Generate AI Images with Stable Diffusion + Audio Reactive Particle Effects - TouchDesigner Tutorial
Added 2022-12-20 06:48:54 +0000 UTCHey, this is an early release of my tutorial and project file on how to use Stable Diffusion in TouchDesigner to turn AI-generated images into a video and add audio-reactive particles for a blending effect.
Make sure to watch part 1 of the tutorial: https://www.youtube.com/watch?v=mRXTR9vcHAs
Project file included below ð
Comments
having the same issue, did you find what causes this?
Beatrice
2025-01-25 23:57:55 +0000 UTCi try to follow the tutorial 100% , but unfortunately the outcome is totally difference than what's in the tutorial ,there are way too many black particles than the original tutorial ,may i know why?
Chee Ming Wong
2024-05-25 06:01:50 +0000 UTCOk. I've gotten most of this going. Quick follow up. Best way to programaticaly pulse the Computerender (Seems like it needs a new Seed # value to produce a new image). I have a button going to the Pulse on computerender so I can Pulse it manually. Can I replace that with ChopExecute Dat? Thank you!
JJ Wiesler
2023-10-18 00:30:20 +0000 UTCThanks so much!
JJ Wiesler
2023-10-17 02:34:41 +0000 UTCHey! The particle system will work if you pass any TOP into it, so it's already setup for real-time. Try connecting the output of the ComputerRender into it. For audio reactivity it'd look nice to pass the channel Low, Mid or High onto the particles_gpu_2022 component for it's Speed parameter in the Settings Tab. You can always place a Lag CHOP in-between. If you want to configure the Computerender component to continuously generate Note that there's a channel called 'loading' that will tell you when Computerender is finished downloading / generating an image. You can create a CHOPExecute DAT to listen to these changes. When it's done loading you can programmatically pulse the Computerender component to generate a new image
Torin Blankensmith
2023-10-14 23:26:21 +0000 UTCI have the same question! Getting the audio is, as described above works well, but how can the image generation and particle system be evolved to work in realtime?
JJ Wiesler
2023-10-14 22:15:02 +0000 UTCHi Torin! Great work - I love the potential here. I'm trying to find a way to make the full project live. Along with live audio in, any thoughts on how to generate the images live so that prompts can be added during the performance or extend for time not limited by the recording? I understand there is the delay as a new image is created but perhaps there is a method for rolling/queueing generation? Would love your thoughts. Thanks!
Deakin Taggart
2023-10-12 15:53:05 +0000 UTCThanks!!
Carlo C
2023-09-06 18:05:42 +0000 UTCHey if you're looking to get Audio from your Serato DJ into your computer you'll need an audio interface connected to your computer. These Scarlett devices work well: https://us.focusrite.com/products/scarlett-solo-refurbished?setCurrencyId=2&gclid=Cj0KCQjwxuCnBhDLARIsAB-cq1rxahYcKQe9t8Yv_H-GJOpeq15YsxmQBRkt2waUSDARNLuTACq3uYQaAh5jEALw_wcB You can then put down an AudioDeviceInCHOP in TouchDesigner and use the audio reactivity from that
Torin Blankensmith
2023-09-06 16:34:10 +0000 UTCHi Torin! I'm new to this community, amazing work. Quick question, is it possible to connect my serato dj audio input in live performances?
Carlo C
2023-09-06 06:08:21 +0000 UTC
