Hey all, after last thread, all your great answers helped me to put together a nice shadowmapping system. I had a few higher level problems though, and would like whatever input ya''ll can provide. Biasing Issues: I''ve read the 2nd depth paper that Yann provided. From what I can gather, We cull the front faces, and use the backfaces for our shadowmaps. (our 2nd surface) then, we cycle through each pixel, and find out it''s illumination vale. (determined by the pixels around it). I''ve tried culling the front face, and haven''t seen that impressive of results. Which makes me wonder if the whole point of the equation is the pixel checking. Is useing a polygon offset (glPolygonOffset()) so bad, tha we have to resort to checking per pixel, to fix the biasing issue? Or am i missing something? Projection Reverse problem: Texgen projects my shadow map not only forwards, but backwards as well. I read somewhere that the solution to this, was to put a "clipping" pixel somehow over the image? Does that sound right? Can someone expand on it more? Fade: I''d like to be able to fade the shadowmap, so that further from the light, the shadow fades out. The only thought I had of this, was to be able to copy a gradient alpha on top of the shadow map. Has anyone else done something similar? Multiple Lights: Charles bloom mentioned the use of a gigantic shadow map that he placed all his shadow maps onto, and then just projected that single map onto the scene. Is this the standard way to do multiple lights? My only other thought was to bastardize the texgen/texture matrix. (that is, by using multiple push/pops and many multitextureARB units) If anyone can shed any light on these issues, it''d help alot. thank you ~Main == Colt "MainRoach" McAnlis Programmer www.badheat.com/sinewave