Re: Efficient image drawing

[Paul Davis <pjdavis engineering uiowa edu>]

Murray Cumming wrote:

> On Thu, 2006-04-27 at 05:20 -0500, Paul Davis wrote:
>  
>
> > Howdy,
> >
> > I'm developing a video playback app and I'm a bit curious to see if 
> > there's a better way to render images to the screen.  As it is now I'm 
> > using the Gtkmm objects and copying images from my non-standard format 
> > in a GdkPixbuf and then drawing to a GtkDrawingArea using 
> > gdk_draw_rgb_image_dithalign.
> >    
> [snip]
>
> This documentation suggests that GdkPixbuf doesn't need to copy the
> data:
> http://developer.gnome.org/doc/API/2.0/gdk-pixbuf/gdk-pixbuf-creating.html#gdk-pixbuf-new-from-data
>  
Definitely noticed this a few hours ago and removed an image copy 
operation.  Not to mention the (assumed) malloc/free calls when each 
pixbuf is created/destroyed.

As it is now, I load a frame, and copy it directly into the pixbuf 
array, expanding to 24bpp as I do.

What I'm mostly wondering about is if its possible to dig deeper into X 
so that I'm bypassing the whole gdk_draw_rgb_image_dithalign call or for 
that matter if any of the gdk_draw_rgb_image_* methods are faster.

> Maybe if you implemented a GdkPixbuf backend for your custom format, you
> could save on some slow memory copying.
>
>  
I have to expand the 8bpp image to 24bpp anyway because of other 
requirements, so the one copy operation is necessary anyway.  But that 
is an idea I hadn't thought of.  I'll take a deeper look into Pixbuf and 
its drawing in a couple hours and hopefully refine my question.