Blender HAIR for Lux

[shadypepe]

I think the integration of exporting hair natively to lux is one of the most important things to make blender/lux an alternative to commercial 3d suites. So is this project still alive? Or are there alternatives?

[SATtva]

It's on the list of priority tasks. Hope to finish it before 0.7 release.

[shadypepe]

hey that's great news, thanks!

[SATtva]

Hair export is in the repo. Implementation uses instanced primitive shapes and is loosely based on e_jackson's work. Added a new section in the LuxBlend's materials tab just below mesh properties to define hair thickness on a per-material basis. You can get the dev version of LuxBlend from here.

With large number and high complexity of hairs be prepared for very long export times, rendering speed can also drop significantly. You've been warned.

There is still some problem with motion blur -- these transform matrices drove me crazy. I didn't had time yet to look into the issue, but will in a couple of days. If you spot the culprit in the meantime, feel free to provide a patch.

hair-test1.jpg (91.41 KiB) Viewed 1315 times

[dougal2]

Crazy and Awesome !

[SATtva]

I'm exporting a testscene with 4000 hair particles (8 segments / 17 parts each) for two hours now on my laptop.

[Lord Crc]

Uhm yeah that's speedy lol.

[SATtva]

And it's still exporting. :D About 3300 particles finished. I really need to look into code optimization in the exportObjects() method as the current state is simply awful. :(

[Dade]

And it's still exporting. About 3300 particles finished. I really need to look into code optimization in the exportObjects() method as the current state is simply awful.

For what I remember it does some O(n^2) check to optimize the mesh, you could try to disable optimization to check if it is the main source of long export time.

[SATtva]

Dade, there is no mesh export here, only spheres and cylinders (just a pair of them per hair material used). However in the long run we have to move to converting hairs to mesh on the fly. First, for GPGPU kernels; second, with the current approach performance drops linearly with the number of hair objects (on QBVH) what is really bad.

Doug already made some optimizations, but not enough I think. The main problem is not in hair export function (this just collects all the needed primitive objects), but in the main object export code which simply chokes on those tenth of thousands of objects.