Aloha,

Rune Allnor schrieb:
> The "obvious" idea hit me that it might be an idea to try and
> write my own MPEG encoder that I can tailor to the type of data
> I'm modeling and want to visualize. I am a bit dubious, though,
> as I suspect it may not be as easy or as straight-forward to implement
> such a thing as I want to believe in my current euphoric state.
> Any comments, suggestions or warnings are welcome.

Don't even think about writing it yourself. Except you have the time
to do it...

For an own Matlab visualization project i used the mpeg-encode
(afair U.C.Berkeley) that is linked on the mpeg.org web-page.
That encoder reads a broad range of input picture formats and
is pretty fast. From Matlab i plotted my data and printed every
single frame to a postscript file (which saves a lot of space).
To get the raster image mpeg_encode needs as input it is possible
to define an input filter to transcode (and in this case render
the .ps via ghostscript) the single frames. 

A large part of the project was to fine-tune the parameters of
the mpeg_encoder to get a reasonable bitrate and picture quality.

Wishing a happy day
		LOBI

p.kootsookos@remove.ieee.org (Peter J. Kootsookos) wrote in message news:<s681xplkfs3.fsf@mango.itee.uq.edu.au>...
> allnor@tele.ntnu.no (Rune Allnor) writes:
> 
> > The "obvious" idea hit me that it might be an idea to try and 
> > write my own MPEG encoder that I can tailor to the type of data 
> > I'm modeling and want to visualize. I am a bit dubious, though, 
> > as I suspect it may not be as easy or as straight-forward to implement 
> > such a thing as I want to believe in my current euphoric state.
> 
> Don't do it!
> 
> There used to exist a MPEG encoder/decoder pair that took single
> frames of .ppm images (I think) and compressed them into MPEG.
> 
> I think I took this from mpeg.org or some related site.
> 
> For the application I had, I needed it to read in .jpg files and
> compress them... it was easy enough to modify the code.
> 
> The other possibility is to use AVI files instead.  Matlab's AVIFILE
> command gives examples of how to do this... but it may not do what
> you're after.

Do you mean an avi -> MPEG converter? That could work. Matlab 
allows, to the best of my knowledge, to make the avi file on a 
frame by frame basis. That way one only needs to deal with one 
frame at the time in RAM. It's not perfect but way better than 
dealing with matlab movies.

Do you have a lonk to such a converter?

Rune

allnor@tele.ntnu.no (Rune Allnor) writes:

> The "obvious" idea hit me that it might be an idea to try and 
> write my own MPEG encoder that I can tailor to the type of data 
> I'm modeling and want to visualize. I am a bit dubious, though, 
> as I suspect it may not be as easy or as straight-forward to implement 
> such a thing as I want to believe in my current euphoric state.

Don't do it!

There used to exist a MPEG encoder/decoder pair that took single
frames of .ppm images (I think) and compressed them into MPEG.

I think I took this from mpeg.org or some related site.

For the application I had, I needed it to read in .jpg files and
compress them... it was easy enough to modify the code.

The other possibility is to use AVI files instead.  Matlab's AVIFILE
command gives examples of how to do this... but it may not do what
you're after.

ciao,

Peter K.

-- 
Peter J. Kootsookos

"I will ignore all ideas for new works [..], the invention of which
 has reached its limits and for whose improvement I see no further
 hope."

- Julius Frontinus, c. AD 84

Hi all.

I have just spent a few days making some small MPEG movies that 
illustrate certain wave propagation phenomena. First of all, 
while I have made movies for my own use for the last 7 or 8 years, 
being able to actually make MPEG movies from time-domain simulated
data has opened up a whole new world to me. All of a sudden I can 
send a small (< 10 MB) movie to just about anyone, and have them 
play the thing on standard software on a standard PC without 
obese requirements to either SW or HW. Even more, they can actually
see for themselves all these more or less obscure waves I talk so 
much about. Which means that all of a sudden I find that people 
think that what I have to say makes some sort of sense. Which is 
a very nice experience.

Now, this revelation has opened some questions to me. First, the MPEG 
generator I got hold of, works via matlab. The generator takes a matlab 
movie as input and spits out an MPEG file. While it works, it is first 
of all slow, and second, it relies on this ridicilously obese matlab 
movie format as intermediate stage between raw data and MPEG movie.
Going via matlab movie does impose some limits that I suspect might 
prevent me form making some movies I would find useful, if I could 
produce them.  

The "obvious" idea hit me that it might be an idea to try and 
write my own MPEG encoder that I can tailor to the type of data 
I'm modeling and want to visualize. I am a bit dubious, though, 
as I suspect it may not be as easy or as straight-forward to implement 
such a thing as I want to believe in my current euphoric state.

Any comments, suggestions or warnings are welcome.

Rune