Indigo Renderer Forum

Texturing to glass surface is hard to be unbiased
But i think textured medium may be easier
because it is just a subdivision of volume and assignment of extinction color to voxels.
Volume texture can be sequentially-numbered images or all images in assigned directory.

vidro(a biased renderer) has this feature.
Volume(is bounding box of medium in vidro) is divided evenly in pixel num(X,Y axis) and image num(Z axis)
extinction color and anisotropy of scattering are assigned to each voxel.

Hello Miso and welcome !

In my opinion parametric volume textures are much simpler to manage, but one would hardly have a control on their placement. That technique you are describing is potentially heavy to generate/manage but offers a good control on texture placement.

Points clouds are a third, intermediate, technique: somehow heavy but more scalable than voxels, interpolation is nicer. Plus a very few 3d apps can generate slices of a volume while most have a particle system.

Finally for the sake of procedural textures, I'll add that they can (oops edit: could) be generated by Indigo itself.

Hi Miso,
interesting idea.
I actually implemented grid-based non-uniform density participating media for smoke datasets etc, so it would be relatively easy to make the grid values control the extinction coefficients instead.

I'm not sure if it would help for stuff like a human face tho, where you need reasonably high resolution for controlling the absorption coefficient.

you finally said something about the particles, again

Thanks for your replying, Men

Points clouds are a third, intermediate, technique

yes, it is appropriate method for smoke and fog, and more flexible than voxels
indigo can render coffee in a cup
and we can add steam on it with this feature.
Procedural texture is very easy to generate and useful for some cases such as clouds, but not for arbitrary cases.
Voxels and points clouds with textures are opposite.

I actually implemented grid-based non-uniform density participating media for smoke datasets etc

That's interesting.
High resolution volume texture is a memory eater and rendering may be very slow.

you finally said something about the particles, again

You could say that .
My suggestion is how to visualize results of parｔicle/grid base simulation
diffusion of gas or convection of coffee and milk.