Gidday there, I have an application that requires me to take still images (nominally jpegs), and pop them into some YUV buffer of a video card so that another piece of software can make funky overlays. Currently the other software expects a video4linux device making use of the XVideo API. I'm a newbie to both of these APIs, and I would like to know whether it would be easier/quicker in one or the other. Unfortunately issues about portability and doing things the Right-Way(tm) are secondary to my horrendous time constraints. Any hints? Clue sticks? I've just downloaded the 40Megs of mailing list archive, which I'm slowly going through. I know absolutely nothing about XVideo, but for v4l, I am guessing that I would base the driver on some sort of USB device which receives individual frames. Then I would figure out some sort of FIFO to userspace in order to allow any sort of networked program to pipe data in. Any help is appreciated. Joshua