n-engine / nvidia-mesh-tools Goto Github PK

What steps will reproduce the problem?
1. run debug mode
2. nvbaker.exe --lomesh zsword.OBJ --himesh zsword.OBJ --tangentSpace
3. crash at --Rasterizing base mesh

What is the expected output? What do you see instead?
normal map、vector displacement map ...

What version of the product are you using? On what operating system?
r1 an r2

Please provide any additional information below.
the program is stop here---->

   bool isTriangle() const { return m_face->isTriangle(); }

What steps will reproduce the problem?
1. lo_ZSword.obj is quade mesh , hi_ZSword_t.OBJ is triangle mesh ( from 
Decimation Master (http://www.pixologic.com/zbrush/features/decimation/)
2. --lomesh lo_ZSword.obj --himesh hi_ZSword_t.OBJ
3. Got  weird normal map

input those tow mesh in xNormal will get fine normal map

Compiling on Linux 64 gives me the following error, due to mismatch in size
between size_t and unsigned int. The attached patch fixes the problem.

[ 39%] Building CXX object
src/nvmesh/CMakeFiles/nvmesh.dir/render/NvTriStrip.cpp.o
/media/data/nvidia-mesh-tools-read-only/src/nvmesh/render/NvTriStrip.cpp:213:
error: ‘bool nv::TriStrip::GenerateStrips(const short unsigned int*,
size_t, nv::TriStrip::PrimitiveGroup**, short unsigned int*, bool)’ should
have been declared inside ‘nv::TriStrip’

What steps will reproduce the problem?

1. start nvmeshanalyzer with museumhallRD.obj from 
http://www.3drender.com/challenges/Natural_History/Natural_History_OBJ.rar
2. entering ParameterizationQuality::processTriangle with p = [0, 0, 0] results 
in an parametricArea of size 0
3. Ss and St calculations divide by zero!

AccMesh does not support triangle tessellation. I've attached a patch to
implement this. Feel free to use it or not, I've tried to follow the coding
standards, but it was mainly useful to me for testing some of the
subdivision methods.

Generation of tangent patches was never reenabled after refactoring the
AccMeshBuilder.