MItGeometry in deform always works in local space. That's why they pass in the world matrix into the deform function. Anyways, if you're really worried about performance, you should write the deformer in C++
Chad. On Thu, Jan 27, 2011 at 8:35 AM, Ofer Koren <[email protected]> wrote: > Maybe you could apply the world-to-local matrix on the deformer? instead of > bringing all the points to world-space and then applying the deformation, > see if it's less computations to bring the deformations into local space. It > might not be feasible, depending on what your deformer does... > > - Ofer > www.mrbroken.com > > > > On Wed, Jan 26, 2011 at 4:37 PM, hotchpotch <[email protected]> wrote: > >> I’m creating a basic deformer using the python api (Ma\ya 2009) - the >> simplest way to get it working the way I want is to convert the in >> mesh to world space using the local_to_world matrix supplied in the >> deform method inside the geom iteration loop. This all works fine - >> except that it really slows down the plugin. Is there any way of >> getting MItGeometry’s .position function to return in worldSpace? The >> default is in local space - and the subsequent matrix mults to and >> from world space seems to add an silly amount of time. >> For example on a 10k vertex mesh the plugin runs at 15fps if I set it >> up to use localSpace - but at 10fps using worldSpace - mainly because >> I have to convert every point position within the iter loop to >> worldSpace and back - whereas if I keep things in localSpace I can >> cull a large proportion of position values before doing the >> deformation calculations. >> >> any help gratefully received >> >> -- >> http://groups.google.com/group/python_inside_maya >> > > -- > http://groups.google.com/group/python_inside_maya > -- http://groups.google.com/group/python_inside_maya
