> Hi! Loic,
> Thanx for the reply.
> I think I got the concept.
> 1.Is this the double buffering???
Yes indeed. Triple would be to have one more buffer:
1 displayed
1 waiting to be displayed
1 in which you render
> 2.I guess can be achieved by a frame buffer to framebuffer copy using the BitBLT
> engine???
> Am I right??
Nope. Thats the whole trick: you do not copy anything. You change one
(or two or three) register(s) and you are done. No bitblitting, no copy.
Fast and efficient. This works only for fullscreen because you need to render
the whole screen.
You can also copy the second buffer in the image, if you are not fullscreen,
using bitblitting, but that's not the same concept (and it's usually far
slower).
> 3.Will this help me in achieving a frame rate of 30fps to play a software decoded
> video.
This highly depends on your hardware. You did not say anything about your
hardware: processor, memory, FSB (of processor and memory). If your processor
can produce 30 frames of RGB images per second, double buffering is THE way
to go if you can afford it: you hardly slow the rendering while doing so. If
you copy the image from offscreen memory to onscreen memory, you slow things
down (usually). If you render your images is system memory, then copy them
to offscreen video memory then blit them onscreen, you make a lot of copies.
(and one more if you do not use shared memory extension). Now if your system
is fast enough, it can render 30fps with many image copies in between. You
should try to see how many frames per second you get with a simple approach:
render in a pixmap and display it.
Lo�c
_______________________________________________
Devel mailing list
[EMAIL PROTECTED]
http://XFree86.Org/mailman/listinfo/devel