Re: [Bf-committers] AAC and ACC3 support
On Nov 19, 2012, at 6:38 PM, Sergey Sharybin wrote: Hi, Well, issue is simple -- in FFmpeg this codecs are marked as experimental and unless you're configuring export context in a proper way (enable experimental codecs) you can not use this codecs. It seems that if FFmpeg is compiled with libfaac this codecs could be used. Bad thing that it makes FFmpeg unable to be redistributed via blender.org. I've made quick tests trying to make AAC/AC3 working without libfaac (seems FFmpeg does have own implementation for this codecs) but it ended up in only noise in output. The question is: is there somebody who're familiar with this stuff and could make it work or we're simply disable this codecs? I just submitted a patch, from the tracker: this patch enables the use of the AAC and AC3 encoders in ffmpeg. These are marked as experimental, yet they seem to be in widespread use and all the information I could find on what makes them experimental is that they create files with suboptimal quality per bitrate, i.e. for a given bitrate the quality achieved could be better. This seems to be an issue only for lower bitrates, any rates 160kbps seem to be transparent anyways. Note: I have not tested the patch more than just quickly encoding one test file, I don't have time to look at this longer unfortunately at the moment. http://projects.blender.org/tracker/index.php?func=detailaid=33242group_id=9atid=127 till then, David. -- With best regards, Sergey Sharybin ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
Re: [Bf-committers] AAC and ACC3 support
Hi David, Thanks for the patch. From quick glance seems ok. Will do some more tests later tonight. On Tue, Nov 20, 2012 at 7:10 PM, David erwi...@gmx.net wrote: On Nov 19, 2012, at 6:38 PM, Sergey Sharybin wrote: Hi, Well, issue is simple -- in FFmpeg this codecs are marked as experimental and unless you're configuring export context in a proper way (enable experimental codecs) you can not use this codecs. It seems that if FFmpeg is compiled with libfaac this codecs could be used. Bad thing that it makes FFmpeg unable to be redistributed via blender.org. I've made quick tests trying to make AAC/AC3 working without libfaac (seems FFmpeg does have own implementation for this codecs) but it ended up in only noise in output. The question is: is there somebody who're familiar with this stuff and could make it work or we're simply disable this codecs? I just submitted a patch, from the tracker: this patch enables the use of the AAC and AC3 encoders in ffmpeg. These are marked as experimental, yet they seem to be in widespread use and all the information I could find on what makes them experimental is that they create files with suboptimal quality per bitrate, i.e. for a given bitrate the quality achieved could be better. This seems to be an issue only for lower bitrates, any rates 160kbps seem to be transparent anyways. Note: I have not tested the patch more than just quickly encoding one test file, I don't have time to look at this longer unfortunately at the moment. http://projects.blender.org/tracker/index.php?func=detailaid=33242group_id=9atid=127 till then, David. -- With best regards, Sergey Sharybin ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers -- With best regards, Sergey Sharybin ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
Re: [Bf-committers] Blender issues on Windows
It seems to me we almost need another lib checkout to deal with vc10, similar to the mingw ones. It is rather annoying to have to download a pile of Boost libs only to have to turn around and compile my own anyways. We could use an svn external to handle libs that both can use (share). Also, the build.bat needs modifications to build libs for vc10 (different folder structure). --Mitchell Stokes On Mon, Nov 19, 2012 at 10:57 PM, Sergey Sharybin sergey@gmail.comwrote: Just for the note, All issues with windows and msvc2008 should be solved now. However, not sure if msvc2010 work now. Also not sure if it's possible to avoid commiting 300 more megabytes of debug libs to make msvc2010 happy. Still would be nice if somebody soublecheck on this :) Thanks Thomas and Brecht for solving windows issues :) On Mon, Nov 19, 2012 at 7:27 PM, Sergey Sharybin sergey@gmail.com wrote: Thanks, I've commited tweaks to CMake so localization _shouldn't_ require debug library anymore. However, i do have quite the same backtrace on startup as with i18n disabled and OIIO enabled. Perhaps that's related somehow. On Mon, Nov 19, 2012 at 7:25 PM, Thomas Dinges blen...@dingto.org wrote: Hi Sergey, I dropped debug boost libs from SVN as they are quite big (160-200 MB per architecture). In case someone needs them, you can find them here (for x32 and x64): http://blender.dingto.org/builds/dev/ Am 19.11.2012 13:18, schrieb Sergey Sharybin: Hello everyone, I've tried to work on some bugs on Windows today and i couldn't do that. First issue was missed debug library for i18n (seems it's substituding by #pragma in boost includes as a recommended). -- Thomas Dinges Blender Developer, Artist and Musician www.dingto.org ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers -- With best regards, Sergey Sharybin -- With best regards, Sergey Sharybin ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
Re: [Bf-committers] Blender issues on Windows
