I have the following method that gets a rgb value and classifies it using a smaller palette:
private static int roundToNearestColor( int rgb, int nrColors )
{
int red = ( rgb >> 16 ) & 0xFF;
int green = ( rgb >> 8 ) & 0xFF;
int blue = ( rgb & 0xFF );
red = red - ( red % nrColors );
green = green - ( green % nrColors );
blue = blue - ( blue % nrColors );
return 0xFF000000 | ( red << 16 ) | ( green << 8 ) | ( blue );
}
The code that annoys me is
red = red - ( red % nrColors );
green = green - ( green % nrColors );
blue = blue - ( blue % nrColors );
I am sure there is an alternate bitwise version of it that will perform faster, but as my bitwise arithmetic is a bit rusty, I have trouble finding such an expression. Any help or comments would be appreciated.