Re: MS standard (pnetlib)

"Peter \"Firefly\" Lund" <[email protected]> Thu, 4 Aug 2005 15:02:50 +0200 (CEST)
Newsgroups gmane.comp.gnu.dotgnu.developer
Message-ID <[email protected]>
On Thu, 4 Aug 2005, Aris Basic wrote:

> hmm to me is not really logical to make f23 = ff2233 at the most 0f0203
> but even that is stretching
> more likely f23 = f2 30 00 or 00 0f 23 :)

It makes perfect sense.  "#f23" specifies an RGB-triplet with 4 bits per 
channel and "#ff2233" specifies it with 8 bits per channel.

When converting from 4 bits per channel to 8 bits per channel one 
multiplies each channel value not by 16 but by 17 in order to cover the 
whole range properly (so F maps to FF instead of to F0).

-Peter