On Thu 15.07.2004 at 05:38:01PM +0800, Mick wrote:
> > On Wed 14.07.2004 at 06:22:33PM +0200, Dirk Meyer wrote:
> > > > Or extend Thomas' work so that one can simulate one variable-length
> > > > buffer using several fixed-length buffers, which can be linked
> > > > together using fields in the header of each buffer. Mmmmhhh, I have
> > > > to think about this :)
> > >
> > > Maybe use the other ringbuffer patch from Jason Tackaberry?
> >
> > The problem with ringbuffers is, that they limit the amount of time you
> > can pause the TV. And if you want to have a large rungbuffer, it will
> > take lots of space all the time. The idea you proposed some time ago was
> > imho better: write the data into several small files of fixed size, and
> > remove the files which have already been read. In this case you have
> > timsehift ability with variable shift length (so if your mother calls
> > and speaks for 2 hours, you will not miss half of the show, unlike in
> > the ringbuffer scenario, where the buffer is limited to e.g. 20 min).
> > And you won't use an awful lot of disk space if you don't pause the show
> > at all (remember: 1 hour = 2GB of data in some cases).
> >
> > > > What I could do also is hack the seek function so that you can
> > > > never seek after the end of the file (e.g. checking eof when
> > > > seeking, and doing nothing if a jump would lead us after the end
> > > > of the movie). This means that you cannot anymore quit mplayer
> > > > by just fast-forwarding (but reaching the EOF while playing
> > > > should work OK). I could also make this feature an option of
> > > > mplayer, so that it would be activated only when timeshifting
> > > > with a growing file (i.e. not using Thomas' timeshift buffer),
> > > > using MPEG PS and TS files.
> > >
> > > That would be cool. An option is what would be best. It avoids
> > > some problems we have right now.
>
> The *real* issue to deal with is how to record something in either
> case..
>
> It would not be wise to make something timeshiftable with no ability to
> record it as well..
This seems pretty obvious to me: for simple recording, use e.g.
{t,s,c}zap -r and cat vdr0 > file.mpeg. For timeshifting, I'm working on
a modified version of tzap which can write to a normal file, to Thomas'
timeshift fixed-length buffer, or to the new kind of buffer I described
earlier, and also remux the TS to a PS and rewrite the PTS so that they
start close to 0. The glue between the recoder process and the player
process can be provided by Freevo, by an external program (script) or by
the recorder process itself.
Besides, if the player doesn't behave properly, there is nothing to
watch what you record...
Matthieu
--
(~._.~) Matthieu Weber - Universit� de Jyv�skyl� (~._.~)
( ? ) email : [EMAIL PROTECTED] ( ? )
()- -() public key id : 452AE0AD ()- -()
(_)-(_) "Humor ist, wenn man trotzdem lacht (Germain Muller)" (_)-(_)
-------------------------------------------------------
This SF.Net email is sponsored by BEA Weblogic Workshop
FREE Java Enterprise J2EE developer tools!
Get your free copy of BEA WebLogic Workshop 8.1 today.
http://ads.osdn.com/?ad_idG21&alloc_id040&op=click
_______________________________________________
Freevo-users mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/freevo-users
