I need to clarify my earlier remarks. I did not meant to say that the programmable pipeline controls hardware blending, but rather allows one to program arbitrary blending models that may not be offered by blending hardware.
It was also pointed out to me that using fixed function blending will be a lot faster than using the programmable pipeline because you will take advantage of blending hardware. The only exception to that, of course, if if you require FP32 blending, in which case it is likely you will have to use the programmable pipeline due to lack of FP32 blending support in most hardware in use at the moment.
I apologize for the confusion.
On May 16, 2008, at 16:17, Richard Salvatierra wrote: Is glBlendMode() the same as glBlendFunc()
This was not working. Continues to show the background (semi opaque)
[vid_A bindCurrentTexture]; glBlendFunc(GL_ALPHA, GL_ZERO); [vid_A compositeChannelInRect:theRect];
[vid_B bindCurrentTexture]; glBlendFunc(GL_ALPHA, GL_ONE); [vid_B compositeChannelInRect:theRect];
-Rich
|
