Sign in to follow this  

Interpolating 1/z in screen-space

This topic is 2584 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

If you intended to correct an error in the post then please contact us.

Recommended Posts

I want to do perspective-correct interpolation. I took the formula (12) from here http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.3.211&rep=rep1&type=pdf and it seems something is a bit wrongly "inversed". The code:

mtx mat = world * cameraTranslate * cameraRotate * persp * ortho;
float s = 0.0f;

for (int i = 0; i <= 10; i++)
{
vec4 q(-5,0,0), Q(5,0,0);
q *= mat;
Q *= mat;
float wt = 1.0f/q.w + s*(1.0f/Q.w - 1.0f/q.w);
wt = 1.0f/wt;
...
s += 0.1f;

(during wt computation, q.w and Q/w contain view-space Z-values [both are < 0 since I have right-handed coordinate system])

And the perspective and ortho matrices:

mtx persp(-near, 0.0f, 0.0f, 0.0f,
0.0f, -near, 0.0f, 0.0f,
0.0f, 0.0f, -near-far, 1.0f,
0.0f, 0.0f, -near*far, 0.0f);

mtx ortho(2.0f/(right-left), 0.0f, 0.0f, 0.0f,
0.0f, 2.0f/(top-bottom), 0.0f, 0.0f,
0.0f, 0.0f, -2.0f/(far-near), 0.0f,
-(right+left)/(right-left), -(top+bottom)/(top-bottom), -(near+far)/(far-near), 1.0f);


Thoeretically, for such a positioning of points in screen-space:

q... . . .Q

I should get values like:

-10 -9.5 -9 -7 -3 -1

But I get:

-10 -6 -4 -2 -1.5 -1

So it looks a bit "inversed" and I can't figure out why

[Edited by - maxest on November 14, 2010 7:46:22 AM]

Share this post


Link to post
Share on other sites

This topic is 2584 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

If you intended to correct an error in the post then please contact us.

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

Sign in to follow this