Thread by @stefan_3d

Testing @intel #OpenImageDenoise inside #b3d. Not fast enough for real-time on my ancient Ivy Bridge machine.

Quick and dirty hack into the drawing code, not taking advantage of albedo and normal buffers yet.

Turned it into a compositing node. Extra inputs are still missing. This should be useful for denoising animations or using it with third party render engines.

Classroom scene, animated at 64spp. Admittedly, 64spp is extremely low and I can't expect a static image denoiser to handle that without artifacts.

I hope OpenImageDenoise gets temporal denoising in the future.

Of course, Twitter had to re-encode that clip, so it's hard to tell codec artifacts from denoising artifacts. Be assured though, the uncompressed footage looks like you've had 100 cups of coffee.

Well that was easy. Normal and Albedo passes are active, HDR and sRGB options exposed. I believe that's feature complete.

Now for the ugly part - satisfying CMake and updating install_deps.sh...

Out of curiosity, this is the @nvidia #optix denoiser in the #b3dviewport. This is work I can't legally share, as the OptiX SDK is incompatible with the GPL.

Split screen #b3d:
@intel #OpenImageDenoise vs no denoising, 32spp