Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
Jaime Vives Piqueres
@jaimevives
Been there, done that... ;
opalite-phpp-90.jpg
POV-Ray media has Rayleigh scattering, which is very similar. It's not wavelength dependent, but that can be simulated with multiple medias for each wavelength.
Jaime Vives Piqueres
@jaimevives
I've been unsuccessfully trying to simulate opalopal-cabachon-05.jpg
Further reading seems to suggest that Tyndall scattering is possible using he Mie scattering in POV-Ray.
Jakob Flierl
@koppi
So, proper simulation of opal is very impractical / impossible then, I guess.
Jaime Vives Piqueres
@jaimevives
Opalite (the one you posted a picture) is OK with media, but true opal requires diffraction, which to my knowledge, no renderer does as of today. I tried several tricks with multiple medias, but none works for animation (the interior remains static, when it should "flow").
I tried some trick using "phase" on the color maps for the medias, but it didin't look natural.
Jakob Flierl
@koppi
Experimenting with BRIO bricks.
bpp-20160519-150200-00-hello-brio-01330.png
Jakob Flierl
@koppi

I just updated the demo/ folder a bit, see: https://github.com/bullet-physics-playground/bpp/tree/master/demo

I have included the citroen gs and the micra k11 into https://github.com/bullet-physics-playground/bpp/tree/master/includes

I have updated https://github.com/bullet-physics-playground/bpp/blob/master/demo/jaimevives/driveable-car.lua – the cars should move much more realistic now.

@jaimevives can you have a look into this file and check the tire radii against the two pov car models? I'm not sure, if I got these correct, see: https://github.com/bullet-physics-playground/bpp/blob/master/demo/jaimevives/driveable-car.lua#L67-L76 and https://github.com/bullet-physics-playground/bpp/blob/master/demo/jaimevives/driveable-car.lua#L90-L104 .

Also: I have re-used the Citroen GS's wheels with the Micra, because I wasn't able to get the correct alignment and rotation of the Micra's original wheels, can you look into this?

https://github.com/bullet-physics-playground/bpp/blob/master/includes/i_micra_k11.inc#L20-L30

Jaime Vives Piqueres
@jaimevives
Ok, will look into it later today....
Jaime Vives Piqueres
@jaimevives
Well... finally a free weekend! Should have plenty of time to test seriously.
About the surface-only liquids video: I didnt listen to it, as my english is mostly read-write, but the graphics look awesome. I guess it's a method for liquids without particles?
About the drivable car test video... how you did the wire graphics?
Jaime Vives Piqueres
@jaimevives
I'm already finding some issues... which is your preferred way: do I post them here or better I open issues on github?
Jakob Flierl
@koppi
Please open new issues on github.
Jaime Vives Piqueres
@jaimevives
This message was deleted
I have updated https://github.com/bullet-physics-playground/bpp/blob/master/demo/jaimevives/driveable-car.lua – the cars should move much more realistic now.
Hmmm... better, but I still feel they behaved better on the previous version.
BTW, I had to get ride of the timestep settings... otherwise the sim hangs on my machine.
And... why did you use a negative initial_angular_velocity???
Also: I have re-used the Citroen GS's wheels with the Micra, because I wasn't able to get the correct alignment and rotation of the Micra's original wheels, can you look into this?
I fixed that... see comments on the changes.
Jaime Vives Piqueres
@jaimevives
src/wrapper/lua_bullet.cpp:404:41: error: ‘addTriangleIndices’ is not a member of ‘btTriangleMesh’
.def("addTriangleIndices", &btTriangleMesh::addTriangleIndices)
cannot compile last changes... maybe bullet-physics-playground/bpp@7cbba56 ?
Jakob Flierl
@koppi
Hi Jaime!
On my Ubuntu 16.04:
$ dpkg -l|grep bullet-dev
ii  libbullet-dev:amd64                                         2.83.6+dfsg-3                                      amd64        professional 3D Game Multiphysics Library -- development files
2.83.6+dfsg-3
The bullet API maybe changed slightly.
I just did a clean checkout and noticed, that I forgot to push some changes..
I did the initial_angular_velocity = -10 to debug the freeze, but did not yet find the cause unfortunately.
Jakob Flierl
@koppi
$ grep -R addTriangleIndices /usr/include/bullet/
/usr/include/bullet/BulletCollision/CollisionShapes/btTriangleMesh.h: void addTriangleIndices(int index1, int index2, int index3 );
bullet-physics-playground/bpp@e50abc0
Jaime Vives Piqueres
@jaimevives
Thanks... it works now.
Jaime Vives Piqueres
@jaimevives
Well, it compiles, but I'm getting a few problems: a black view on most demos once I click on the window, then an non-functioning open dialog.
I guess I will have to update to 16.04... or do you prefer that I remain on 14.04 to try and fix the problems?
Now I remembered I've a second, unused disk on this computer: I will install 16.04 there to see if the problems are related to 14.04.
Jakob Flierl
@koppi
Screenshot from 2016-06-04 15-42-13.png
This is how the 00-hello.lua should look with the current version.
The mesh rendering is still WIP as I need to figure out a proper way of doing the vertex data caching in opengl.
Jaime Vives Piqueres
@jaimevives
00-hello runs fine, but almost any other fails when I click and drag on the view to move the camera: everything disappears and the view goes black.
Also, see how 03-trans looks on load:
Captura de pantalla de 2016-06-04 17:30:07.png
And with drivable-car, the car disappears when I press F2... too many strange things. I guess there is something fundamentally wrong on 14.04.
Jakob Flierl
@koppi
Thank's for the info. I can confirm all your observations here, too on Ubuntu 16.04.
I'm right now doing a little exercise with Valgrind to find all the memory leaks. Also trying to find out the cause of the blocking in the drivable-car-demo.