Home
Artists
Search Recent Random
Posts
Search DMs Popular Hash Lookup Tags Random
Importer
Import FAQ
Register
Partychan Telegram Find&Fuck Waifus ThePornDude
Home Artists Posts Import Register

PLUGIN UPDATE: Alive v70 (files) (Patreon)

Published:
2024-02-21 14:46:06
Imported:
Tags:

Downloads

Content

INFO

A few people reported yesterday issues with ElevenLabs Text-to-speech in Alive. I tested it and I had problems as well. It's something related to their SSL certificate I think, something not working properly with Vam's unity version being a little out of date. It's something I had trouble getting to work in the first place. It also retroactively stopped working on older versions that used to work before. To fix the issue and to prevent other similar issues in the future, I moved the elevenlabs api calls directly in the TextAudioTool script. Until now they were done directly from inside VAM.

For people that use TextAudioTool, I recommend updating to the latest version 0.3 (first zip here https://github.com/5PQR/TextAudioTool/releases/latest ) . You only really need to replace this file with the newest one: /SPQR.TextAudioTool/SPQR.TextAudioTool.script.py . Starting with this Alive version, v70, the elevenlabs TTS needs running TextAudioTool along with vam, same as the offline voices and the speech-to-text functionality.

CHANGELOG


CREDITS: https://www.patreon.com/posts/61736572

INSTRUCTIONS


AI



DEPENDENCIES

ALIVE should work without these vars too but they're recommended for the best experience and more morphs & animations. They're all free VAM addons.


Files

Comments

Dakoto

Hey, for the vision part, it ll be cool to add the parameter to the max turn up, turn down, and side head like on your earlier version. Do you think it ll be possible to re-add this ?

SPQRAeternum
>>130362220

I'll think about it and ways to make it better but no promises on sliders. Some of that is driven by the character now, it's dynamic and somewhat context-based. Sometime the character rotates more, sometimes less, which to me is more realistic, or at least gives more variety. I don't want it to be a robotic automaton that becomes predictable and constant once you get used to its movement

Anonymous

Is it possible to use a local TTS such as xtts v2/allstars with the plugin. Or is it possible to somehow wrap their API to trick your plugin into thinking it's calling and sending request to elevenlabs?

SPQRAeternum
>>130445238

there's a modified TextAudioTool that a dude did here: github.com/marcoseduardopm/TextAudioTool_Coqui-ai_TTS that should allow coqui voices , it should replace the local tts voices with that Generally speaking yes, the way it's designed, people can easily change what alive connects to if they want to code. There's people here that connected it to chatgpt too. For STT and TTS, it's possible to change the textaudiotool script which is a very simple script doing very basic stuff, and use different services instead of whisper for STT or windows voices for TTS. The elevenlabs calls have more calls (to get the AI models, to get the available tokens, etc). For replacing TTS stuff, changing the offline voices is the way to go, there are just two calls there, one to return a list of available voices, and one to generate the audio and return the file name to vam. INB4 why i don't add coqui voices or other offline AI voices: licenses troubles, there's lots of headaches with this AI stuff licenses

Anonymous

How do people like to use Alive? It's my favorite plugin I'm never sure what to do with.

Anonymous
>>130445238

Thank you for answering, this is very useful. I guess it's true so many AI programs are free to use but don't let you use it on any platform that could let you generate income. Ill check out that fork thanks

SPQRAeternum
>>130555827

It's more of a question of how you use vam I think. I initially made this plugin as a quick alternative to ExpressionRandomizer++ for photography purposes. If you wanted to make a MetArt-like photoset with VAM, a few years back I think all solutions required for you to basically do softcore game developement to trigger randomized expressions, or manually go through menus and alter yourself the morphs of the character from photo to photo so that the photos won't all look the same and lifeless as is often the case with 3dcgi expressions. I liked ER++ but it too required manually doing stuff every time, picking morphs. So the main purpose of Alive is to add a bit of life to the characters, with 0 work or DIY setup required. What you do with that, it's up to you, and people use it in various ways: - interaction with vam models - using the body actions to get a very quick sexy pose set up - 3dcgi photography - photorealistic AI photography and art. You can take photos in VR of almost anything you can imagine, something that a few years ago would have seemed straight out of science fiction. It blows my mind still how under the radar this feature went around here - to do quick videos without having to animate expressions yourself or with some of the animations included, like some of the videos I do for fun as trailers sometimes - AI companions or assistants and AI scenarios, play out your fantasies - For advanced DIY enthusiasts there are lots of triggers, voice commands and things to play with - now you can do makeup and tattoos with the aesthetics editor so it's a tool to customize VAM models in ways that weren't possible before, or to build your own textures, draw on models, etc - i use it all the time now for hair customization, changing the settings on all hair items at the same time, also as a tool to create random looks or turn down all morphs on a model without having to merge them in a single morph - some use it for quick sound background. without it again you have to do softcore game developement just to add some atmosphere sound, every time in every scene. With Alive you have background noises and music playlists too

