crossproduct


    Sep 27 2007 | 9:45 am
    Just a quick question to all the mathbuffs out there...
    I want to get the crossproduct of 2 vectors, and for this I set up 2 matrices going into jit.expr with the following expression :
    expr "(in[1].p[2]*in[2].p[3])-(in[1].p[3]*in[2].p[2])" "(in[1].p[3]*in [2].p[1])-(in[1].p[1]-in[2].p[3])" "(in[1].p[1]*in[2].p[2])-(in[1].p [2]*in[1].p[3])"
    I was wondering if this is correct, and maybe if there's a much more simple expression or object for this...?
    grtz.
    Brecht Debackere
    www.autofasurer.net M+32(0)498319302 F+32(0)70429559 Skype : Bdebackere brecht@visualantics.net

    • Sep 27 2007 | 10:56 am
      Hi Brecht,
      I'm not sure but maybe a shader can help. Parts of the shaders derivatives are don with matrix calculations. It's only an idea. Maybe a bit more complicated. But matrices calculations are much faster on gpu than on cpu.
      grr. Ranja
      2007/9/27, Brecht Debackere : > > Just a quick question to all the mathbuffs out there... > I want to get the crossproduct of 2 vectors, and for this I set up 2 > matrices going into jit.expr with the following expression : > > expr "(in[1].p[2]*in[2].p[3])-(in[1].p[3]*in[2].p[2])" > "(in[1].p[3]*in[2].p[1])-(in[1].p[1]-in[2].p[3])" > "(in[1].p[1]*in[2].p[2])-(in[1].p[2]*in[1].p[3])" > > I was wondering if this is correct, and maybe if there's a much more > simple expression or object for this...? > > grtz. > Brecht Debackere > > www.autofasurer.net > M+32(0)498319302 > F+32(0)70429559 > Skype : Bdebackere > brecht@visualantics.net > > > > > >
    • Sep 27 2007 | 11:50 am
      The idea is to create quad_strip ribbon my filling up a matrix, with the width of the ribbon being calculated using the crossproduct using previous positions :
      vector 1 being : current position - previous position
      vector 2 : previous position - the position before that
      and then calculating the crossproduct of those 2 vectors, normalizing it, and then some stuff with adding the current position to the crossproduct and using that value for the even vertices of the quad_strip. (I think it sounds clearer in my head, but it is for doing something like this : http://www.youtube.com/watch?v=fM1qj_OFY4w )
      On 27 Sep 2007, at 12:56, Ranja Achbar wrote:
      > Hi Brecht, > > I'm not sure but maybe a shader can help. Parts of the shaders > derivatives are don with matrix calculations. > It's only an idea. Maybe a bit more complicated. But matrices > calculations are much faster on gpu than on cpu. > > grr. > Ranja > > 2007/9/27, Brecht Debackere : > Just a quick question to all the mathbuffs out there... > > I want to get the crossproduct of 2 vectors, and for this I set up > 2 matrices going into jit.expr with the following expression : > > expr "(in[1].p[2]*in[2].p[3])-(in[1].p[3]*in[2].p[2])" "(in[1].p[3] > *in[2].p[1])-(in[1].p[1]-in[2].p[3])" "(in[1].p[1]*in[2].p[2])-(in > [1].p[2]*in[1].p[3])" > > I was wondering if this is correct, and maybe if there's a much > more simple expression or object for this...? > > grtz. > > Brecht Debackere > > www.autofasurer.net > M+32(0)498319302 > F+32(0)70429559 > Skype : Bdebackere > brecht@visualantics.net > > > > > >
      Brecht Debackere
      www.autofasurer.net M+32(0)498319302 F+32(0)70429559 Skype : Bdebackere brecht@visualantics.net
    • Sep 27 2007 | 5:50 pm
      Xray.jit.crossproduct maybe can help you. www.mat.ucsb.edu/~whsmith/
    • Sep 28 2007 | 10:04 am
      wonderfull. this does exatly what I want, and normalizes as well :) thx.
      On 27 Sep 2007, at 19:50, jasmin wrote:
      > > Xray.jit.crossproduct maybe can help you. > www.mat.ucsb.edu/~whsmith/ > -- > jaz
      Brecht Debackere
      www.autofasurer.net M+32(0)498319302 F+32(0)70429559 Skype : Bdebackere brecht@visualantics.net