BEST FREE TTS AI Voice Generator | Any Voice, RVC Model + More!

what's up I'm troubleshoot welcome back to another video in this quick one I'll be showing you how you can use any RBC voice model to create text to speech why is this so special and why does it deserve a separate video well essentially you can train voice models on anyone and you're able to transfer your voice or someone else's into anothers I've previously covered videos on not only using your own voice but also creating those AI song covers where you replace someone singing in a song all the way to changing your voice live on Discord Etc you'll find links to these in the description down below this video is going to cover creating text to speeech in someone's Voice using one of these RVC voice models if you followed either of the previous two videos you probably already have some voice models but we'll go through everything in this video here just to keep you all up to speed first of all let's download the text to speech voice model program in the description down below you'll find a link to Appo RVC Fork this allows you to do a whole bunch of things including most importantly text to speech I haven't seen many bits of software offering this other than maybe like 11 Labs where you type in some text and it spits out someone speaking in someone else's voice but this is run completely on your own system for free all we need to do is look on the right side for releases and just below you'll see an installer simply click this to get to the latest version then we'll be downloading the windows. bat file here obviously if you're on Linux you'll be downloading that instead then when it's done in instead of just opening it we're going to be moving it to a folder where we'll install our program so I'll place it on my desktop and call it maybe aits you can call it whatever you want I'll move the bat file into here and all we need to do is run this if you'd like you can right click and choose edit or open a notepad to see exactly what this bit of code does if you're prompted about Windows security simply click run anyway much like the rest of this project is all open source so you can see exactly what this installer file does etc assuming you're able to understand what's going on here to install this program simply run this file and a window will open up where the installation will begin it'll all be downloaded to this folder here the first question is what kind of graphics card do you have I'll be choosing the top one as it's probably the best for most users these other two are only if you really know what you're doing so I'll enter one and press enter now a rather large download will begin and it'll take some time to complete I'll skip forward until this is done there we go the download is done and the window closes now we have a folder in here where the actual program is installed you can move this folder to a different place but for now I'll just be leaving it here opening this inside we'll find a whole bunch of files and what we're looking for is goao dobat simply run this and you'll be able to launch the software the first time you fire it up you'll be asked a question more than likely it'll be number one if you're using Nvidia two for an Intel and 3 for AMD that's it I'll enter one and press enter now a few more downloads will be done including a few models to actually get the text to speech to work and things like that and shortly after the program itself will start up just a quick note if you get any errors about python or something like that in the GitHub page you can scroll down to the installation section and read some more info about fixing possible issues more than likely you won't need to do anything but if you need to download python you'll find a link for it in the description down below as well shortly after a few gigs of downloading models the program itself will open up in your browser just like this this model looks very similar to the previous video that I did on the original voice too RVC program you'll find this link down below essentially we can use the first tab to convert one voice to another the second tab to train our own model the third Tab uvr 5 separates voices from music Etc then TTS is the tab that we're going to be focusing on in this video you also have the resources to download models and import them which we'll get to in a moment the extra tab to merge audios Etc and and settings to change the theme of the program but once again TTS is where we'll be spending most of our time here first of all we'll need to select a voice model for which I currently don't have any let's go ahead and download some in the description down below you'll find a link to voicey models.com but there's quite a few places that you can get voice models from this is just one of them all you'll do is you'll search for someone or something so for example I don't know let's go for Obama and it'll be looking for RVC models that usually have RVC written some where in them or in the download this one looks about right so I'll click the link to download it or be taken to a page where I can download it then after the roundabouts 300 Meg download is complete we can open up the zip and we'll need to extract these to somewhere temporary so for example just extracting them to my downloads folder here as such we can delete the zip and with these two files here let's actually import them so back to the web UI head to the resources tab at the very top and where it says drag your path file here simply drag your path file as such into