I didn't understand what step 2 refers to, or what a,b,c,d are... how does it relate to the original polygon?
barath
So, I think 'c' and 'd' are the face coordinates of polygon understudy and one of its neigbouring faces. Those two faces have a common edge which is 'ab' whose end points are 'a' and 'b'.
So, do we take every single polygon of the mesh and do this calculation and connect points to subdivide. Is that right? Also, why are we doing the step 0? is it to have an initial estimate of the control point?
keenan
@barath has it right: a and b are the endpoints of the edge; c and d are the face coordinates computed in step 1.
keenan
@barath Step 0 is updating the connectivity; steps 1 through 3 are assigning the vertex positions.
jifengy
Does Q in the equation refer to step 1, and R refer to step 2?
eryn
Wait if the edge coords are the avg of a,b,c,d, are they guaranteed to be on the edge? I think no? What if c and d are not equidistant to the edge?
keenan
@jifengy: Q is the average of all the face points from Step 1, around a given vertex. Likewise, R is the average of all the edge points from Step 2, around a given vertex. So: loop over neighbors, add them up, divide by number of neighbors.
I didn't understand what step 2 refers to, or what a,b,c,d are... how does it relate to the original polygon?
So, I think 'c' and 'd' are the face coordinates of polygon understudy and one of its neigbouring faces. Those two faces have a common edge which is 'ab' whose end points are 'a' and 'b'.
http://www.rorydriscoll.com/2008/08/01/catmull-clark-subdivision-the-basics/
This is a nice read.
So, do we take every single polygon of the mesh and do this calculation and connect points to subdivide. Is that right? Also, why are we doing the step 0? is it to have an initial estimate of the control point?
@barath has it right: a and b are the endpoints of the edge; c and d are the face coordinates computed in step 1.
@barath Step 0 is updating the connectivity; steps 1 through 3 are assigning the vertex positions.
Does Q in the equation refer to step 1, and R refer to step 2?
Wait if the edge coords are the avg of a,b,c,d, are they guaranteed to be on the edge? I think no? What if c and d are not equidistant to the edge?
@jifengy: Q is the average of all the face points from Step 1, around a given vertex. Likewise, R is the average of all the edge points from Step 2, around a given vertex. So: loop over neighbors, add them up, divide by number of neighbors.