Re: how a w=0 vertex is tranformed and rasterized?
- From: "Dave Eberly" <dNOSPAMeberly@xxxxxxxxxxxxxxx>
- Date: Tue, 14 Feb 2006 14:43:30 GMT
<seayoung@xxxxxxxxx> wrote in message
news:1139902867.624206.252190@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
if in the projection matrix ,the Zn is 10, the Zf is 110,the Vw is
10,then the vector (5,0,10,0) after the projection
transform is (10,0,11,10),after divided by w,it is (1,0,1.1,1),how can
the graphics card knows it is a vector and not a
position. a point of that value is possible.How the graphics card does
with this?
PosView might have w = 0, but its projection does not.
The output 4-tuple position of a vertex shader must be in
"clip space" coordinates. The rasterizer uses these and does
the perspective divide itself, and eventually calls the pixel
shader for each pixel visited during rasterization.
Regarding "How the graphics card does with this?". Your vertex
and pixel shaders *are* telling the graphics card how to deal
with this :)
--
Dave Eberly
http://www.geometrictools.com
.
- Follow-Ups:
- Re: how a w=0 vertex is tranformed and rasterized?
- From: seayoung
- Re: how a w=0 vertex is tranformed and rasterized?
- References:
- how a w=0 vertex is tranformed and rasterized?
- From: seayoung
- how a w=0 vertex is tranformed and rasterized?
- Prev by Date: very fast graphics library
- Next by Date: Fast polygonal mesh aproximation with spheres
- Previous by thread: how a w=0 vertex is tranformed and rasterized?
- Next by thread: Re: how a w=0 vertex is tranformed and rasterized?
- Index(es):
Relevant Pages
|
