[MPlayer-users] using mplayer in DRT/Video reverse engineering tool

[Amir Michail]

Hi,

We would like to use mplayer in the DRT/Video tool for design
recovery/reverse engineering of real-time graphical applications such
as arcade games and flight simulators.

To do this, we need to be able to do several things:

* generate a video from individual image files; although mencoder can
do this, you need to specify the number of frames / second.  For our
purposes, the file timestamps indicate when that image should occur in
the video.  Since file timestamps have second granularity,
linear interpolation is performed to get a more accurate time of when the
frame should be shown.  (Basically, we use xvidcap to capture the
video frames of the application while it is demonstrated.
Unfortunately, the xvidcap capture rate is not very precise,
particularly if the application is CPU intensive.  Moreover, we need
to know when the frame was taken to match it with code instrumentation
data.)

* we need to be able to zoom in and out of various parts of the video
and annotate/highlight various regions of the video; what is the best
way to do this efficiently as the video is playing in real-time?  We
cannot anticipate the zooming/annotation ahead of time since that
depends on what the user wants to do with DRT/Video.

* we need to be able to extract an individual frame at time t from a
video file 

Are these things easy to do?  If not, would it be easy to modify
mplayer/mencoder to do them?  

Amir Michail
UNSW DRT/Video Group

P.S.  You can find out more about DRT/Video here:

http://www.cse.unsw.edu.au/~drt/screenshot
http://www.cse.unsw.edu.au/~amichail/drtvideo.pdf