Shopping Cart
Still trying to get xN VDM's working in MAX!
November 27, 2011 - 4:07am
- 410 reads
Hi Jogshy,
Just downloaded the new xNomal version 3.17.6. I have found a few issues but for the moment I'll stick to the ongoing problem I have been having.
You probably remember me from previous threads trying to get VDM (direction) maps working in Max 2012.
I had eventually accepted that it was just not possible to do this and that the reason was un-smoothed UV's in my obj exported from ZBrush.
So I thought I would have another look and loaded your VDM example settings in xNormal to find that it rendered what looks to be a valid map.
However this map still did not work in Max with mental ray so I think that this is a problem with how xN encodes the direction and height information in the map. Seems not to be compatible.
I am hoping that you can find the time to look into this issue as it affects the workflow possibilities of many wanting better displacement in Max.
I really appreciate all of your hard work with xNormal so far!
Tristan.
PS: what software did you use to model the VDM example sphere?
Jogshy,
Thanks for taking the time to clarify. I understand that xNormal is mainly designed for the game market so I won't push the issue any further.
I have (to my knowledge) tried everything you have suggested with the same errors so I guess this is a dead end.
Anyway, keep up the good work - many people appreciate it!
Cheers.
I used ZB to model the spiked ball.
The problem with 3dsmax/MR is that it expects the directions in RGB and the heights in other file. xNormal packs all in one file with 4 channels ( RGB= direction, A=heights ) which is better for game engines. So, you probably need to separate the RGB and alpha in Photoshop and then save the files separately.
Another thing I've realized is that some programs need the vectors in tangent space ( Modo for example ). so, you must render your VDM map with the tangent-space option enabled in xNormal.
Also, pls notice MR subdivides the model while xN not. My VDM method is really aimed to game engines so I just apply a cheap and fast planar tessellation+ vertex displacement with water-tight sampling.
Tessellation != Subdivision, don't forget that. Tessellation, strictly speaking, just creates sub-triangles in the same plane that the original triangle.
Subdivision creates new faces and vertices in a way that the model is curved and smoothed ( == Catmull-Rom, Loop, Butterfly, PN triangles, etc... ). So, to get smooth results in xNormal, you must pass your LP mesh already subdivided a few times or you'll get some strange "faceting" artifacts.
Finally, to get good results you should use the MatchUV feature. For that, your HP mesh should have to be created using your LP mesh as subdiv basis ( == create your LP mesh. subdivide it. Sculpt it ), so the HP's and LP's UVs could be shared. At that point, if you render your UV layout you will notice are identical in form, except that the HP's one contains much more faces. The MatchUV ( aka dual-parameterization ) makes the calculations much more accurate and you won't need to setup any cage or ray distances.
------------------------------------------------------
http://www.xnormal.net
http://santyhammer.blogspot.com
http://www.ratgpu.com