Mailing List Archive

On 26/10/12 10:34 PM, Tim Phipps wrote:
> Quoting Mark Greenwood <fatgerman@gmail.com>:
>>
>> Unfortunately it's not working for me Tim (on SD channels), it seems
>> from what I can tell that it exits very quickly so I'm thinking it's
>> probably not doing anything. However it's also not generating any log
>> messages anywhere that I know to look (/var/log/mythtv) so I'm stuck
>> for where to go next.
>
> You can run it from the command line (same user as mythbackend):
>
> mausc-wrapper.py --chanid 1010 --starttime 20121025205800
>
> will operate on the file 1010_20121025205800.mpg in the recordings
> directory and it should print out the breaks that it finds.
>
> Cheers,
> Tim.
>
I'm really keen to give this a go (in Australia), but I appear to be
having some python troubles:

# ./mausc-wrapper 24367
Traceback (most recent call last):
File "./mausc-wrapper", line 106, in <module>
main()
File "./mausc-wrapper", line 97, in main
runjob(jobid=args[0])
File "./mausc-wrapper", line 18, in runjob
db = MythDB()
File "/usr/lib/pymodules/python2.6/MythTV/database.py", line 965, in
__init__
dbconn = xml.getConnectionInfo(pin)
File "/usr/lib/pymodules/python2.6/MythTV/connections.py", line 619,
in getConnectionInfo
dat = self._request('Myth/GetConnectionInfo', Pin=pin).readJSON()
File "/usr/lib/pymodules/python2.6/MythTV/connections.py", line 607,
in _request
for k,v in keyvars.items()])
File "/usr/lib/python2.6/urllib.py", line 1234, in quote
res = map(safe_map.__getitem__, s)
TypeError: argument 2 to map() must support iteration

What version of python are you running (and maybe what distro)? I'm on
mythbuntu lucid and I can't seem to get past this problem.
I've checked that my python2.6 packages are all up to date, as well as
my libmyth-python (2:0.25.2+fixes.20121002.139bd59-0ubuntu0mythbuntu1)

Cheers,
Andrew
_______________________________________________
mythtv-users mailing list
mythtv-users@mythtv.org
http://www.mythtv.org/mailman/listinfo/mythtv-users

On 27 Oct 2012, at 07:56, Andrew Leech wrote:

> On 26/10/12 10:34 PM, Tim Phipps wrote:
>> Quoting Mark Greenwood <fatgerman@gmail.com>:
>>>
>>> Unfortunately it's not working for me Tim (on SD channels), it seems from what I can tell that it exits very quickly so I'm thinking it's probably not doing anything. However it's also not generating any log messages anywhere that I know to look (/var/log/mythtv) so I'm stuck for where to go next.
>>
>> You can run it from the command line (same user as mythbackend):
>>
>> mausc-wrapper.py --chanid 1010 --starttime 20121025205800
>>
>> will operate on the file 1010_20121025205800.mpg in the recordings directory and it should print out the breaks that it finds.
>>
>> Cheers,
>> Tim.
>>
> I'm really keen to give this a go (in Australia), but I appear to be having some python troubles:
>
> # ./mausc-wrapper 24367
> Traceback (most recent call last):
> File "./mausc-wrapper", line 106, in <module>
> main()
> File "./mausc-wrapper", line 97, in main
> runjob(jobid=args[0])
> File "./mausc-wrapper", line 18, in runjob
> db = MythDB()
> File "/usr/lib/pymodules/python2.6/MythTV/database.py", line 965, in __init__
> dbconn = xml.getConnectionInfo(pin)
> File "/usr/lib/pymodules/python2.6/MythTV/connections.py", line 619, in getConnectionInfo
> dat = self._request('Myth/GetConnectionInfo', Pin=pin).readJSON()
> File "/usr/lib/pymodules/python2.6/MythTV/connections.py", line 607, in _request
> for k,v in keyvars.items()])
> File "/usr/lib/python2.6/urllib.py", line 1234, in quote
> res = map(safe_map.__getitem__, s)
> TypeError: argument 2 to map() must support iteration
>
> What version of python are you running (and maybe what distro)? I'm on mythbuntu lucid and I can't seem to get past this problem.
> I've checked that my python2.6 packages are all up to date, as well as my libmyth-python (2:0.25.2+fixes.20121002.139bd59-0ubuntu0mythbuntu1)

I'm using mythbuntu 12.04.1, which has python 2.7.3. I'm no python expert I'm afraid so I don't know what your error means.

