PDA

View Full Version : GL_NORMALIZE



nib
08-15-2008, 02:38 PM
If I recall, GL_NORMALIZE is there to renormalize when there is a scale in the model?

Is setting GL_NORMALIZE when only using shaders necessary? Or is this really only an older "fixed" pipeline thing? I remember it would hurt performance.

Running my game on the ATI card in fixed pipeline mode it seemed to be an issue. Sometimes certain polygons drawn after the scale would turn white. But on the NVidia cards this did not appear to be a problem when GL_NORMALIZE was disabled in fixed mode. But when I use shaders I'm sampling from normal maps or normalizing anyways and it did not appear to be an issue.

He he, took five days to figure it out. I thought it was blending. But when I make explosions I scale them differently in a sequence. :)

arekkusu
08-15-2008, 02:49 PM
Read FAQ #1 (http://www.opengl.org/resources/features/KilgardTechniques/oglpitfall/).

Auto normalization is bypassed when using shaders, as documented (http://www.opengl.org/registry/doc/glspec21.20061201.pdf) in sec 2.15.4 Shader Execution.

nib
08-15-2008, 08:47 PM
I enable normalize now. All is well. Think I've been burned by this before. Its such a simple mistake.

I like that default fixed pipeline mode if people have older macs etc. Though, in the future I'll go the shader only approach. Enough time has passed to start depreciating the fixed pipeline stuff in my future applications.

Thanks again.