ali

import numpy node = hou.pwd() geo = node.geometry() # Add code to modify contents of geo. # Use drop down menu to select examples. #get array of all points attr "P" as string buffer... string_buffer = geo.pointFloatAttribValuesAsString("P", float_type=hou.numericData.Float32) #create numpy array from our string buffer... array_of_pos = numpy.fromstring(string_buffer, dtype='float32' ).reshape(-1, 3) # and reshape to make it 2dimentional (array of vectors) # doing something with your array... (google the numpy broadcast rules) array_of_pos -= (20, 4, 2) #per-element operation #and then you need to copy your numpy array back to points attr "P" geo.setPointFloatAttribValuesFromString("P", numpy.getbuffer(array_of_pos), float_type=hou.numericData.Float32) # WORKED (h13)!

The ability to add parameters to several selected nodes. And multi-threading everywhere )

http://rapidshare.com/files/427474043/SIM_SolverBullet_h11_winX64.rar

sdk: http://bullet.googlecode.com/files/bullet-2.77.zip source: http://gitorious.org/bullet-physics-...tarball/master SIM_SolverBullet_h11_winX64.rar

Could you help me please with compilation of RSL_Ocean.cpp for windows x64 ? I have compiled RSL_Ocean.cpp with command cl /I"c:/src/3rdparty/win64" /I"c:/src/3rdparty/include" /I"c:/src/3rdparty/include/loki" /I"c:/src/3rdparty/include/loki/flex" /I"c:/src/3rdparty/include/loki/yasli" /I"%DELIGHT%/include" "c:/src/3rdparty/win64/blitz.lib" "c:/src/3rdparty/win64/libfftw3f-3.lib" "%DELIGHT%/lib/3delight.lib" /LD /EHsc "c:/src/RSL_Ocean.cpp" the builded file RSL_Ocean.dll and libfftw3f-3.dll were placed to the folder /3Delight/shaders/ but when I tried to compile shader with the command shaderdl ocean_eval.sl , it has crushed. What am I doing incorrectly?

How I can make work "ragdoll_start.3.hip" on houdini9.5?

Thanks. I recreate this example using wire and constraints. But when I add box for collisions test I see that only wire is collide with it. The ball is not collide. So were is mistake? chainball.zip

In 8.x I can connect any constrain to multy-input of ApplyData node. In 9.5 this is don't work. Is this a bug?

ok. thanks. But what about time of this simulation? I looked at this example in XSI 7 and was very impressed. This scene was simulated in realtime!! Maybe this is dirty calculations, but chain isn't break and animation looks realistic.

Yes. This is fast and didn't break because in your example is only 5 chains and isn't have a ball on the end of chain. Please, download my last example and try it (22 chains).

>Use Volume Based Collision. un check this now simulation became slow. very slow. And this is not collada's bug. I recreate geo in houdini but nothing changed. chains1_new.hipnc

Just get "dynamic chain" example from XSI 7 and try to recreate it in Houdini. But chain is always break to pieces. I try to tune different parameters from RBD solver but problem is still here. In maya and XSI this scene simulate correctly with default solver parameters. So what magic parameter can help me to resolve this problem? chains1.zip

big thanks!

I have some strange reflections when I render glass sphere on reflectable grid. Can anybody help me? XP, 8.2.154, mantra image and scene in attachment glass_bug.zip

I try to render light with z-map shadow in intel core2duo pc (with option -n 2 or -H localhost, localhost), but z-map is not generate. What a solution?