i cant edit my post :) i v made litle step, but now i cannot get the matrixoutput from jit.gl.mesh.

help help @:)

hi napentro.

i'm a little unclear on what you are trying to accomplish. if you set jit.phys.body @shape concave @mass 0, you can send your geometry matrix directly in to the phys.body.

you don't need a matrixoutput of gl.mesh (and gl.mesh does not support matrixoutput) just send the same matrix you are sending to gl.mesh, to phys.body.

hope that helps.

:) thanks for answer!

so maybe i will start from the begin. i would like to use binary image from webcam, to controll the phys object. i would like to do something like this http://www.youtube.com/watch?v=f_u3sSffS78.

unfortunately when i try to send the matrix to jit.phys, like in tutorial with matrix, thinks dont work as they should. i v got problems with the position rotation and scale of extruded shape in 3d, and everything is terribly slow. :(

help help @:))

please post your patch showing how you are sending image data to the physics object.

thank you!

any reason your not just using a jit.phys.body @shape cylinder and jit.phys.body?

other than that, your patch seems to be behaving as expected, so i'm not entirely sure what you need help with.

if it's a fps issue, disable the phys.gl.draw and decrease the dimensions of your input matrices.

for what it's worth, a future release will introduce a new physics object specifically designed to handle streaming matrices. but for now, a jit.phys.body @shape concave is the way to go.

I... think? I can't really tell, I'm having issues understanding Napentro - but I think i'm having a similar problem.

Whenever I pass my geometry matrix into a jit.phys.body, and it's set to either @shape concave or @shape convex, rotation and position totally die - the object resets itself to 0. If I disconnect the geometry matrix for jit.phys.body, rotation comes back online, but position acts... strangely - I can move it around with jit.phys.picker, but it doesn't have any inertia. It seems to be affected by outside forces (a jit.phys.ghost with a negative central force will pull it towards the center, but veeeeeery slowly compared to a simple phys object).

Anyway, I don't know if this is all the way it's 'supposed' to work, there's not that much information that I can find about the convexhull/concave setting in the literature or on the forums. Am I supposed to also be putting in a position matrix, somehow? That's sort of what I want to use jit.phys.body for in the first place, so that strikes me as odd.

My patch is a bit too big (and currently disorganized) to post, sadly. Hopefully what I typed above made sense - it's 7am and I haven't slept.

OK so it would appear that regularly updating the vertex matrix also forces the jit.phys.body object to reset completely, hence all my issues. Which is sad, I wanted to change the shape of it on-the-fly, not just at first load! (I was toying with an example someone posted on another thread).

Any reason that feeding jit.phys.body a constantly-updating matrix of vertex points isn't an available thing? I suppose if a vertex suddenly moves from within world bounds to out-of-bounds it could cause some mathy-crashy issues? Anyway if anyone has some insight I'm interested to know. For now I guess I'll use spheres.

Slightly spammy - but another fun thing. Setting a jit.phys.body to convexhull, feeding it a matrix, then changing it back to sphere, makes the physics system act all strange until I quit max and relaunch. It's like all my physics objects are moving in jello.

any update on this project? i'm having some issues as well in finding a way to have a webcam input and use it as collider body

hi, this can be a good starting point : https://cycling74.com/forums/cameratophysic2d

i found it yesterday in my research, but as i'm going to use multiple 3d cubes, i have huge problems in having the multiple reacting with my shape (while the poly int he patch reacts amazingly good)...