| 00:34.29 | brlcad | starseeker: uhm, because "The code embodies decades of research and experience by Pixar, and a more recent and still active collaboration on fast GPU drawing between Microsoft Research and Pixar." :) |
| 00:36.46 | brlcad | interesting font, looks like adobe themselves published it: http://www.tested.com/news/443102-adobe-releases-monospaced-open-source-font-source-code-pro/ |
| 00:38.09 | brlcad | elf11_: I guess I'm still not seeing the problem -- if you know the values being fed to bullet are always a specific unit, you are providing bullet with a model sized consistently |
| 00:39.01 | brlcad | elf11_: is the problem that small objects will traverse a large distance with very few simulation steps or something like that? |
| 00:48.28 | elf11_ | brlcad, the thing is that you are feeding bullet what's stored in brl-cad, so if for example the simulation I did in mm will be translated to m, then since brl-cad stores sizes in mm, bullet will get 50000 m and not 50 000 mm from brl-cad, hence the scaling |
| 00:49.36 | elf11_ | and small geometry results are not reliable because, bullet has a tolerance of 0.04m that's like 40mm, the simulation I did the biggest geometry was 30mm, lower than the tolerance |
| 00:49.42 | elf11_ | so the cube was penetrating the surface |
| 00:50.33 | brlcad | elf11_: okay, but the mm to m translation would just be a one-time bug to get fixed |
| 00:51.14 | brlcad | that also seems like a very high tolerance -- is it configurable? |
| 00:52.15 | brlcad | moreover, what would be the impact of scaling the geometry up so that it's always an order bigger than that tolerance? |
| 00:52.52 | elf11_ | the mm to m translation is a few lines of code fix |
| 00:53.23 | elf11_ | about the configurable part looked into it, didn't find yet a way to get it lower |
| 00:53.27 | brlcad | or always scaling 10^6 even unless values are >10^6 already |
| 00:55.56 | elf11_ | hmm I tried the simulation I did for values like 10^2 bigger than the ones used, and it worked - the cube didn't fall through the plane, like just meter scaling all the geometry under the tolerance |
| 01:08.46 | brlcad | elf11_: the issue might be the rate at which it simulates |
| 01:09.29 | brlcad | you want X time steps to equate to the same amount of time and for the mass being simulated to reflect the sizes in brl-cad, not whatever size it might have been scaled to |
| 01:10.12 | brlcad | you could compensate with a 10^6 scaling with a similar mass scaling, but it'd have to be done carefully to be correct |
| 01:11.11 | brlcad | best to start as simple possible, unscaled and just document that you the model should be larger than 40mm in order for collision detection to be evaluated correctly |
| 01:11.58 | elf11_ | okay |
| 01:45.28 | *** join/#brlcad PrezKennedy (~DarkCalf@2002:ade7:2862::ade7:2862) |
| 02:33.58 | *** join/#brlcad elf_ (~elf@92.80.57.5) |
| 03:04.56 | *** join/#brlcad yukonbob (~bch@c-50-135-255-34.hsd1.wa.comcast.net) |
| 03:04.59 | yukonbob | hello #brlcad |
| 03:17.47 | brlcad | howdy |
| 03:18.46 | yukonbob | :) |
| 03:19.13 | yukonbob | just saw news of Sony OSL, looked at license. Very impressed. |
| 03:42.57 | yukonbob | brlcad: how're things going? |
| 03:57.25 | *** join/#brlcad yukonbob (~bch@c-50-135-255-34.hsd1.wa.comcast.net) |
| 05:04.06 | *** join/#brlcad abhi2011 (~chatzilla@117.200.84.29) |
| 08:40.26 | *** join/#brlcad archivist (~archivist@host81-149-189-98.in-addr.btopenworld.com) |
| 13:34.37 | brlcad | starseeker: interesting, rhino mostly punted: http://wiki.mcneel.com/rhino/meshtonurb |
| 13:35.02 | brlcad | though that base case does have some interesting implications that could be useful to support |
| 13:40.37 | *** join/#brlcad crdueck (~cdk@24-212-219-10.cable.teksavvy.com) |
| 16:34.34 | *** join/#brlcad Al_Da_Best (Al_Da_Best@cpc2-shep12-2-0-cust21.8-3.cable.virginmedia.com) |
| 16:43.49 | *** join/#brlcad tobiasu_ (~tobiasu@tin.tmux.org) |
| 17:10.26 | *** part/#brlcad PrezKennedy (~DarkCalf@2002:ade7:2862::ade7:2862) |
| 17:11.35 | *** join/#brlcad DarkCalf (~DarkCalf@2002:ade7:2862::ade7:2862) |
| 17:35.39 | *** join/#brlcad Yoshi47 (~jan@64.235.102.210) |
| 20:31.14 | *** join/#brlcad elf_ (~elf@92.80.33.182) |
| 20:53.42 | starseeker | Hmm - CIA still not working |
| 21:13.28 | *** join/#brlcad Yoshi47 (~jan@d24-204-236-81.home4.cgocable.net) |
| 23:24.28 | *** join/#brlcad tobiasu (~tobiasu@tin.tmux.org) |