Mailing List Archive

Your summary format requires a bit of guessing as to what you mean. It
***might*** mean that show 3 did not record because there was insufficient
time between the end of show 2 and the start of show 3 for Myth to start a
new capture. But this really depends on what you mean by "Time" in your
chart ... it might mean start time or length, and we really need to see both.

So please repost this report, this time replacing "Time" with two columns:

Start Time
Length in Minutes (or End Time if that is easier)

Also please tell us if you have the "record past end" option set, and if
you do, for how many minutes.

At 12:25 PM 5/19/2003 -0300, Jeff C wrote:
>Hello all,
>
>Here is some more data on the tagged but not recorded problem I'm having.
>
>I am running sunday afternoon CVS code.
>
>I set it to record four shows last night as a test:
>
>Local Channel/Network Time Show Status
>14/TNN 00:30 C&D TV Recorded ok
>55/Prime 02:00 Entertain Tnght Recorded ok
>40/Life Network 03:00 Kink NOT RECORD & Still tagged
>in guide view
>46/History Network 04:00 Around World Recorded ok
>
>I am using a PVR 250 card. Unfortunately I lost the -v console output due
>to an untimely reboot his morning but I'll try things again tonight and
>see if I can get the -v output back to you.

On Monday 19 May 2003 11:49 am, Ray Olszewski wrote:
> Your summary format requires a bit of guessing as to what you mean. It
> ***might*** mean that show 3 did not record because there was insufficient
> time between the end of show 2 and the start of show 3 for Myth to start a
> new capture. But this really depends on what you mean by "Time" in your
> chart ... it might mean start time or length, and we really need to see
> both.
>
> So please repost this report, this time replacing "Time" with two columns:
>
> Start Time
> Length in Minutes (or End Time if that is easier)
>
> Also please tell us if you have the "record past end" option set, and if
> you do, for how many minutes.

And it's all pretty useless without all the information that the backend
prints out. =)

Isaac

At 08:49 AM 5/19/2003 -0700, you wrote:
>Your summary format requires a bit of guessing as to what you mean. It
>***might*** mean that show 3 did not record because there was insufficient
>time between the end of show 2 and the start of show 3 for Myth to start a
>new capture. But this really depends on what you mean by "Time" in your
>chart ... it might mean start time or length, and we really need to see both.
>
>So please repost this report, this time replacing "Time" with two columns:
>
> Start Time
> Length in Minutes (or End Time if that is easier)
>
>Also please tell us if you have the "record past end" option set, and if
>you do, for how many minutes.

This option is not set - my NTP set time appears to be in synch with my
cable compay.

>At 12:25 PM 5/19/2003 -0300, Jeff C wrote:
>>Hello all,
>>
>>Here is some more data on the tagged but not recorded problem I'm having.
>>
>>I am running sunday afternoon CVS code.
>>
>>I set it to record four shows last night as a test:

Network Start/Length Title Status
TNN 00:30/30m C&D TV Recorded OK
Prime 02:00/60m Entertain Tnght Recorded OK
Life 03:00/30m Kink Did not record
History 04:00/60m Around t/World Recorded OK

There were no conflicts displayed in the scheduling conflicts screen and
oddly, when I scroll back in the guide today, the show that did not record
still appears as Orange (tagged).

-Jeff

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

> Local Channel/Network Time Show Status
> 14/TNN 00:30 C&D TV Recorded ok
> 55/Prime 02:00 Entertain Tnght Recorded ok
> 40/Life Network 03:00 Kink NOT
> RECORD & Still tagged in guide view
> 46/History Network 04:00 Around World Recorded ok

Where the last three programs consecutive, or were there gaps in
between because the programs < 60 minutes?

-----BEGIN PGP SIGNATURE-----
Version: PGPfreeware 7.0.3 for non-commercial use <http://www.pgp.com>

iQA/AwUBPskBhfc1NpCTlP0JEQKVHQCgxd3KlocxXVlkLYfR452gKyI+lHoAn182
i4cbFqaakYjg7CCCSx62nPKR
=Qm6j
-----END PGP SIGNATURE-----

Well ... Isaac is really the expert here, of course, but it appears to me
that the problem is what I guessed at first ... that the time between
stopping the second show and starting the third is insufficient to let the
third show start recording.

