Hi David,
>> 1) new pixels = (old pixels & mask1) | (new pixels & mask2) >> >> Where mask1 and mask2 are the negated forms of each other. That one only works for boolean masks, but it works on 386. >> This even works for alpha masks: >> >> 2) new pixels = (old pixels "*" mask1) "+" (new pixels "*" mask2) That one needs a CPU and compiler which supports at least MMX or some newer vector unit, so you are right that there is no effective way to do this on 386 with real alpha masking. You are right that you could first check your mask for areas which can be done the 1) boolean way and then only do the more costly computations with multiplications for those pixels where the mask is neither 0.0 nor 1.0, but whether that really helps will depend on which fraction of your pixels has 0.0 or 1.0 (or 0 and 100 percent, if you prefer) alpha masks and which fraction has more complicated values... Regards, Eric PS: As http://locklessinc.com/articles/sat_arithmetic/ shows, even saturated addition is tricky to do without having actual hardware support for it (MMX has saturated operations, too). Also notable: http://locklessinc.com/articles/booleans/ (testing for 0, 1 or -1) ------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot _______________________________________________ Freedos-devel mailing list Freedos-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/freedos-devel
