Using a redshift shader to control Shader Effector in mograph
I'm using a Matrix object to generate particles. The particles are spread aross the surface.
I'd like to generate more particles where there's more detail.
I figured I'd use a shader effector to try to control the density of particles to specific parts of the geometry. I can't seem to do that, but I can contol particle color and scale.
This works fine with a checkerboard shader or a gradient in the color channel on a classic C4D material as a test. But it won't work with something like an AO shader.
What I'm really looking for would ideally work with the curvature shader as found in Redshift materials (which in effect highlights sharp details).
It seems that the Shader Effector doesn't work with Redshift materials at all.
So I'm interested in other solutions to either control the density of the particles in the Matrix object based on the amount of detail (so flat surfaces don't get a lof of particles but corners and details do), or at least control their color/transparency/scale.
The Shader Effector was developed for the Standard Material. To the best of my knowledge, and without mentioning in the Documentation: Bake what you need and go from there. (This can generate a lot if that is an animation, I'm aware of that.)
Redshift calculates the materials when the render process has been started. By then, all MoGraph parts have to be calculated. Baking is an option to have that data available before MoGraph starts its process.
All the best
Thanks, I suspected as much. Some of the geometry is uhm... elaborate. So I'm not sure baking the curvature may be feasible for part of the project. I might try a different route by overlaying the curvature pass in compositing...
Thanks for the reply, Bar3nd.
Since I have no idea about your project, I have to guess here, and that is my least favorite way of working. Typically one works on the file or project that is in question. Guessing here is (hopefully not) wasting your time.
How about working with Vertex Maps, eventually selecting edges, converting those to Splines, and using the splines as fields? In other words, paint or define your areas of interest.
Those edges could be selected based on Phong Break Angle.
My best wishes for your project.
I completely understand.
The thing is that in the end it will be a lot of very complex scenery with fairly complex geometry (think trees, cathedrals). So I'm currently working on developing a pipeline that allows me to bring in the models/geometry, and develop the final look in a fairly procedural way. And ideally in a single multipass render. This is why I'm hoping to avoid baking and vertex maps.
But I've figured out how to keep the base geometry invisible while still sending the curvature pass to a custom AOV - which allows me to use the curvature output (which works better for my purpose than the AO pass) to tweak the look in the composite. Not in particle distribution per se, but in luminance. Which seems like a workable compromise.
Here's an impression of the direction I'm heading in.
Thank you, Bar3nd,
It sounds like you found a workable solution. If the per frame generating of data isn't slowing down the process, then that is certainly great.
My best wishes for your project.