12 bit RGB to 16 bit RGB conversion

[jimst]

Could anybody tell me how to convert 12 bit RGB to 16 bit RGB?
Thanks for the help in advance.

[joe_bruin]

here's code that does pretty much the same thing for 16->32 conversion (taken from directfb). adjust it to do what you want.

#define RGB16_TO_RGB32(pixel)  ( (((pixel) & 0xF800) << 8) | \
                                 (((pixel) & 0x07E0) << 5) | \
                                 (((pixel) & 0x001F) << 3) )

[jimst]

thank you, joe_bruin!
Looking at the code, I guess it'll be

#define RGB12_TO_RGB16(pixel)  ( (((pixel) & 0x0F00) << 4) | \
                                 (((pixel) & 0x00E0) << 3) | \
                                 (((pixel) & 0x001F) << 1) )

[joe_bruin]

not quite (but close).

a 12bit RGB pixel is composed of 4 bits of red, 4 of green, and 4 of blue. if you have a 16bit variable holding the pixel, it's 0000RRRRGGGGBBBB (where every letter is 1 bit).

a 16bit RGB pixel is:
RRRRRGGGGGGBBBBB (5 reds, 6 greens, 5 blue bits).

the reds you have correct. the greens, you're a little off. you're missing one green bit. for the blue, you've captured an extra green bit.

[jimst]

Thank you very much.
But I don't understand about the greens... Do you mean I have to shift more?

[joe_bruin]

no. your shifting is right, your masks are a little off. here's what's being evaluated right now:

#define RGB12_TO_RGB16(pixel)  \
((((pixel) & 0x0F00) << 4) | \
 (((pixel) & 0x00E0) << 3) | \
 (((pixel) & 0x001F) << 1) )

#define RGB12_TO_RGB16(0000RRRRGGGGBBBB)  \
(((0000RRRR00000000) << 4) | \
 ((00000000GGG00000) << 3) | \
 ((00000000000GBBBB) << 1) )

[jimst]

Now I see, thank you very much for good explanation!