Hi,
I am using POP operators to divide a 3D object along the Z-axis and map the rendered result onto another grid-based 3D model. However, the Group POP is very performance-heavy and introduces noticeable latency.
I would be very grateful for any recommendations on a more efficient workflow or alternative approach.
2DsliceTest.4.toe (760.8 KB)
Hello @marugotobanana0120,
The group POP shouldn’t be particularly heavy, on my system I see 0.04ms for cpu and 0.02ms for GPU.
You do have a lot of group POP + LookupTexture POP.
I think one thing you could try to make the system a lot faster is to offset the texture coordinates to match layout1 and use a single lookup texture POP, that would also avoid all the crop TOPs.
Let me know if that makes sense.
Best,
Vincent
Thank you for your help.
I’m not sure how to configure the Lookup Texture POP and the Texture Map settings.
Could you please explain the correct way to set them up?
Hello @marugotobanana0120,
Something like that, I think the calculation may be a little off since I’m not quite getting the same result, I would start with a simpler test case.
Since all the slices are packed in X in the layout TOP, the x tex coordinate is remapped from 0 to 1/72 (you have 72 slices), and an offset is added based on the z coordinate, so each slice in the POP looks up the correct part of the large texture.
2DsliceTest_singleLookup.toe (19.7 KB)
Thank you very much for the detailed explanation.
I understand the concept of remapping the X texture coordinate from 0 to 1/72 and offsetting it based on the Z coordinate so that each slice reads the correct region of the texture.
However, I’m still not able to reproduce the expected result on my side. I will try creating a simpler test case as you suggested and check the calculation again.
If you have any additional suggestions or notice something I might be doing wrong, I would greatly appreciate your advice.
Thanks again for your help.