An alternative to avoid huge libraries is to add cmake/scons support to build BOOST from source, at least for Windows. Just copy the BOOST source code in blender/extern for example. Of course this can stay pure optional, so it will be ignored if you use precompiled libraries. On 20 November 2012 14:08, Mitchell Stokes moguri...@gmail.com wrote: t seems to me we almost need another lib checkout to deal with vc10, similar to the mingw ones. It is rather annoying to have to download a pile of Boost libs only to have to turn around and compile my own anyways. We could use an svn external to handle libs that both can use (share). Also, the build.bat needs modifications to build libs for vc10 (different folder structure). ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
Re: [Bf-committers] Blender issues on Windows
ho, I voice up rarely because I use pretty aged versions of blender in our production line. Building Blender for Windows from the 'hot rod' is a challenge every time I try. Most annoying is downloading the entire lib/windows mess for an/some hour(s) on a clean (virtual) machine. Yes .. we do use virtual machines .. just to tame the windows version zoo on the customer side :) IMHO that directory needs to be structured like lib .. windows (..target switch) ..32 ..common ..development system ..Unix clones ..ms likes ..64 ..common ..development system ..Unix clones ..ms likes so I can check out the things I really need old man's grumbling ends here .. almost, A symbolic hug at the flying Dutchman I know it is hard to control a pile of ants. Am 20.11.2012 23:28, schrieb Erwin Coumans: An alternative to avoid huge libraries is to add cmake/scons support to build BOOST from source, at least for Windows. Just copy the BOOST source code in blender/extern for example. Of course this can stay pure optional, so it will be ignored if you use precompiled libraries. On 20 November 2012 14:08, Mitchell Stokes moguri...@gmail.com wrote: t seems to me we almost need another lib checkout to deal with vc10, similar to the mingw ones. It is rather annoying to have to download a pile of Boost libs only to have to turn around and compile my own anyways. We could use an svn external to handle libs that both can use (share). Also, the build.bat needs modifications to build libs for vc10 (different folder structure). ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
[Bf-committers] Cursor to center of mass
Hi! I've written some center of mass calculations in python for closed meshes, but there is an issue with using tessfaces in my code; I run mesh.update with tessfaces=True but when I loop over mesh.tessfaces there are only 6 faces in default scene cube so the center of mass is offset. It seems to work when the mesh is triangulated http://www.pasteall.org/37396/python # Code import mathutils from mathutils import * import bpy tess_count = 0 # Handle tri as a tetrahedron with fourth point at origo def handleTri(v1, v2, v3): global tess_count tess_count = tess_count + 1 temp = (v3-v1).cross(v2-v1) nor = temp.normalized() area = 0.5*temp.dot(nor) vol = area*nor.dot(v1)/3.0 centroid = (v1+v2+v3)/4.0 return (centroid, vol) for me in bpy.data.meshes: me.update(calc_tessface=True) sumc= Vector() summ = 0 for f in me.tessfaces: v1 = Vector(me.vertices[f.vertices[0]].co) v2 = Vector(me.vertices[f.vertices[1]].co) v3 = Vector(me.vertices[f.vertices[2]].co) centroid, mass = handleTri(v1, v2, v3) sumc += centroid * mass summ += mass print(tess_count) print(sumc/summ) ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
Re: [Bf-committers] Cursor to center of mass
Tessfaces includes both triangles and quads, it doesn't triangulate entirely, so you'd need to handle the quads still. Brecht. On Wed, Nov 21, 2012 at 4:30 AM, Ummi Nom ummi...@gmail.com wrote: Hi! I've written some center of mass calculations in python for closed meshes, but there is an issue with using tessfaces in my code; I run mesh.update with tessfaces=True but when I loop over mesh.tessfaces there are only 6 faces in default scene cube so the center of mass is offset. It seems to work when the mesh is triangulated http://www.pasteall.org/37396/python # Code import mathutils from mathutils import * import bpy tess_count = 0 # Handle tri as a tetrahedron with fourth point at origo def handleTri(v1, v2, v3): global tess_count tess_count = tess_count + 1 temp = (v3-v1).cross(v2-v1) nor = temp.normalized() area = 0.5*temp.dot(nor) vol = area*nor.dot(v1)/3.0 centroid = (v1+v2+v3)/4.0 return (centroid, vol) for me in bpy.data.meshes: me.update(calc_tessface=True) sumc= Vector() summ = 0 for f in me.tessfaces: v1 = Vector(me.vertices[f.vertices[0]].co) v2 = Vector(me.vertices[f.vertices[1]].co) v3 = Vector(me.vertices[f.vertices[2]].co) centroid, mass = handleTri(v1, v2, v3) sumc += centroid * mass summ += mass print(tess_count) print(sumc/summ) ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers
[Bf-committers] Leap Motion
Has anybody been thinking about the Leap Motion input device? https://leapmotion.com/ I'm only asking because as part of my research I want to add support for it to Blender, but if anybody else is interested I do not want to duplicate work when we could work together. ___ Bf-committers mailing list Bf-committers@blender.org http://lists.blender.org/mailman/listinfo/bf-committers