pseudo color and direct color

[Brian Paul]

Jon Smirl wrote:
> On 5/17/05, Brian Paul <brian.paul at tungstengraphics.com> wrote:
> 
>>Jon Smirl wrote:
>>
>>>On 5/16/05, Brian Paul <brian.paul at tungstengraphics.com> wrote:
>>>
>>>
>>>>Jon Smirl wrote:
>>>>
>>>>
>>>>>Pseudo color and direct color - neither of these are support in EGL, right?
>>>>>Everything is true color?
>>>>>
>>>>>8,16,24,32 true color are all allowed, just no palettes.
>>>>
>>>>EGL, as-is, has no notion of color index mode (PseudoColor).  It could
>>>>be easily added with a new extension if we need/want it.
>>>
>>>
>>>What about DirectColor? Is that grouped with PseudoColor?
>>
>>DirectColor requires lookup tables for each of the R, G, B channels.
>>We'd have to add functions for specifying those tables.  Similarly for
>>PseudoColor.
> 
> 
> Is there a need for DirectColor support, or do everyone just use true color?
> Does current hardware support DirectColor?

I've seen very few applications that need/use DirectColor.  I think 
it's still supported by most modern cards though.

With a GL/compositing window system, the DirectColor-style lookups can 
be done with OpenGL commands.

Similarly, color indexed windows can be treated as color index / 
paletted textures and converted to RGB on the fly.

-Brian