On Wed, Jul 19, 2000 at 11:53:30PM +1000, David Hodson <hodsond@xxxxxxxxxxxxxx> wrote:
> > No:
>
> No? If I render an object, and the edge of that object only covers
> half of a pixel, why does it need more than half the colour range?
I was talking about precision and not resolution.
> is that I could load an image file containing pre-multiplied alpha
> without being asked if the alpha should be un-multiplied. (Although
> I still think there's a useful distinction between pre-mult alpha
> and non-mult masks, and I would like to load an rgba image into
> Gimp and then add a mask layer.)
Pre-mulitplied alpha is a data representation which contains as much or less
information than the un-multiplied version.
The only reason to use pre-multiplied alpha is for speed, or ease-of-use.
(unless you argue that out-of-gamut values make sense). Since pre-multiplying
rgba images looses information (maybe a lot) there should be very good
reasons to switch the representation used to store the data.
As opposed to pre-multiplying image data to speed up display.
--
-----==- |
----==-- _ |
---==---(_)__ __ ____ __ Marc Lehmann +--
--==---/ / _ \/ // /\ \/ / pcg@xxxxxxxxxxxxx |e|
-=====/_/_//_/\_,_/ /_/\_\ XX11-RIPE --+
The choice of a GNU generation |
|
