Why is it that the size of a face/poly doesnt matter? Why does a large polygon not take up more processing power than a smaller polygon? In the real world it would take longer to construct a larger object, so why not in virtual reality?
The only answer I can come up with is the number of vertices matter, but that theory has flaws...
If its the number of vertices, why do double sided faces take up twice as much processing power than a single face? it has the same amount of vertices right? Gah! I dont get it...
welding doesnt reduce poly count, but can dramatically reduce the data size of the model. cuse a bit of info is neaded to describe the location of each vertice. so less vertices means less data to describe the mesh.
did you "target weld" or did you use "weld selected". if you used Weld Selected then you had it set to high and you welded vertices that would close faces, its not meant to close faces just to get rid of overlapping vertices. the faces that were closed were most likely very small.