Subject: Mpeg movies

Jon,

Try,

	oas% cd ~shuki

	oas% mpeg_play dmotdmo.mpg

	oas% mpeg_play dmotdmo_close_up.mpg

If the files were ftp-ed okay,
and if your mpeg_play is compatible with
mine you'll see movies.
(Maybe you want to check it before you forward
this message to msgs.)

The media, in this case,
is more important than the content.

But the content is a movie of (DMO^T DMO)**p
where p goes from 0 to 99
(not always in increment of 1).
At the SEG you mentioned a similar movie
with DMO replaced by velocity stack,
and you said it turned into all noise for p>6.
I won't tell you what you'll see in the DMO movies.
But a hint is that I went back
and fixed something in my DMO
and what you have is after that fix
(before there was nothing after frame number 6),
and I know about one more thing I yet need to fix.

My observations are:

1.  Mpeg is a good media for visualizing seismic data.
    One nice thing is that you can get annotations,
    wiggles, variable area, overlays of well data
    on the seismic, and leverage all the plotting
    programs you already have
    without having to duplicate all that
    in your movie program.

    It is extremely nice that we don't have to write
    movie programs ourselves but can leverage
    the healthy entertainment industry.

    The mpeg variable intensity movie is about
    a factor of 10 compressed compared to our movie
    (really byte) files.
    However, this is a lossy compression and you
    can see hash marks artifacts.


2.  Preparing mpeg from seismic data needs improvements.
    I used window dumps of each frame then xwdtopnm,
	    ppm2cyuv, and mpeg.  Found it arduous.

    It would be extremely nice if one of your students will
    make a Seplib program that will input a seplib cube
    and output an mpeg movie in batch.  Just like Taplot
    if you still call it by this name.


3.  We have been mixing two programs:

	movie -
		that is what mpeg_play is evolving to.

	cube_view -
		can cut a cube in various directions,
		give you chair diagrams.  Should be able
		to make some cells transparent.  Maybe have
		horizon picking.
		That is what Landmark's, GeoQuest's and Charisma
		programs are all about.  They are seismic specific.

		Cube_view programs can also make movies but
		are not a replacement for movie programs.

4.  (A^T A)**p movie are good to find bugs.

Shuki