videolanclient and iPAQ and Opie

[Jean-Paul Saman]

Jean-Paul Saman wrote:
> 
> Christophe Massiot wrote:
> >
> > Hello,
> >
> > À (At) 23:14 +0200 26/04/2002, Jean-Paul Saman écrivait (wrote) :
> >
> > >  > I'm interested in the status of videolan for Opie. Is there
> > >already some work
> > >  > done? If so, how much needs still to be done?
> >
> > I don't know what Opie is, but I understand it is some kind of
> > embedded system. I have first started the port on iPAQ familiar Linux
> > a few months ago, but since then I have quit my company and I can no
> > longer work on it, so that the ipkg isn't maintained any more. I hope
> > it still works on familiar, has anyone tested it recently ?
> >
> Opie is based on qpe and uses QT-embedded from Trolltech. It stands for
> Open Palmtop Integration Environment. Opie, Qtopia (qpe), Intimate and
> familiar are the four most favourite distro's for the iPaq. Opie and
> Qtopia both use familiar as there base, but without X11. Compaq Research
> is busy with getting Opie and Qtopia on top of X11.
> Opie (and Qtopia) are better equiped as Personal Information Manager
> then a complete familiar is at the moment. They are clearly ahead with
> respect to functionality. I think it is worth to get vlc working on that
> (and as well as on familiar).
> 
> Last week I had a demo with iPaq and vlc-0.3.1. I used as streaming
> server vlms and that worked fantastic. The audio did not work perfectly
> it sounded like audio underruns, so I guessed we still had to do the
> video part. But as you write below that a lot of work has already be
> done. I guess an extra look at the audio part can't hurt.
> 
> So I think, I'll have to look at the audio again and do some testing.
> 
Just did some testing with vlc-0.3.0 and indeed is a buffer underrun. I
changed in include/config.h the define AOUT_BUFFER_DURATION and now the
audio is smooth. No more work to be done here for audio, just make sure
when compiling for ipaq AOUT_BUFFER_DURATION will be (50000*2) selected.

However a good user interface is still a nice to have ;-)

> A pitty that you do not have as much time for that kind of work as you
> did have. I'll try to spend some more time on this and get an up to date
> ipk out. Do you have some patches for SDL on iPaq? Or is the libsdl.org
> good enough.
> 
> > >- video decoder should be integer based, this probably needs a rewrite
> > >(needs guru)
> >
> > The video decoder is already integer-based. A floating point
> > implementation would be far too slow, and video IDCT doesn't need to
> > be as accurate as audio... I even think we got rid of all floats in
> > VLC, except in the native MPEG and ac3 decoder. What were you
> > thinking of ?
> >
> So the complete video path is not integer based yet?
> If that is so, then I think we should get the native MPEG and ac3
> decoder integer based. Is that possible?
> 
> A vout_qte based on purely QT-embedded and a qte native UI too would be
> nice. I have some ideas for the interface and only need time to do it.
> Some help can not hurt either ;-)
> 
> > --
> > Christophe Massiot.
> >
> > --
> > This is the vlc-devel mailing-list, see http://www.videolan.org/vlc/
> > To unsubscribe, please read http://www.videolan.org/lists.html
> > If you are in trouble, please contact <postmaster at videolan.org>
> 
> --
> This is the vlc-devel mailing-list, see http://www.videolan.org/vlc/
> To unsubscribe, please read http://www.videolan.org/lists.html
> If you are in trouble, please contact <postmaster at videolan.org>

-- 
This is the vlc-devel mailing-list, see http://www.videolan.org/vlc/
To unsubscribe, please read http://www.videolan.org/lists.html
If you are in trouble, please contact <postmaster at videolan.org>