[directfb-dev] Re: BGR16 and BGR15 colourspace

[armin.gerritsen@philips.com]

Hello,

Sorry for the late response.

> this sounds interesting. Is it related to a project at Philips?
> I haven't heard much about the Philips MHP implementation for quite
> some time. Are these embedded platforms in any way related to it?

No, I'm afraid not (Philips is a huge company :-) ). This is for prototyping
some mobile audio/video platforms.

> Do you have a hybrid endian system, i.e. RGB16 but byte swapped?
> Or do you really have BGR16 as in BBBB BGGG GGGR RRRR ?

Actually it is worse; It is both, but I didn't want to complicate my mail,
since the solution is essentially the same.
I have BGR15, BGR16 and BGR16 with mixed endian depending on which display
device I hook on.

> Adding another pixel format is quite simple. You implement ~13 tiny
> span functions in "src/gfx/generic/generic.c" and all possible
> combinations of blitting flags, destination/source formats etc. are
> supported. Apart from
> the DFBSurfacePixelFormat enum, there are some places where the pixel
> format has to be added (mostly switch statements). Searching all
> source files for RGB16 should be sufficient.

Ok, that's what I did. Unfortunately, our test-system decided to fail right
during testing, so I have to wait to get it repaired to give the final
results ... ;-(

 I only have a question about the file src/gfx/convert.c

DFBSurfacePixelFormat
dfb_pixelformat_for_depth( int depth )
{
     switch (depth) {
          case 8:
               return DSPF_LUT8;
          case 15:
               return DSPF_ARGB1555;
          case 16:
               return DSPF_RGB16;
          case 24:
               return DSPF_RGB24;
          case 32:
               return DSPF_RGB32;
     }

     return DSPF_UNKNOWN;
}

I now just replaced the RGB16/ARGB1555 by BGR16/BGR15, but that doesn't make
me feel too well, since it makes usage of RGB16/ARGB1555 with the same
binary/code impossible. (For instance in fbdev.c the function is being used
when initialising the primairy layer.) I may want to come up with a more
flexible way.

Regards,

Armin




-- 
Info:  To unsubscribe send a mail to ecartis@directfb.org with 
"unsubscribe directfb-dev" as subject.