Re: 24-bit-per-pixel limit to Gdk RGB ?

[individual <individual mi cl>]

On Tuesday, February 18, 2003, at 02:09 PM, Michael L Torrie wrote:

> On Tue, 2003-02-18 at 08:08, individual wrote:
>
> > Wait... we must be misunderstanding each other. Last night I realised
> > that for Gtk to be limited to 8-bit colour is absolutely ludicrous! 
> > Gtk
> > is the Gimp Toolkit, and the Gimp is for hight quality images!
>
> No.  Gimp is limited to 24-bit color.  Film gimp, on the other hand,
> uses 16-bit per channel (I think)--48-bit, just like what you're trying
> to do--making it more suitable for work on film.  See
> http://filmgimp.sourceforge.net.  However, on XFree86, even Film Gimp
> will still have to render to a 24-bpp display.

That's interesting. I was always under the imprsesion that we have left 
24bpp displays far behind. Here's what my system says:

Resolution        Video Memory
                                   8MB
                             (SGRAM)

640 x 480               32-bit
800 x 600             32-bit
1024 x 768          32-bit


> > Ok, last attempt: I made a 1000 pixel wide gradient going from 255 red
> > to 254 red. In the colour selection panel, I can clearly see the
> > difference between the two colours. When I draw the gradient, there is
> > no dividing line. HA! I have therefore proved that The Gimp and my X
> > server  are capable of displaying colour at more than 8 bits per
> > channel. Or have I?
>
> Now you have 2 bands of color, each about 500 pixels wide.  So you'll
> definitely see it.  The less colors you have and the wider the bands,
> the more noticable it is.

I didn't see the difference on screen, though I clearly saw the 
difference in the foreground/background colours in the tool palette. 
Someone else explained that this because of dithering.

> Nope you haven't proven anything.  You're still drawing 8-bits per
> channel.  Do an 'xdpyinfo' and you'll see the visual is 24 bpp.

Yes, as a matter of fact the xserver does report a display depth of 24.


>  Nothing
> you can do will change that.  To display a 48-bpp image ona 24bpp
> screen, you need to chop the least significant bits off each channel
> (shift right) until each channel is 8-bits.  Then you can render it to 
> a
> 24 bpp screen and get a decent approximation.


Eek.... I was hoping for greater bit depth, but I'm getting more and 
more convinced that it's pointless to try and render a 48 bpp image at 
48 bpp screen depth when screens are at most 24bpp deep.


Thanks for the info!


Paul