Half-Life 2 uses a method that combines lightmaps and standard ppl, there's a paper on the nvidia developers site, perhaps you should take a look.

Quote:

Original post by python_regious Half-Life 2 uses a method that combines lightmaps and standard ppl, there's a paper on the nvidia developers site, perhaps you should take a look.

Do you have a link to that?I can't find it.

Bugger, I can't find it either. It was certainly there...

It's at ATI's site;

http://www2.ati.com/developer/gdc/D3DTutorial10_Half-Life2_Shading.pdf

Odin

Ah, well. Wasn't too far off. [lol]

It seems interesting.They use a technique called "Radiosity Normal Mapping",but I can't figure it out.What are they storing in the lightmap?

Quote:

Computing Light map Values: •Traditionally, when computing light map values using a radiosity preprocessor, single color value is calculated •In Radiosity Normal Mapping, we transform our basis into tangent space compute light values for each vector. At the pixel level. . . Transform the normal from a normal map into our basis Sample three light map colors, and blend between them based the transformed vector lightmapColor[0] * dot( bumpBasis[0], normal )+lightmapColor[1] * dot( bumpBasis[1], normal )+lightmapColor[2] * dot( bumpBasis[2], normal )

When they say "compute light values for each vector",do they mean that,when building the radiosity map,they use T,B,N vectors as normals and calculate 3 different light values?And then,at the pixel level,perform some kind of weighting between these 3 values based on the direction of the normal obtained from the normalmap?


[Edited by - mikeman on September 5, 2004 10:24:24 AM]