Michael Fread

Is there any way to set the voice to text to be voice activated like that supported on the windows speech recognition? Since after pounding my head against the wall for awhile I realized that had to be enabled I assume it is being used under the hood. We could then hide the UI and converse with the models. That allows for much more natural and immersive interactions with the models, especially in VR without a keyboard.

Michael Fread
>>130555827

That history explains a lot. The plugin pushes away the uncanny valley and almost serves as a sort of essentials pack and with the TTS the models really do seem to come alive. I wondered at the effort put into stills which are essentially outside the game and bogged down even my 4090 running stable diffusion along with the rest in VR. Now I understand. With voice activated speech recognition and a few extensions so the AI can trigger poses and mocaps/animations plus something like toyserialcontroller loaded and passthrough this thing could enable a full blown virtual sex bot.

SPQRAeternum
>>130799584

Hi, speech recognition in Alive is push-to-talk. In VR you should be able to use the controller to enable speech recognition, you can bring a controller next to the mouth like a microphone and press trigger, it should enable it when the Alive UI is in chat mode. It is possible but I'll have to do some polls to see if interest in AI has grown and if I can put more time into it. So far AI stuff has been bad for this project and it's very low priority for me to work on it

SPQRAeternum
>>130555827

You can for example trigger speech recognition by putting that alive event in a lookat trigger on the character, or on some place in the scene. You look at something and you talk. You can also trigger anything in vam with "trigger buttonname". If there's a button with the name "button name" or "buttonname" or "Button Name" etc in the scene, Alive will (should, haven't tested it in months) trigger it. There's a video on my yt that demoes it. So you can trigger anything triggerable in vam already. There's also custom commands that can be created and there's some that I added already like changing a few poses. Like "turn around" for example

Michael Fread
>>130555827

I'll have to look at those features. Can I suggest tossing up a wiki for setup and howto type information? You are obviously very active on development and Alive has so many features. I think you've got a lot of information on setup and usage but as with most folks using patreon its scattered around on different posts or the notes under different releases. With a wiki you could just put it in a central place and link it on every release. Unless you already have something like that and I'm just blind lol

Michael Fread
>>130799584

I can see that. Those features involve a lot of moving parts and I'd bet most people don't know how to set them up and use them properly. Now that the projects involved have swept their setups into batch files for windows I wonder if it couldn't be all be simplified with a master batch file that grabs what you need and sets it all up within a folder inside VAM and grabs a lower memory footprint model and swap in a replacement text with the right start flags so people can just run either CPU bat or NVIDIAGPU bat and it loads the AI in the background along with your TTS tool. Of course they still need an elevanlabs account if they don't want it to sound like a robot. It really is great functionality. I mean tying to your own mentioned use case one could load a model up in an Alive scene and speak more or less naturally to the model to direct her to pose/dress in different ways with custom commands/buttons and adopt different expressions in the same way. Even engage in a bit of chitchat and get responses from the AI roleplaying as a model in the shoot.

SPQRAeternum
>>130555827

there is a guide posted in every alive release: https://www.patreon.com/posts/guide-alive-87116262 also, in the ui you have information about each setting if you press on the "?" next to them. some apps have extra tabs with help sections " but as with most folks using patreon its scattered around on different posts or the notes under different releases" yeah, i don't know why patreon creators post patreon posts still. they should learn from meshedvr

latter_ganache

Hey! I keep getting an error when it tries to call for alive63.audioasset - im assuming it's a me problem, but do you have any idea on how to resolve the issue?

SPQRAeternum
>>131095306

Hi! it's an old bug in alive that I couldn't fix. It halppens when you load different versions of Alive in the same session, the sound will work only for the first version. If you load v69 it will work. If after that you load v68 or v70, you'll get that error. The fix at the moment is to hit hard reset to restart vam and then load directly the version you want to use. I'll take another look at fixing it, maybe now I find something to get rid of that

latter_ganache
>>131095306

Awesome! Thanks - i kept forgetting to set the defaults but once i did, and reset it - no problems! Had you not said it was a bug, I wouldn't have thought anything about it and just assumed it was Vam weirdness.