Often I get into the situation where itâ€™s more efficient to combine multiple displacement maps at render time rather than sculpting/applying all of them in zbrush. A common example is the following: I want to create some kind of fabric. I would sculpt the displacement map in Zbrush but leave out the high frequency details. Just big shapes. I prefer to paint theÂ high frequency stuffÂ so I can use the same info in my color map. It also allows me to go much higher res â€“ Iâ€™m not bound to my 8 â€“ 12 million polygons in zbrush. (.. Or just want to combine two tiled mapsÂ in the shading tree).
So we end up with a 32bit displacement map in the [inf, inf] range and an 8 bit map in the [0, 1] range image. How do we combine these? First you have to remap the painted displacement map so the mid point falls at 0. Every render engine provides some kind of control node for this. alRemapFloat for Arnold, pxrRemap for prman, etc. Change the output min and output max to something like -0.25 and 0.25. These values obviously depend on what youâ€™re doing. Make sure to disable clamping. We definitely donâ€™t want to lose half our information in the negative values.
Then just layer it on top using an alCombineFloat node (or something similar if youâ€™re not using Arnold) using the add operation.