I'm also confused about step 2. It doesn't seem to guarantee that the edge coordinate is actually on the edge from a to b.

MrRockefeller

step 0 is converting polygons to quadrilaterals? but the pictures showed are triangles. I'm very confused of step 2 as well, cause from a subdivision like the picture in step 1, there is no analogy to point a and b in step 2's picture. Maybe in future lecture it would be easier to understand to make the 3 steps the same pictures.

(edit) I went look up catmull-clark and found the source code here:
http://rosettacode.org/wiki/Catmullâ€“Clark_subdivision_surface#Python

tacocat

What exactly is happening in step 2? What coordinates is it using?

dchen2

Step 2 seems very confusing. Are we computing the middle vertex of all faces first and then for each edge, we compute where the edges split by looking up the middle vertices of its neighboring faces? Isn't this very taxing on space? I guess alternatively, we could recompute the middle vertex each time, but then why is this step 2 and not step 1?

snaminen

Is the diagram in step 0 a possible outcome of Catmull-Clark subdivision?

snaminen

Does Catmull-Clark subdivision always result in new edges from the center of a polygon to every midpoint of its boundary edges?

ScreenTime

What do face coords and edge coords mean ?

manchas

In step 2 why wouldn't we just use the vertex positions of a and b?

superbluecat

Why is step 3, resetting the original vertices necessary for a nice subdivision?

David

I am also confused with step 2. When you keep one polygon the same but stretch the other one, the center of the other one would change. This doesn't seem to ensure the point staying on the line

large_monkey

In step 2, the resulting point need not lie on the segment from a to b. Is this intended? I suppose if c and d are equidistant from the line (like in this case) this will be true, but in general it seems this shouldn't hold.

anag

This visualizations of step 2 and 3 don't make a lot of sense, why exactly are we updating like this?

viceversa

What would be the value of a,b, and c, and would the weighting process being produced?

viceversa

How would the weighting process be produced, won't there be negative value here?

How does step 2 find the a, b, c, d edges?

I'm also confused about step 2. It doesn't seem to guarantee that the edge coordinate is actually on the edge from a to b.

step 0 is converting polygons to quadrilaterals? but the pictures showed are triangles. I'm very confused of step 2 as well, cause from a subdivision like the picture in step 1, there is no analogy to point a and b in step 2's picture. Maybe in future lecture it would be easier to understand to make the 3 steps the same pictures.

(edit) I went look up catmull-clark and found the source code here: http://rosettacode.org/wiki/Catmullâ€“Clark_subdivision_surface#Python

What exactly is happening in step 2? What coordinates is it using?

Step 2 seems very confusing. Are we computing the middle vertex of all faces first and then for each edge, we compute where the edges split by looking up the middle vertices of its neighboring faces? Isn't this very taxing on space? I guess alternatively, we could recompute the middle vertex each time, but then why is this step 2 and not step 1?

Is the diagram in step 0 a possible outcome of Catmull-Clark subdivision?

Does Catmull-Clark subdivision always result in new edges from the center of a polygon to every midpoint of its boundary edges?

What do face coords and edge coords mean ?

In step 2 why wouldn't we just use the vertex positions of a and b?

Why is step 3, resetting the original vertices necessary for a nice subdivision?

I am also confused with step 2. When you keep one polygon the same but stretch the other one, the center of the other one would change. This doesn't seem to ensure the point staying on the line

In step 2, the resulting point need not lie on the segment from a to b. Is this intended? I suppose if c and d are equidistant from the line (like in this case) this will be true, but in general it seems this shouldn't hold.

This visualizations of step 2 and 3 don't make a lot of sense, why exactly are we updating like this?

What would be the value of a,b, and c, and would the weighting process being produced?

How would the weighting process be produced, won't there be negative value here?