step videos on this on my Membership page if you are interested :) ua-cam.com/users/AllAboutAIjoin
cool things, but where the copy colab link to play with?
It will be for my members now first :) ua-cam.com/users/AllAboutAIjoin
@@AllAboutAI there any page or video with explanation what i will get with be a member? i did not found anything.
can you send link for python script ?
Sorry for the comment off topic... WinAmp, You're Cool Man!😂
For someone who doesn’t know code how would I build a similar application?
Hello! I have beginner friendly tutorials and step-by-step videos on this on my Membership page if you are interested :) ua-cam.com/users/AllAboutAIjoin
@@AllAboutAI Thank You. I need that badly
Hi, can you add in your API that everytime a new character talks (in a dialogue) it calls a new voice? ;)
I think that can be dont with Python yes! Splitting the text outputs
@@AllAboutAI
# Read input text from file
with open('input.txt', 'r') as f:
input_text = f.read()
# Split input text by character name
dialogue = input_text.split('
')
dialogue = [d.strip() for d in dialogue if d.strip()]
Does it sounds about right?
How about a video that shows how to use AI Voice to table read a screenplay, with different characters. You could also add images (from Mid-Journey) to run with the dialogue, so that it runs as a full on film production.
Cool idea:) noted:) and thnx for tuning in!
Hello my Winamp brother.
Whooot? Someone is still using Winamp? I thought that software was old, out dated and dead long time ago. Amazin! 😆
Is there a character limit in ChatGPT Api ?
Yes, 4K :S
Can you use this app to turn a book into audible?
You can of course, but i think I would check the price for that first with Eleven Labs:)
why is there no real time audio voice?
I just wanna shout out still using winamp I've not seen anyone else use that in years.
haha, it reminds me of this AI era, so i wanted to use it :)
You can tell this guy has never been on VRChat or he would get more accurate feedback on how people actually perceive "AI" voices 🤣
This will be amazing for people with learning disabilities, dyslexia or hearing aids and are in need of assisted reads during tests and exams etc though ☺
So hey mate, how can I do all this? Your video shows off what can be done and how can i now do it? I already subscribed to your newsletter. Do I have to pay to join? tell me sir.
Helloe mate :) it is a members feature! Send me a mail if you wanna know more :)
Ok so i guess i need to just join. Thanks
@@AllAboutAI Which one of you memeber should i join to et access to this info for this video?
I'd love to see a tutorial using the 11labs api
It is up on my membership now
@@AllAboutAI I am also very curios here. Being new to colab as well. What membership do you refer to? The UA-cam?
its 117 pages because movie scripts are formatted in a such a way that 1 page represents 1 minute of the movie.
Will ai create phone calls,animated,cgi from movies like they did in fat albert 2004 and green screen to put yourself in movies and shows and meet young people from old movies and shows like they want flashback back like 50s 60s 70s 80s 90s 00s 10s 20s and stuff like that? I want ai to bring it back to hear old voices from elevenlabs.
can i have python notebook code ???
What about the cost for the use of the API's ?
Its well documented on their website. OpenAI is very cheap, Eleven Labs is a bit more exp
@@AllAboutAI i thought it would be a nice addition to at least include something like a screenshot of the piercing in the video but that's just a suggestion. I love your content keep up the great work ❤️
How to use eleven labs api please make a video
*machina* is pronounced like _mechanism_ , not like _machine_ . It is stressed on the middle syllable: muh-KAI-nu.
See if the voice generator gets it right if it sees the entire idiom: "deus ex machina".
How much Python scripting is necessary for a non-programing background person aged over 40 to learn advanced level prompt engineering. Basic,Intermediate and Advance?
I've followed the links why do they lead no where?
great
thnx for tuning in :)
Would the Rookie membership get me access to the code used in this video?
Hello:) The script is for Tier 3 members, but all tutorials is for Tier 2. T1 is more of a support tier:)
The best voice to use for ChatGPT would be that of Optimus Prime.
(and it got the joke with a little prodding)
Share the Google colab please!
Thanks for the great video btw I'm wondering, is eleven labs voice API actually free?
It has a free version with 10K characters yes :)
Hello can you share this google Colab? Thanks
Hello! I have tutorials and step-by-step videos on this on my Membership page if you are interested :) ua-cam.com/users/AllAboutAIjoin
LMAO, did anyone else feel weird when he used Elsa and Spiderman in his children's story since those were the main characters in the UA-cam controversy ELSA GATE?
Great, but where is the code?
This video is awesome but wait... WINAMP still exist? 😍
Haha i found it so yes 🤩
@@AllAboutAI Dude I just download it haha 🥳
Got all excited and the web site doesn't work, won't let me create an account. Hey AI, how can I improve this website? LOL