How the videoclip was created
The first video clip 9 MB was created using ffmpeg:
=> ffmpeg -i phdvideo-dvd.mpg -s qntsc -sameq -ab 64 -ar 32000 -ac 1 phdvideo-small.mpg
Output #0, mpeg, to 'phdvideo-small.mpg':
Stream #0.0: Video: mpeg1video, 352x240, 29.97 fps, q=2-31, 200 kb/s
Stream #0.1: Audio: mp2, 32000 Hz, mono, 64 kb/s
phdvideo-small.mpg has variable bit rate, plays okay with mplayer and xine
but has trouble with the vlc player on the audio stream.
It seems that the ffmpeg mplex subroutine does not put correct
SCR/PTS timing information on the output stream (ffmpeg version 0.4.8),
and vlc depends on that information for A/V sync.
The second video clip 22 MB was created using mjpegtools:
=> extract_mpeg2 < phdvideo-dvd.mpg > phdvideo-dvd.mpv
=> ffmpeg -i phdvideo-dvd.mpg -ab 64 -ar 32000 -ac 1 -acodec mp2 phdvideo-audio.mp2
mplayer -benchmark -nojoystick -nolirc -identify -noframedrop -vo yuv4mpeg -v phdvideo-dvd.mpv > foo.log
yuvscaler -v 0 -O VCD -n n < stream.yuv | mpeg2enc -v 0 -f 1 -I 0 -n n -F 4 -a 2 -o phdvideo-vcd.mpv
mplex -v 1 -f 1 -o phdvideo-vcd.mpg phdvideo-vcd.mpv phdvideo-audio.mp2
The original video was done in 1994. During the CFD simulation run (lasted a few days on 20 workstations),
save output computed data as needed. Once finished, X11-graphics display data snapshots as needed for video.
Tried camera on computer screen, not too bad but flicker.
Then, used silicon graphics X11 display that had video capture. Then beta-cam and mini-8
in the AI video lab, editing, added audio. Finito in 1994, used tape for talks,presentations
including Ph.D. thesis defense.
10 years later. found old betacam and mini-8, digitized into miniDV. put into computer with kino,
did a dvd export from kino, did the above mpeg conversions, and voila.
Back to my main science page
Last modified: Thu May 27 15:18:46 EDT 2004