[Bf-committers] M-Def Binding Accelerator
Juan Pablo Bouza
jpbouza at hotmail.com
Tue Sep 25 16:09:35 CEST 2012
Yep, no difference at all... I've tested it with a 40k human model and there is no noticeable difference between the two bindings at level 7.
I took the rig far away from its center, rotated it 180 degrees and everything was cool.
The BVH binding took less than a minute, against 40 minutes with the traditional binding.
+1 for Zanqdo!
> From: zanqdo at gmail.com
> Date: Mon, 24 Sep 2012 13:26:00 -0600
> To: bf-committers at blender.org
> Subject: [Bf-committers] M-Def Binding Accelerator
>
> Hi, there has been a patch for dramatically accelerating MDef binding
> using BVH for around a year and a half. You can see the code review
> here http://codereview.appspot.com/4529048/
>
> And this is the updated patch for r50858 (thanks dfelinto)
> http://www.pasteall.org/35522/diff
>
> There where some concerns about precision so I proposed to do some
> tests. Here's a comparison between the deformation result of a
> character binded with trunk vs patched build
>
> Active object is binded with the patch and orange (selected) object is
> binded with plain trunk
>
> http://www.3developer.com/tmp/mdef.png
>
> I have highlighted the areas that showed some deviation from trunk.
> This mesh is highly distorted by the mdef and moved many meters away
> from the original position.
>
> The strongest deviations are of about 0.002m
>
> The binding times are 0:14s vs 1:32s, that's a speedup of 650%
>
> On the practical side I have been using this patch since it first
> appeared to bind many complex characters and I've got no problems with
> it
>
> Could we get it in for 2.65?
>
> Cheers!
>
> Daniel Salazar
> patazstudio.com
> _______________________________________________
> Bf-committers mailing list
> Bf-committers at blender.org
> http://lists.blender.org/mailman/listinfo/bf-committers
More information about the Bf-committers
mailing list