At 07:02 �� 28/3/2002, you wrote:

>Phoebus Dokos wrote:
> > ie. 01111111 00011111 11111111 as a starting colour - but calculating that
> > wouldn't be difficult).
>
>If you just want to halve the colour with each step as in your example
>this is fairly easy. My 16bit shadow routine does it this way: (d5 is
>preset to %1110111101111011 for mode 32 and %0111101111011110 for mode
>33)
>         move.w  (a1)+,d0                ; gggbbbbbrrrrrggg (mode 32 example)
>         ror.w   #1,d0                   ; ggggbbbbbrrrrrgg
>         and.w   d5,d0                   ; ggg0bbbb0rrrr0gg = 1/2 RGB factors
>
> > Would ORing with a predetermined bit mask would be best or something else?
>
>OR always adds something to some data. You can't fade something out
>with it ;-)

Did I say OR? It's getting late and I am up to my shoulders in Hydrostatic 
problems in Physics :-)))

So the masking is indeed what's the best solution... I see... :-) An added 
benefit as I see it is that you don't have to wait to equalise the colours 
before you start a full fade... The mask takes care of that although it 
should appear like "hesitating" before getting to a point where all colours 
bits are phased out together (Okay it may not even happen if for example 
all Green bits are 0 :-)

Nonetheless I think I got it this time :-D (Okay minus the OR ;-)

Thx Marcel,


Phoebus

Reply via email to