Multi-sample antialiasing optimization via edge tracking
View Patent ↗An efficient method and system for multi-sample antialiasing in graphics processing is described. Geometric edges as well as implicit edges of primitives in a bin are identified by iteratively rendering bins of pixels. Selective multi-sample antialiasing is applied to pixels that are touched by either a geometric edge or an implicit edge; pixels that are fully covered are not antialiased.
1. A method comprising:
rendering an image space by using iterations over a plurality of bins which correspond to screen regions, each said bin including a plurality of pixels;
for a given fragment over a respective bin, attempting to identify geometric edges and attempting to identify implicit edges; and
rendering said given fragment within said respective bin, the rendering being done differently, either with or without multi-sample antialiasing, wherein if the given fragment lacks any geometric or implicit edges, the given fragment is rendered without antialiasing.
2. The method of claim 1 , wherein said geometric edges are identified from a coverage mask produced during rasterization.
3. The method of claim 1 , wherein fragments having either geometric or implicit edges are rendered using multi-sample antialiasing.
4. The method of claim 1 , wherein at the start of a bin, depth buffers and color buffers are cleared and any edge tracking reset.
5. The method of claim 1 , wherein said implicit edges are identified by comparing the minimum and maximum depth values of the pixels to which the fragment is added against the depth values of said fragment.
6. The method of claim 1 , wherein edges are tracked to a 4×4 pixel level.
7. The method of claim 1 , wherein information is collected concerning the presence or absence of edges that affect a target pixel in order to guide down-filtering of the colors of the pixel such that only those pixels that have been multi-sampled need to be down-filtered.
8. A computer system comprising:
a display;
a processor;
memory; and
one or more programs stored in the memory containing instructions for:
rendering an image space by using iterations over a plurality of bins which correspond to screen regions, each said bin including a plurality of pixels;
for a given fragment over a respective bin, attempting to identify geometric edges and attempting to identify implicit edges; and
rendering said given fragment within said respective bin, the rendering being done differently, either with or without multi-sample antialiasing, wherein if the given fragment lacks any geometric or implicit edges, the given fragment is rendered without antialiasing.
9. The system of claim 8 , wherein said geometric edges are identified from a coverage mask produced during rasterization.
10. The system of claim 8 , wherein fragments having either geometric or implicit edges are rendered using multi-sample antialiasing.
11. The system of claim 8 , wherein at the start of a bin, depth buffers and color buffers are cleared and any edge tracking reset.
12. The system of claim 8 , wherein said implicit edges are identified by comparing the minimum and maximum depth values of the pixels to which the fragment is added against the depth values of said fragment.
13. The system of claim 8 , wherein edges are tracked to a 4×4 pixel level.
14. The system of claim 8 , wherein information is collected concerning the presence or absence of edges that affect a target pixel in order to guide down-filtering of the colors of the pixel such that only those pixels that have been multi-sampled need to be down-filtered.
15. A non-transitory computer readable medium storing instructions for:
rendering an image space by using iterations over a plurality of bins which correspond to screen regions, each said bin including a plurality of pixels;
for a given fragment over a respective bin, attempting to identify geometric edges and attempting to identify implicit edges; and
rendering said given fragment within said respective bin, the rendering being done differently, either with or without multi-sample antialiasing, wherein if the given fragment lacks any geometric or implicit edges, the given fragment is rendered without antialiasing.
16. The non-transitory computer readable medium of claim 15 , wherein said geometric edges are identified from a coverage mask produced during rasterization.
17. The non-transitory computer readable medium of claim 15 , wherein fragments having either geometric or implicit edges are rendered using multi-sample antialiasing.
18. The non-transitory computer readable medium of claim 15 , wherein at the start of a bin, depth buffers and color buffers are cleared and any edge tracking reset.
19. The non-transitory computer readable medium of claim 15 , wherein said implicit edges are identified by comparing the minimum and maximum depth values of the pixels to which the fragment is added against the depth values of said fragment.
20. The non-transitory computer readable medium of claim 15 , wherein edges are tracked to a 4×4 pixel level.