I just set my Myth host (running latest .deb binaries, not CVS) up to test
this sort of timing problem here, but of necessity, the test takes 2 hours
to complete. I'll get back to the list later with its results. I did try
something similar last week, but it involved capturing several consecutive
shows on the **same** channel, not as demanding a task since it does not
require that the vidcap card be reset ... just that the capture output
switch to a new file.

Do allow for the possibility that the latency in stopping capture depends
on the hardware. I'm using an AverTV card here, and with other capture
software (var/avifile), I always had to leave at least a minute, better
two, between stop and start of capture. This was partly (though I think not
entirely) due to the limited time resolution of crond ... I expect Myth
uses an internal timer that is capable of timing to the second, not just
the minute. But still, your PVR 250 and its driver may introduce latency
that Myth does not allow for.

If I see the same problem in my test, though ... then I wonder if Myth
needs to refine its calculations of when two captures conflict?

At 01:04 PM 5/19/2003 -0300, Jeff C wrote:
>At 08:49 AM 5/19/2003 -0700, you wrote:
>>Your summary format requires a bit of guessing as to what you mean. It
>>***might*** mean that show 3 did not record because there was
>>insufficient time between the end of show 2 and the start of show 3 for
>>Myth to start a new capture. But this really depends on what you mean by
>>"Time" in your chart ... it might mean start time or length, and we
>>really need to see both.
>>
>>So please repost this report, this time replacing "Time" with two columns:
>>
>> Start Time
>> Length in Minutes (or End Time if that is easier)
>>
>>Also please tell us if you have the "record past end" option set, and if
>>you do, for how many minutes.
>
>This option is not set - my NTP set time appears to be in synch with my
>cable compay.
>
>>At 12:25 PM 5/19/2003 -0300, Jeff C wrote:
>>>Hello all,
>>>
>>>Here is some more data on the tagged but not recorded problem I'm having.
>>>
>>>I am running sunday afternoon CVS code.
>>>
>>>I set it to record four shows last night as a test:
>
>Network Start/Length Title Status
>TNN 00:30/30m C&D TV Recorded OK
>Prime 02:00/60m Entertain Tnght Recorded OK
>Life 03:00/30m Kink Did not record
>History 04:00/60m Around t/World Recorded OK
>
>There were no conflicts displayed in the scheduling conflicts screen and
>oddly, when I scroll back in the guide today, the show that did not record
>still appears as Orange (tagged).

I don't think that this is the problem. In my
experience when mythtv is set to record during a
certain time period, then it will start recording at
the start time or as soon after as possible.

i.e. if you record show X from 10:30 - 11:00 and show
Y from 11:00 - 11:30 with a "record past ending" value
of 2:00 then show Y will essentially record from 11:02
- 11:30.

to make it even more simple. a recording isn't missed
if it can't start right at the begin time. I've done
hundreds of recordings like this and never missed a
single one.

I would guess the problem is something else.

--- Ray Olszewski <ray@comarre.com> wrote:
> Well ... Isaac is really the expert here, of course,
> but it appears to me
> that the problem is what I guessed at first ... that
> the time between
> stopping the second show and starting the third is
> insufficient to let the
> third show start recording.
>
> I just set my Myth host (running latest .deb
> binaries, not CVS) up to test
> this sort of timing problem here, but of necessity,
> the test takes 2 hours
> to complete. I'll get back to the list later with
> its results. I did try
> something similar last week, but it involved
> capturing several consecutive
> shows on the **same** channel, not as demanding a
> task since it does not
> require that the vidcap card be reset ... just that
> the capture output
> switch to a new file.
>
> Do allow for the possibility that the latency in
> stopping capture depends
> on the hardware. I'm using an AverTV card here, and
> with other capture
> software (var/avifile), I always had to leave at
> least a minute, better
> two, between stop and start of capture. This was
> partly (though I think not
> entirely) due to the limited time resolution of
> crond ... I expect Myth
> uses an internal timer that is capable of timing to
> the second, not just
> the minute. But still, your PVR 250 and its driver
> may introduce latency
> that Myth does not allow for.
>
> If I see the same problem in my test, though ...
> then I wonder if Myth
> needs to refine its calculations of when two
> captures conflict?
>
> At 01:04 PM 5/19/2003 -0300, Jeff C wrote:
> >At 08:49 AM 5/19/2003 -0700, you wrote:
> >>Your summary format requires a bit of guessing as
> to what you mean. It
> >>***might*** mean that show 3 did not record
> because there was
> >>insufficient time between the end of show 2 and
> the start of show 3 for
> >>Myth to start a new capture. But this really
> depends on what you mean by
> >>"Time" in your chart ... it might mean start time
> or length, and we
> >>really need to see both.
> >>
> >>So please repost this report, this time replacing
> "Time" with two columns:
> >>
> >> Start Time
> >> Length in Minutes (or End Time if that is
> easier)
> >>
> >>Also please tell us if you have the "record past
> end" option set, and if
> >>you do, for how many minutes.
> >
> >This option is not set - my NTP set time appears to
> be in synch with my
> >cable compay.
> >
> >>At 12:25 PM 5/19/2003 -0300, Jeff C wrote:
> >>>Hello all,
> >>>
> >>>Here is some more data on the tagged but not
> recorded problem I'm having.
> >>>
> >>>I am running sunday afternoon CVS code.
> >>>
> >>>I set it to record four shows last night as a
> test:
> >
> >Network Start/Length Title Status
> >TNN 00:30/30m C&D TV Recorded OK
> >Prime 02:00/60m Entertain Tnght
> Recorded OK
> >Life 03:00/30m Kink Did
> not record
> >History 04:00/60m Around t/World Recorded OK
> >
> >There were no conflicts displayed in the scheduling
> conflicts screen and
> >oddly, when I scroll back in the guide today, the
> show that did not record
> >still appears as Orange (tagged).
>
>
>
> _______________________________________________
> mythtv-users mailing list
> mythtv-users@snowman.net
>
http://lists.snowman.net/cgi-bin/mailman/listinfo/mythtv-users


