> I tested it...

Ah. How clever of you. =)

> What's done in Compensate_For_Alpha_AA() is not really accurate
> scientific method and therefore it is not working properly for all kind
> of input values.  It is not yet decided what will ultimately be the
> solution but of course there will be one.

It seems to me that the correct thing to do is (a) treat the background as
rgb <0,0,0> (which I believe is already done) and (b) at the end, divide the
color by one minus the amount of transparency (if zero then just leave the
color as is). Are there cases where that doesn't work? The hardest thing, I
would think, is determining what the transparency value really is in the
case of things like emission media, but that seems to be handled already.

 - Slime
 [ http://www.slimeland.com/ ]

Post a reply to this message