Any news about supporting atmosblender? Or at least supporting multiple krakatoa atmospherics in a scene? I think this would be a nice way of setting different Length’s steps etc.
Thanks,
Ben.
What sort of things would this enable you to do? I’m thinking the most useful thing would be get different voxel spacings for some particle objects. If that was the only benefit, it might be simpler to allow the spacing to be set on a per-object basis in the atmospheric. Having multiple atmospherics try and communicate with each other seems destined to cause headaches.
All the settings in the atmosphereic could be different. And it should also work with fumeFX, Hairfx, afterburn, v-ray. I’m not sure what the state of that project is. It may be defunct. We should contact Chaos group. I think the ablity to render with other atmos systems is more important than supporting multiple krak atmos’s.
But while we’re talking about improving the krak atmos so we only need one instance, I’d like per object settings, for length, steps, and render style (shlick, iso, phong, other). And also an interface to the volume data object(s) so it can be processed before the render, and exported and/or cached to a file.
Thanks
B.
Considering those other atmospherics’ objects can be converted to render in Krakatoa, it actually might make for a better use case as you say to handle multiple Krakatoa atmospherics. But I suppose supporting AtmosBlender would allow you to render Krakatoa plus an unsupported (by Krakatoa) atmospheric.
If we were to have per-object settings, wouldn’t you likely need to do it as multiple atmospherics (with a include/exclude list for each one)? Like if the voxel size or step size or whatever was per object, what would you do when the objects intersected? They’d have to be two separate voxel arrays anyway, right? And you’d end up having to step through the same space twice and blend the results together, right?
If we had multiple objects with different spacing values, I’d have to loop through each of them each time I went to sample along the ray. That is somewhat easier than trying to compute an entire ray and then combine the results afterwards.
While we’re on the topic, i’d like some diagnostics about the memory and total dimensions of the voxel grid.