here and on the right side your index file so index I'll drag here just like that as soon as the uploading vanishes it's now been copied into the actual program itself and we can delete these temporary files here as well on top of the zip otherwise you'll have multiple copies of the same thing everywhere heading back to the DTs tab you should see in RVC models and clicking refresh you should see in the RVC models we now have someone here and in the index file we can select the appropriate one here I probably should have renamed this to be something like Obama to match the path file just so we know which index file is which after selecting both of these we'll need to head across to the left side where it says TTS model and choose a voice type if we expand this you'll see tons of different languages followed by names so for example AF is Africans Za stands for South Africa followed by a name or a style and a gender if we scroll all the way down or simply type in E en hyphen will'll find all of the English accents so it'll filter it we have Australian South African British Etc some of these sound much better than other ones all you're going to do is look for something that matches the gender of the person you're trying to recreate so for example let's go for maybe I don't know Eric mail then we'll type something in here such as this then we'll click convert at the very bottom here this will then convert our text into a text to speech audio file and shortly after run it through the Audio model we have on the far right and pump something out at the end we'll have an audio TTS version that you can listen to and hear how the raw raw unprocessed file sounds hello I am Obama and I am enjoying a hot dog doesn't sound all that good but the processed file on the far right should sound a bit better hello I am Obama and I am enjoying the hot dog much better if you'd like to save it click the three dots and choose download that's it if we have a look at TTS method we can expand this and choose Google TTS if you'd like to use Google's text to speech engine to create maybe a more realistic sounding voice so if we try convert this we should hopefully get something realistic sanding back hello I am Obama and I am enjoying a hot dog okay yeah well uh definitely not but let's see how it came out on the other side hello I am Obama and I am enjoying the hot dog and now you can see why it's important to match or at least closely match the gender of your text to speech and The Voice type of your model as for changing the gender or how the Google TTS sounds you can't do that and of course it's online Edge TTS however should be run off of your system so en let's go for us this time and go for maybe Roger if we try convert once more we should get a different result and this is what it sounds like hello I am Obama and I am enjoying a hot dog then hello I am Obama and I am enjoying a hot dog obviously it could be improved with pulling on an accent or different spacing between words Etc you could edit this to sound a bit better but obviously the most realistic way of getting a good sounding result is to record yourself saying something and run it through on the model inference tab over here so for example we can refresh here choose the voice and the index file then using the original text to speech let's say let's drag it into here to select it we can change the pitch and convert should change our text speech audio into whatever model we have selected here so playing this hello I am Obama and I am enjoying a hard dog obviously it sounds very similar to the text to speech tab as we used a text to speech sample however we can instead record from a microphone so I'll exit this out and record from a microphone so oh hello I am Obama and I am enjoying a hot dog doing a terrible accent let's convert it and oh hello I am Obama and I am enjoyed a hard dog obviously uh marginally better but you get the point it's a great little fun tool to play around with and you can import practically as many text speech models as you want to import more just head to the resources tab X out of both of these here add drag in A New Path file and index file select them on the text speech Tab and just like that things are working as you'd hope that's really it it's a simple piece of software that just works uses your own local resources and is completely 100% free assuming you have the computer hardware to run it you can run this on much lower end systems without such powerful graphics cards to speed up your process but it is going to be a little bit slower in that case we can close out of this command prompt window to exit out of the program and as soon as we've done that you'll notice that some resources are freed up on your system and in order to launch it again simply head back to the folder Appo RVC Fork followed by goao dobat that's it hopefully you found this video useful it's a really fun tool to mess around with and that's really it so thank you all for watching my name is been troubleshoot and I'll see you all next time ciao

Channel: TroubleChute

Views: 4,146

Keywords: ai, tts, ai tts, ai voice generator, ai voice, tts ai, free tts ai, ai tools, ai text to speech, text to speech ai, free ai tts, ai tts tools, eleven labs ai, ai robots, aitts, ai news, ai video, twitch ai tts monster, ai voices, gpt tts, bark ai, bark tts, presidents ai, tts tool, free tts, ai robot, ai agents, ai voice cloning, ai text to speech voice, ai avatar, twitch tts, decoder tts

Id: b4Lke9Z6Hyo

Length: 10min 14sec (614 seconds)

Published: Sat Dec 23 2023