shader - OpenGL ES 2.0 Per Fragment Lighting for Untextured Generated Geometry -
I am currently producing geometry instead of importing it as a model. It is necessary to calculate all normals within the application.
I have successfully implemented Gouraud shading, and now want to apply fol shading (per piece / pixel).
I have taken a look at the relevant tutorials online and have two camps: A simple code-of-a-phong reheifting of a shard code, which provides better lighting, in fact per-pixel The second does not work by using normal embedded maps inside the right texture, but they are produced in a modeling toolkit such as RenderMonkey.
My questions are:
How should I go about creating programs for normalization of my geometry, this is the form of a vertical set Is it considered? In other words, a set of discrete polygon points have been given, should the general rules of interpolated manual be calculated?
Should I store generalized normals online as textures in the form of texture, and if so, how can I do within the code instead of modeling software?
Dissection per piece at the top normals The calculation of light will definitely get better results (assuming the normal streamed normals in the fragment bag normal), and this is actually per-pixel, though the strength of the difference depends on the model nozzles and light variation . Have you tried to take it out?
As long as no norms change between your face (think of collision mapping) and only rotate per-normal norms between each other, a normal map is totally unnecessary, as You get any kind of normal built-in rasterizer while the normal mapping can give a good effect if you really have normal per-pixel variations (such as a lot Tin surface), it is not right way to per-pixel lighting.
If you calculate per-light light and rotate the colors or if you calculate the light of each piece (even if you add the per-vertex Normals between each other, So what is this classical fond shading?), Especially you have a large triangle or very bright surface (too much frequency light variation).
As you have said, if you do not have high frequency normal variations (a triangle), you do not need a normal map, and neither per-specific normals Repeat. You generate per-head normal as you did for per-vertical lighting (like the normal average of the adjacent face) the rasterist interpires for you.
You should try simple per-pixel light before entering the techniques like normal mapping. If you have not got such a thinly spotted geometry or very bright surface, you will surely see the difference in simple per-vertical lighting system. Then when it works, you can try normal mapping techniques, but to work, they definitely need to understand the meaning of per-pixel lighting and fong shading unlike guarding shadows.
Comments
Post a Comment