自适应多采样扫描转换算法

随着片元着色器的可编程能力不断提高和新型图元的出现,扫描转换必须加强多采样时的处理能力.经典的边方程扫描转换算法便于实现多采样,但采样点测试的计算量大并且存在冗余测试,为此提出一种自适应多采样扫描转换算法.该算法的多采样集中在"边界片元",并结合点采样和区域采样技术根据自适应细分规则细分边界片元进行采样点的测试;避免了边方程算法中进行采样点测试时的乘法运算,减少了采样点的冗余测试,提高了多采样时扫描转换的性能.实验结果表明,文中算法在较小代价下达到或接近有相同采样点数的超采样反走样能力.

Adaptive Multi-sampling Scan Conversion Algorithm

With the increasing programmability of the fragment shader and the emerge of new graphics primitives,scan conversion needs improving its processing ability for multi-sampling.The edge equations algorithm is easy to implement multi-sampling,but the cost of sample testing is expensive and there is redundant testing.To solve these problems,this paper presents an adaptive multi-sampling scan conversion algorithm,which performs the sampling inside "boundary fragment" to reduce the redundant computation.The algorithm performs sample testing by combining the point sampling with area sampling techniques,and subdivides the boundary fragments recursively according to the proposed adaptive rules.The algorithm substantially reduces redundant testing and avoids multiplication operation,and improves the performance of scan conversion with multi-sampling.The experimental results show that the algorithm achieves results that are comparable to super-sampling anti-aliasing with the same number of samples at minimum cost.