whl@cl.cam.ac.uk (Wing Hung Lau) (01/23/91)
I have been trying to implement the A-buffer (Siggraph 1984) recently. However, I have a problem. I hope someone in the net who has experience on it or know the answer can give me a clue. In the paper, the fragment structure contains two fields zmin and zmax which, according to the author, are used for intersection detection and for visibility approximation when it occurs. The method used for doing these is also described in the paper. However, zmin and zmax are not included in the Pixelstruct definition. Only a single z field is inlcuded in it meaning that if a pixel is fully covered by two polygons and if these two polygons intersect each other there, then we would not be able to use the method. In other words, since we create a fragment only when there is a polygon edge crosses a pixel, it seems to me that we can only use the method to detect surface intersections which occur around polygon edges. Am I right or have I missed out something? I hope my question is clear! Rynson whl@uk.ac.cam.cl