xkwooo
August 7th, 2007, 10:54 AM
Hi,
How is the logics behind matrix concatenation? Most CG books I've read say the matrix concatenation take a right to left order, so the rule is V = M * V0 = Mn * Mn-1 * ... * M0 * V0. And it also means the earlier the matrix be applied, the further be it from the equal-sign.
But one book about D3D says the homo vector representing a point or vector should be a ROW vector, and it should be the left hand multipier when used for multiplication, so there always be World * View * Projection tranformation sequence applied to position vector. Does D3D treat concatenation form left to right?
And the document states the vector will be packed into column vectors in intrinsically by default. Is this only about physical format or logically?
How is the logics behind matrix concatenation? Most CG books I've read say the matrix concatenation take a right to left order, so the rule is V = M * V0 = Mn * Mn-1 * ... * M0 * V0. And it also means the earlier the matrix be applied, the further be it from the equal-sign.
But one book about D3D says the homo vector representing a point or vector should be a ROW vector, and it should be the left hand multipier when used for multiplication, so there always be World * View * Projection tranformation sequence applied to position vector. Does D3D treat concatenation form left to right?
And the document states the vector will be packed into column vectors in intrinsically by default. Is this only about physical format or logically?