krohmerNV

the USD Preview Surface spec says the outputs of the texture reader should be: ``` r - float g - float b - float a - float rgb - float3...

changing the type of `rgb` to from `color3` to `vector3` makes the MDL code gen produce correct results. Giving the USD spec, I would consider this to be a good...

I'll investigate, thanks!

thanks @jstone-lucasfilm, I need to make time for this in the next week

@jstone-lucasfilm updated the default parameters of the script to behave like before. Sorry it took so long.

MDL is broken in this state! I'll prepare a fix and if you are okay, I'll try to push it your branch. One question though. It seems you copied all...

[MaterialX-mdl-fractal2d-node.patch](https://github.com/user-attachments/files/18559513/MaterialX-mdl-fractal2d-node.patch) given the time zone difference and the scope of the patch, you can apply it easily yourself. If you agree with the proposal above to use the same function...

I only looked into the MDL implementation of the fractal3D and fractal2D functions and the only difference I saw was a +17 floating point add. What I was suggesting is...

Saw the renaming. Thanks. This looks more consistent now. About unifying the 2D and 3D code I was asking for. It's not a blocker. If you don't want to align...

tracked down the [GLSL implementations](https://github.com/AcademySoftwareFoundation/MaterialX/blob/6032c4ce1b3c32ff782a3fb51d3a0a9c5a17beb5/libraries/stdlib/genglsl/lib/mx_noise.glsl#L249) and you are right, there is a difference. 2D is using bilinear interpolation, 3D trilinear. I also checked the MDL implementation and the PR changes...