MATLAB Central Newsreader - Creating a movie with varying time steps

Creating a movie with varying time steps

Alexander Erlich — Thu, 11 Jun 2009 13:56:40 -0400

Hi there,

suppose I have a solution of a differential equation which I would
like to visualize with surf. The numerical solution tells me exactly
when to plot the object in which coordinates. Now, I would like to
create a movie out of these surf plots. I cannot use getframe, as it
does not enable me to consider the varying time steps. Is there
another way to do that?

Thanks in advance!

Alexander

Re: Creating a movie with varying time steps

Steven Lord — Thu, 11 Jun 2009 14:14:57 -0400

"Alexander Erlich" <alexander.erlich@gmail.com> wrote in message
news:54d79ffd-aeb6-4244-ac04-445c12c94c06@z7g2000vbh.googlegroups.com...
> Hi there,
>
> suppose I have a solution of a differential equation which I would
> like to visualize with surf. The numerical solution tells me exactly
> when to plot the object in which coordinates. Now, I would like to
> create a movie out of these surf plots. I cannot use getframe, as it
> does not enable me to consider the varying time steps. Is there
> another way to do that?

Solve the system of ODEs to generate a solution struct array (using the "sol
= ..." syntax described in the help rather than the "[t, y] = ..." syntax.)
Once you have the solution struct, use DEVAL to evaluate the solution at
equally-spaced time steps. Then create the movie from the surfaces created
by the equally-timestep-spaced data.

--
Steve Lord
slord@mathworks.com

Re: Creating a movie with varying time steps

Alexander — Thu, 11 Jun 2009 18:11:17 -0400

Thank you Steve,

I will try that. But I still have a question: Suppose I had equally-
spaced time steps. Getframe won't know how long one time step is, will
it? It might play the movie much faster than it really (=as given by
the solution of the ODE) is, or slower, probably depending on my
computer. If I'm right with that problem, what could I do?

Alexander

On Jun 11, 4:14 pm, "Steven Lord" <sl...@mathworks.com> wrote:
> "Alexander Erlich" <alexander.erl...@gmail.com> wrote in message
>
> news:54d79ffd-aeb6-4244-ac04-445c12c94c06@z7g2000vbh.googlegroups.com...
>
> > Hi there,
>
> > suppose I have a solution of a differential equation which I would
> > like to visualize with surf. The numerical solution tells me exactly
> > when to plot the object in which coordinates. Now, I would like to
> > create a movie out of these surf plots. I cannot use getframe, as it
> > does not enable me to consider the varying time steps. Is there
> > another way to do that?
>
> Solve the system of ODEs to generate a solution struct array (using the "sol
> = ..." syntax described in the help rather than the "[t, y] = ..." syntax.)
> Once you have the solution struct, use DEVAL to evaluate the solution at
> equally-spaced time steps. Then create the movie from the surfaces created
> by the equally-timestep-spaced data.
>
> --
> Steve Lord
> sl...@mathworks.com

Re: Creating a movie with varying time steps

Steven Lord — Thu, 11 Jun 2009 21:35:03 -0400

"Alexander" <alexander.erlich@gmail.com> wrote in message
news:749aa82c-1702-4c71-b5c2-209488908cee@h11g2000yqb.googlegroups.com...
> Thank you Steve,
>
> I will try that. But I still have a question: Suppose I had equally-
> spaced time steps. Getframe won't know how long one time step is, will
> it? It might play the movie much faster than it really (=as given by
> the solution of the ODE) is, or slower, probably depending on my
> computer. If I'm right with that problem, what could I do?

You're correct that GETFRAME doesn't know how long one time step is.
Technically, MOVIE and AVIFILE don't know how long one time step is either.
But both MOVIE and AVIFILE know how many frames to show per second, and
that's just as good, isn't it?

http://www.mathworks.com/access/helpdesk/help/techdoc/ref/movie.html

http://www.mathworks.com/access/helpdesk/help/techdoc/ref/avifile.html

--
Steve Lord
slord@mathworks.com

Re: Creating a movie with varying time steps

Alexander — Fri, 12 Jun 2009 14:31:40 -0400

Thanks a lot Steve,

it makes sense :-)

It would have taken me ages to figure this out myself, I didn't know
about deval! Now, with a movie of let's say 20fps, I can use deval to
obtain intervals of 1/20th of a second and render the movie! This
method will probably enable me to make my computer render a movie of
60s for three hours or so, hopefully making it both correct and
optically refined!

Thanks!

Alexander

On Jun 11, 11:35 pm, "Steven Lord" <sl...@mathworks.com> wrote:
> "Alexander" <alexander.erl...@gmail.com> wrote in message
>
> news:749aa82c-1702-4c71-b5c2-209488908cee@h11g2000yqb.googlegroups.com...
>
> > Thank you Steve,
>
> > I will try that. But I still have a question: Suppose I had equally-
> > spaced time steps. Getframe won't know how long one time step is, will
> > it? It might play the movie much faster than it really (=as given by
> > the solution of the ODE) is, or slower, probably depending on my
> > computer. If I'm right with that problem, what could I do?
>
> You're correct that GETFRAME doesn't know how long one time step is.
> Technically, MOVIE and AVIFILE don't know how long one time step is either.
> But both MOVIE and AVIFILE know how many frames to show per second, and
> that's just as good, isn't it?
>
> http://www.mathworks.com/access/helpdesk/help/techdoc/ref/movie.html
>
> http://www.mathworks.com/access/helpdesk/help/techdoc/ref/avifile.html
>
> --
> Steve Lord
> sl...@mathworks.com