Can Jogamp's Ardor3D various Particle effect be used to render something like a 3D probability distribution cloud, or fog and
haze with a concept of light extinction coefficient? Or like a light scattering effect in a field of many small particles?
The fog is doable too. There is another example with numerous lights.
I just warn you that I won't accept any request for enhancement for the version 1.0, I'm extremely late and I still have to reproduce at least 2 important bugs you reported. I know dynamic light scattering, is it what you want to achieve?
Understood. Dynamic light scattering yes would be great: we could render clouds based on droplet size distribution. But that might be more than I need for this particular aspect. I will make a more precise description.
I'm looking for a way to interactively visualize a scalar value with 3D domain, f = f(x,y,z) so that, for example, higher field values are brighter or less transparent, but with the field value -> (color, alpha) user controllable to focus on other features like gradients. One use case would be like a 3D electron probability density surrounding a atom, where the electron cloud would be semi-transparent with a opaque solid object at the center. We have method which I described in a previous post, "render order of children" what we call volume rendering:
but it has some limitations: the underlaying data must be very high resolution, or at least be able to be upscaled, or the rendering looks very blocky. I'm convinced though, if I can resolve the child draw order issue, the texture update capability of Jogamp Ardor3D will be a huge boost to the interactivity.
I've been wondering, and experimenting a little, if the ParticleSystem and Sprites might be used especially in the use case of 3D probability distributions. Is each point like a separate light emitter? The more points per unit volume would make that volume brighter to the eye, and density of points could represent the higher values of the field (probability). Or could there be a single point per unit volume independently mapped to color and alpha so that lower field values would be become more transparent?
Is the smoke effect (NewDynamicSmoker) a "many lights" model, like what the ParticleSystem and PointSprites example demo? My description above (individual light emitters) is correct, right?
You had mentioned 'dynamic light scattering' w.r.t to fog. Do you have a bit of time to discuss? Is this particle based (like proteins in a solution, or water droplets suspended in the air), or some property is specified per unit volume which effect the light?