hi GDNet, if i have vertex of object stored in this matrix: {50,0,0,0}, {0,50,0,0}, {0,0,50,0}, {0,0,0,1} ; x=50, y=50, z=50 and then translate this with this matrix(translate by x for 20): {1,0,0,0}, {0,1,0,0}, {0,0,1,0}, {20,0,0,1} then i get this matrix: m = {50,0,0,0}, {0,50,0,0}, {0,0,50,0}, {20,0,0,1} so to actual use this i must sum up m[0][0] and m[3][0] is this the right way?

I am a little confused about what you are trying to do. Normally you don''t need to store a vertex in a matrix, just a vector. So, you could say that your vertex has a location of
[ 50, 50, 50, 1 ].

Your first matrix is more like a scaling matrix that would increase an objects dimensions by a factor of 50 on all three axes.

Then to translate the vertex, you would multiply it by your matrix. m = vertex vector, N = your translation matrix

new location = m * N = [ 70 50 50 1 ]

A vertex is not a 4x4 matrix, it is a 4x1 matrix (also known as a "vector"). The multiplication of a 4x4 matrix with a vertex is another vertex -- the transformed vertex.

So you should be writing your vertices as {x, y, z, 1} and not in the matrix format you have. So, for your example, it would be { 50, 50, 50, 1 }.

Go to your favorite college/bookstore/library/friend/sample code and look for "linear algebra" which will talk about the process for multiplying matrices and vectors. For your sample matrix (that translates x by 20), the results of multiplying that by the vertex {50, 50, 50, 1} is {70, 50, 50, 1} -- exactly what you would expect.



---- --- -- -
Blue programmer needs food badly. Blue programmer is about to die!

thank you for reply, so if i have vertex and i want it first to translate it and then rotate i can do something like:
v..vertex
TM..translation matrix
RM..rotation matrix
NM..new matrix
new v = v*tm*rm
if i have 100 vertices and i must do all the same tranformation on them i can do first NM=TM*RM and then for all the vertices i do v(all)*NM
that is cool and it even makes it faster because it doesnt need so much multiplication,
can you correct me if i am wrong
bye

The NM matrix that you create is sometimes ( always ? ) called a transform matrix. What you need to be careful about is that

TM * RM != RM * TM

in most cases. You can even build up nice a transform matrix by using multiple RM''s and TM''s. For example, you could have TM1, RM1, RM2, and use these to get:

NM = RM1 * TM1 * RM2

thank you again for additional information, but lets consider one more situation...
if i get it right with camera if we have some object at z = 200 and we say that camera is at z = 50 then object_world_coordinates - 50, then in some sense we get as that object is 150 units from us.

so...

WC..world coordinates(if i get this right these newer move if they are stable..like walls, houses)
VC..view coordinates...there are world coordinates but transformed by camera matrix
CM..camera matrix

VC=WC*CM (in CM we translate,rotate in negative direction)

(WC consist of x,y,z,xr,yr,zr)

then in some program

loopingaround{
we_press_up->we update CM z coordinate
/*..like z=10
(CM={1,0,0,0}
{0,1,0,0}
{0,0,1,0}
{z,0,0,1}) */

VC=WC*CM;
}

ok is this correct, sorry for bother i know there is all in books but why not use this forum as great learning tool of the future
bye!