At 12:58 �� 10/11/2001 +0100, you wrote: > > I've started to have a look at this. QPC (in 65536 colour mode, i.e. > > Mode32) stores the colour values in the screen memoey as follows: > > > 16 bits per pixel > > gggrrrrrbbbbbggg > >On the Q40, it is gggggrrrrrbbbbbl >where l is a luminance value, You in fact use 6 bits for each of the >colours, and when the lower bit is set on two or three of the colours, then >l is true. Then build the colour using the most significant five bits only >(for some demo code, look in the PROforma sources, pf_driv_QVME_c and the >the ColourNearest routines.
I see, however do the colour codes from S*Basic create the same results for comparable modes in Q40 and QPC? (ie is there a "translation table" of sorts INSIDE SMS? Or should we construct one to be accurate? Additionally do COLOUR_PAL in its default setting display the same colours from 0-255 for all Platforms supporting colour drivers? >I personally don't think there is any point in having 24 bits colour >precision except for some applications. We don't come near to having >anything which properly uses 16 bit colour. So sticking to 16 bits makes it >much more memory efficient, which is very importent considering the way the >Pointer Environment works. I disagree that 24bit colour precision is not necessary. Many file formats support that and it would be best if we could be "compatible". Not to mention ported applications ;-) >Joachim >-- >Joachim Van der Auwera >http://www.triathlon98.com/Joachim/