__________________________________
Do you Yahoo!?
The New Yahoo! Search - Faster. Easier. Bingo.
http://search.yahoo.com

On Mon, May 19, 2003 at 10:21:40AM -0700, Allen T. Gilliland IV wrote:
> I don't think that this is the problem. In my
> experience when mythtv is set to record during a
> certain time period, then it will start recording at
> the start time or as soon after as possible.
>
> i.e. if you record show X from 10:30 - 11:00 and show
> Y from 11:00 - 11:30 with a "record past ending" value
> of 2:00 then show Y will essentially record from 11:02
> - 11:30.

Actually, in that situation, it will ignore the "record past ending"
value, and start recording the second show right at 11:00.

> to make it even more simple. a recording isn't missed
> if it can't start right at the begin time. I've done
> hundreds of recordings like this and never missed a
> single one.

I've actually started the backend 15 minutes after a show was supposed
to start recording, and it started the record automatically. Which
reminds me of the one thing I'd like to see work differently. Myth adds
a program to the database as being recorded as soon as it starts
recording instead of when it ends. In cases like this where I don't get
a whole show, or if the backend dies, that show is marked as recorded,
and until I remove it from the DB, it will never record that episode
again. (which in my opinion is the wrong behavior. It never completed
the episode, so it should still be recorded at a later date)

Rob

At 01:45 PM 5/19/2003 -0400, Rob Rogers wrote:
>On Mon, May 19, 2003 at 10:21:40AM -0700, Allen T. Gilliland IV wrote:
> > I don't think that this is the problem. In my
> > experience when mythtv is set to record during a
> > certain time period, then it will start recording at
> > the start time or as soon after as possible.

I can now confirm, based on a test this morning ... at least for a system
with 0 seconds "past end set" ... Myth records two consecutive shows on
different channels just fine.

OTOH, simply ignoring the fact that the one show out of 4 that fails is the
one with this "quick-change" property seems to me unwise.

The possibility that remains is that there is some special problem
involving the PVR 250 and quick changes. Are any of you who are posting
reports that consecutive tapings work using a PCR 250?

>I can now confirm, based on a test this morning ... at least for a system
>with 0 seconds "past end set" ... Myth records two consecutive shows on
>different channels just fine.

Thanks for running the test.

>OTOH, simply ignoring the fact that the one show out of 4 that fails is
>the one with this "quick-change" property seems to me unwise.
>
>The possibility that remains is that there is some special problem
>involving the PVR 250 and quick changes. Are any of you who are posting
>reports that consecutive tapings work using a PCR 250?

I can't comment on the BT8x8 cards but I know the PVR takes a good 2
seconds or so to switch channels. I really need to get the -v output and
also any feedback from the driver to see if it wigs out if myth tries to
change the channel to fast. I have recorded show back to back on two diff
channels before with success so it might be some kind of buffer overflow or
something... not sure... I'll get more debug info when I run it all again
tonight.

-Jeff