[mythtv-users] scheduling not using all three tuners?
schachte at csse.unimelb.edu.au
Wed Apr 12 14:50:29 UTC 2006
On Wed, Apr 12, 2006 at 09:53:42AM +1000, Max Barry wrote:
> A global soft buffer that does everything you want, but also
> unnecessarily pads shows from the two reliable TV networks (ABC & SBS),
> is still pretty great!
Agreed, as long as it's controlable.
OK, right now Myth has two padding features: global pre- and
post-roll that is sometimes ignored, and per-recording (and
overridable) start early and end late settings that are always
respected. It sounds like your fix is to give users some global
control over when the pre- and post-roll are ignored. But there's a
lot to be said for the direct control of start early and end late
settings. For one thing, they can be explicitly overriden. That's
how I resolve conflicts now, and I like being able to decide for one
program to risk not recording late, but for another to decide to
simply not record a conflicting program rather than risk a WAF crisis
if the end of a favourite program is missed.
So perhaps the nicest/simplest solution would be to:
1) Allow configuration of global default start early and end
late values. Newly scheduled recordings would get these values
by default, but of course they would be changable.
2) In the "upcoming recordins" list, highlight programs that are
conflicting only because of start early and end late settings
in some way that distinguishes then from conflicts of the
programs themselves. Provide a key binding and menu item that
will immediately override the schedule removing enough of the
start early and end late values so the recordings do not
conflict. Also have a different highlight for programs that
are scheduled to record, but have had their start early/end
late settings overridden, and have a key/menu item to remove
those overrides. This would make it easy to see the state and
to toggle the override on and off.
3) Provide a global configuration setting that would allow
automatic (no user intervention) overriding of start early and
end late values with the same options you provided in your
earlier patch, but implemented by automatically generating
program overrides as needed.
Each of these changes is independent of the others, and all of them
leave users in full control over each program's start early and end
late settings. They all just make it a bit more convenient to do what
they can already do.
> > And when consecutive programs are to be recorded on
> > the same channel and the same tuner, the pre- and post-roll can be
> > ignored, and the start early and end late values shouldn't prevent
> > consecutive programs from being recorded on the same tuner (though
> > ideally the overlapping end late/start early time would be included in
> > both recordings).
> But if you're saying that MythTV should prefer to drop the soft buffers
> in order to record consecutive shows on the same channel *even if
> another tuner card is idle*, I would disagree.
No, what I mean is that there's no reason that the signal coming in
from a tuner can't be recorded into *two* files at once. This should
allow you to happily record two programs in a row on the same tuner
and still respect the end late setting for the first program and the
start early setting for the second, as long as both programs are on
the same channel. So this is another improvement, also independent of
4) Allow myth to record program data into two (or more) files at
once. When the start early time of one program arrives while
another is already recording on the same channel and tuner,
open the recording file for the second and start writing every
byte from the tuner into both files. When the end late time
of the first program arrives, close the first file and
continue writing only to the second file. Then modify the
scheduler to know that multiple programs at the same time on
the same channel and tuner are not in conflict.
I believe there's already a TRAC number for exactly this idea.
With all of these changes together, I think Myth would handle
unreliable TV station schedules about as well as it possibly could.
> You a coder, Peter? :)
Yeah, though my C++ is pretty rusty, and was never great. I started
looking into implementing item #4 recently, but haven't gotten very
far. I'm happy to work on any of these ideas, with a little coaching
on where to look.
Peter Schachte In the post-Cold War era, the United States
schachte at cs.mu.OZ.AU needs to promote the development of democracy
www.cs.mu.oz.au/~schachte/ and human rights, not militaries that view their
Phone: +61 3 8344 1338 own citizens as the enemy. -- Senator Richard
More information about the mythtv-users