Hi, I’m very new to TouchDesigner. I’m looking to create a network where on one end I input audio, and on the other a 3D object is generated. It could be either that the object is generated from scratch or the object is the result of modifying a subdivided 3D primitive.
I’m envisaging creating vertices using numeric data (which could come from the processing of the audio signal) then connecting them to create the mesh of a 3D object. Is there any way of realising this?
If not, what would be a workable approach to generate 3D mesh from audio?
Thanks in advance!
A 3d object is more than just vertices. You’ll also need to specify how those vertices are connected to each other (edges), how they form faces (triangles or quads) and how they are orientated (normals). So I’m not sure when just having audio signal data how you would want to visualize that as a mesh. Or perhaps you’re aiming to render audio visualizing lines that deform based on the frequency or amplitude (no faces/normals, just edges between points)?
In any case it might be a better solution to start with a 3d mesh and deform the mesh based on the signal. For example using FFT (audiospectrumCHOP) to receive the frequencies and use that
to deform a gridSOP using noise. (either noiseSOP (CPU), or using a displace texture in the PBR material (GPU)).
Or you could for example instance boxes at different positions to build up some 3d shapes (though this will have hard edges since the boxes will intersect with each other)
Hope this helps!
Thank you Tim!
I think the idea of visualising lines sounds good. Would this give me a wireframe 3D shape that I could view from all angles in TouchDesigner (and maybe model according to in other softwares)? If possible may I be given some further instructions on how to do this, or be pointed to resources that can guide me to do this?
I will try the other 2 approaches you’ve mentioned too. Thank you so much!
It’s quite a big world out there what is possible in TouchDesigner, so there is not really one straight-forward path to master everything. Which is wonderful, since it’s an endless adventurous rabbit hole
Perhaps check out alltd.org (td tutorial website) for some tutorials on for example audio analysis/visualizations.
I made a simple example to get you started. It’s a tube that’s reacting to the FFT analysis of audio. If you view the AudioVisualizeCOMP, you can control the camera and look around.
AudioVisualize.tox (28.8 KB)