Coqui TTS/ Local Machine TTS

kwannie

I am curious if anyone has experimented with Coqui TTS an open source voice cloning platform (for local computer) to add speech to any DAZ animations. I have been searching around to find more modern TTS engines that can be run on a local computer. Couqui seems to have  a very technical setup process that relies heavily on Python, which I would like to learn nut am an absolute novice about. Any info on Coqui or TTS that can be run on the local machine (no internet dependancy) would be appreciated, Thanks.

kwannie

So which forum would be the best place to get some interaction on TTS (Text To Speech) topics?   Hoping there are a few of you that do some animation  experimenting here. I got a bit of AI suggestions in the past, but I am really looking for a few Local Machine TTS possibilities.............Anyone???   Thanks!!!    (pnsc)

pctech4ny

I think there are some animation threads in the "Art Studio" part of the forum.

NylonGirl

It seems like all of the AI stuff depends on Python. It's possible to download and install Python on a computer. Then you have to start it from a command line, at least on Windows. I've been trying to figure out how to get the Python scripts to run through a web interface. And as far as I know, that involves downloading and installing Apache. I figured that would be a good way to go since all of the AI language models and image generators seems to have web interfaces. But I guess it's possible to run them from the python command line if you get python going from the operating system's command line. None of it seems easy and it seems like an RTX 4090 is a minimum system requirement.

kwannie

   Thank you  for the feedback!! Well, I found a couple of resources that explain how to install Coqui TTS, but because it requires so many dependencies such as installing python and a "CUDA Toolkit" I am very cautious to proceed. I have never tried to tackle this type of interface before. Accordin to the YT video in the link below, Coqui can run local on a Windows machine and has the ability to do voice cloning. I am not so much interested in cloning as I am natural sounding TTS with a bit of expressiveness. If anyone could ad a bit of expertise to the links I provided I would greatly appreciate it.

https://stackoverflow.com/questions/66726331/how-can-i-run-mozilla-tts-coqui-tts-training-with-cuda-on-a-windows-system

 

 

N-RArts

I got a UI called OOGABOOGA to run on my ancient (10 y/o) AMD APU. It was slow as heck, but it did work. 

I don't know if the link will stay up, but here's a video about it, and how to install it:

NOTE: IT'S NOT NSFW (AS IN PR0N), IT JUST MEANS THAT YOU'RE ABLE TO DO WHAT YOU WANT TO DO WITH THE AI. THERE'S NO NUDITY FEATURED IN THE VIDEO. I WOULDN'T HAVE POSTED/SHARED IT IF THERE WAS.

You might want to also check out something called PINOKIO. It's a hub where you can access AI tools. I also had this installed on the same computer as OOGABOOGA, but I needed the space for other stuff.

I hope that you're eventually able to do what you want to do. 

kwannie

N-RArts I appreciate the information and I did check out the video and exploe the other AI tools from the YT link. It looks amazing, but as always my concern what data is collected the anonymous  overseeing corporate interests, if any at all.

I have received snarky comments in the past for voicing my concerns, but data harvesting is a thing, they don't give out free stuff without getting some kind of return on their investment.

Are there any registration requirements to use OOGABOOGA or PINOKIO? Do they require an email to download or a persistant internet connection to run? My concerns are sheerly based on my ignorence about what has or does not have nefarious ramifications. I mean if it comes down to it, I use DIM and Iclone Hub to download assets all the time so paranoia doesn't completely interfere with progress, just a bit catious here.

From what I can tell the Coqui installation is all done on a local machine once the required programs are downloaded.