QUOTE(shakebox @ 12/06/05, 10:50 AM)
Just try it: add a reorder node to your script with "rrrn" as value. This converts your image to a black and white only image, which has only one channel. Then measure the time it takes to render the same nodes with and without this reorder node.
[snapback]221704[/snapback]
I'm going to post this shake example of what you're talking about. There are 3 outputs, one has a reorder(rrrr) and then a randomize, one has reorder(rrrn) and then randomize, and one has randomize then reorder(rrrn). All 3 outputs take the same amount of time. Also, if you view the Randomize2 (on the right) it takes the same amount of time that it does to render after the reorder. Render times are the same with flipbook render as well. Randomize2 is full color, so it obviously will operate on all channels. Based on render times, even the gray versions are rendering all channels no matter when the reorder takes place. I was hoping I could do something to make it take 1/3 or 1/4 the time when I only want to operate on gray.
Here is my script:
___________________________________________________
// Input nodes
ColorWheel1 = ColorWheel(720, 486, 1, 0, 1, 1, 1);
// Processing nodes
Randomize2 = Randomize(ColorWheel1, 4, 57, 60, xAmplitude, 5.4,
0);
Reorder1 = Reorder(ColorWheel1, "rnnn");
Reorder3 = Reorder(ColorWheel1, "rrrr");
Randomize1 = Randomize(Reorder1, 4, 57, 60, xAmplitude, 5.4,
0);
Randomize3 = Randomize(Reorder3, 4, 57, 60, xAmplitude, 5.4,
0);
Reorder2 = Reorder(Randomize2, "rrr0");
____________________________________________________________