i am getting spikes in my displacement map. this is caused by black and white values in the map that is generated, but thankfully, they’re only at the edges of my uvs. however, they are showing up in the renders, and doing a great disservice to the final result. i checked, and to my knowlege, and i don’t believe mental ray is mipmapping or filtering the image. i’m using custom uv’s (i.e. not AUV or GUV tiles), i have the displacement rendering at 4096, don’t have any holes in my mesh, etc. there is one spot that shows up as an overlap on the uvs, but it’s nowhere near the problem area, and in maya the uvs for that polygon are fine.
if anyone can figure out what’s going wrong, i would appreciate it.
It’s possible that your map is not lining up quite right with the model when it’s wrapped onto it in Maya. Try very, very small offset values in the horizontal and/or vertical directions.
try using a 32 bit map from disp Exporter. I had a similar issue on my Rhino and it disappeared when I moved to 32Bit maps.
S
alright, i’ll give that a try. how do i convert the displacement into a *.map file though? you’ve mentioned it’s a command line program, but i’m a total dummy when it comes to stuff like that. is it something that is done through windows or through maya?
In a pinch you can use the tif file it just tends to be slower and in the worst case causes MR to crash. The best way to convert is to get the .bat file here
and just drag and drop the file onto it. It will convert it and save a copy in the same directory.
Scott
THX to atwooki for posting the utility.
Hello,
What I have tried:
- Every configuration under the DE. As of now, I have written down about 50 combinations. There are always spikes, but some do better than others, still a horrible artifact.
-
I have also tried offsetting my uv (all diff. directions) in Zbrush and then output but I get same effect. Too much offset and my texture goes jigsaw.
-
I tried that image conversion tool listed above but that didn’t seem to change anything…
I am using XSI 5.0 with a subd level of 3-4. UV is AUV. I followed the ZB to XSI pipeline and by now I have it memorized.
It looks as though the spikes are occuring at the edges of my geometry (the openings) - inside of lips, eyes, bottom of bust. Is there a way to fix this?
Could the problem be when -making- the map or -exporting- or -applying-?
I think this is more of an issue of how ZB makes the DMaps rather than application. Oh well, my trash bin gets another ZB project.
first of all, Scott, i never thanked you for your advice. i’m slowly learning my way around a command prompt, and haven’t used the batch file yet. however, i used something similar at one point within Maya called “memory_mapped_textures_scripts” (-.mel), which lets the user convert every texture file within the scene to memory mapped format, which is very useful and helpful. also lets you swap them back and forth with a couple simple commands. i recommend looking it up.
as for the actual topic of this thread: i’ve found a couple things reduce the presence of spikes.
like scott said: 32 bit displacements do tend to have fewer spikes. however, if you have spikes already, especially badly, it usually won’t get rid of them completely.
try to avoid open faces, as that’s where they are generated. if i know i need something open later, i’ve started filling it in the external program, and then simply deleting the polys after generating the maps. if you put them outside the uv range of all the other sets you can use uv sets to hide and/or crease the edges around them.
smooth uv: i use it as a last resort or when i’ve left something open and it’s producing spikes. i’m not too fond of the distortion it produces along the edges of uv seams, but i’ll take it over spikes if i have to.
photoshop: absolute last resort. you can always blur them out, though since you can’t see anything beyond 8-bit, you never know what kind of damage you’re actually doing. but, in in inconspicuous places, it shouldn’t be too bad.
Charkin,
you’re spikes are happening on an edge, but they also seem to be happening where the surfaces curve inward. This may very welll not be it- but what do your normals look like? A fried of mine was having a very similar problem the other day, and his normals on the inner surfaces were facing the wrong way so MR didn’t know where to put the displacement as it switched from one face to another and freaked out. Might wanna check that.
The normals are what they should be. Thanks though. =]