Cheers,

Mark

>
> Cheers,
> Andrew
> _______________________________________________
> mythtv-users mailing list
> mythtv-users@mythtv.org
> http://www.mythtv.org/mailman/listinfo/mythtv-users

_______________________________________________
mythtv-users mailing list
mythtv-users@mythtv.org
http://www.mythtv.org/mailman/listinfo/mythtv-users

On 27/10/12 9:24 PM, Mark Greenwood wrote:
> On 27 Oct 2012, at 07:56, Andrew Leech wrote:
>
>> On 26/10/12 10:34 PM, Tim Phipps wrote:
>>> Quoting Mark Greenwood <fatgerman@gmail.com>:
>>>> Unfortunately it's not working for me Tim (on SD channels), it seems from what I can tell that it exits very quickly so I'm thinking it's probably not doing anything. However it's also not generating any log messages anywhere that I know to look (/var/log/mythtv) so I'm stuck for where to go next.
>>> You can run it from the command line (same user as mythbackend):
>>>
>>> mausc-wrapper.py --chanid 1010 --starttime 20121025205800
>>>
>>> will operate on the file 1010_20121025205800.mpg in the recordings directory and it should print out the breaks that it finds.
>>>
>>> Cheers,
>>> Tim.
>>>
>> I'm really keen to give this a go (in Australia), but I appear to be having some python troubles:
>>
>> # ./mausc-wrapper 24367
>> Traceback (most recent call last):
>> File "./mausc-wrapper", line 106, in <module>
>> main()
>> File "./mausc-wrapper", line 97, in main
>> runjob(jobid=args[0])
>> File "./mausc-wrapper", line 18, in runjob
>> db = MythDB()
>> File "/usr/lib/pymodules/python2.6/MythTV/database.py", line 965, in __init__
>> dbconn = xml.getConnectionInfo(pin)
>> File "/usr/lib/pymodules/python2.6/MythTV/connections.py", line 619, in getConnectionInfo
>> dat = self._request('Myth/GetConnectionInfo', Pin=pin).readJSON()
>> File "/usr/lib/pymodules/python2.6/MythTV/connections.py", line 607, in _request
>> for k,v in keyvars.items()])
>> File "/usr/lib/python2.6/urllib.py", line 1234, in quote
>> res = map(safe_map.__getitem__, s)
>> TypeError: argument 2 to map() must support iteration
>>
>> What version of python are you running (and maybe what distro)? I'm on mythbuntu lucid and I can't seem to get past this problem.
>> I've checked that my python2.6 packages are all up to date, as well as my libmyth-python (2:0.25.2+fixes.20121002.139bd59-0ubuntu0mythbuntu1)
> I'm using mythbuntu 12.04.1, which has python 2.7.3. I'm no python expert I'm afraid so I don't know what your error means.
>
> Cheers,
>
> Mark
Thanks for that, it gives me an avenue to try seeing as you're on a
different version. I'll try getting a backport of 2.7.3 as a first step,
and if I get some time I might look at upgrading my system, it's been a
while since it's had one.

Regards,
Andrew

_______________________________________________
mythtv-users mailing list
mythtv-users@mythtv.org
http://www.mythtv.org/mailman/listinfo/mythtv-users

Quoting Andrew Leech <coronasensei@gmail.com>:
> On 27/10/12 9:24 PM, Mark Greenwood wrote:
>> On 27 Oct 2012, at 07:56, Andrew Leech wrote:
>>> File "./mausc-wrapper", line 18, in runjob
>>> db = MythDB()
>>> <deleted>
>>> as my libmyth-python (2:0.25.2+fixes.20121002.139bd59-0ubuntu0mythbuntu1)
>> I'm using mythbuntu 12.04.1, which has python 2.7.3. I'm no python

I don't think that will help. I'm on crufty old Debian stable and it
has python version 2.6.6. The MythDB() call is where the failure is
rooted. Can you run mythpython and type 'db=MythDB()' into it?

Cheers,
Tim.







_______________________________________________
mythtv-users mailing list
mythtv-users@mythtv.org
http://www.mythtv.org/mailman/listinfo/mythtv-users

Quoting Mark Greenwood <fatgerman@gmail.com>:
> On 26 Oct 2012, at 16:04, Mark Greenwood wrote:
>> rec.storagegroup, db=db) returns 'None'.
>
> Right, found what I suspect is a bug in the python bindings for 0.26
>
> sg.dirname+filename == /var/lib/mythtv/recordings1029_20121023230500.mpg
>
> So, change the call to
>
> sg = findfile('/'+rec.basename, rec.storagegroup, db=db)
>
> and the script runs. (Haven't yet checked whether it has found the
> breaks correctly)
>
> Looks like sg.dirname is stripping a trailing slash where it
> shouldn't. Where should I report this?

I just upgrades to 0.26 to see if I get this problem and I don't. One
thing I noticed though is that all my storage groups have a slash at
the end. Do yours? I just created a new one in mythtv-setup and it put
the slash on the end even though I didn't type it in. It could be that
they are meant to have slashes at the end, a bit odd but won't kill
anything.

Cheers,
Tim.


_______________________________________________
mythtv-users mailing list
mythtv-users@mythtv.org
http://www.mythtv.org/mailman/listinfo/mythtv-users

On 27 Oct 2012, at 22:15, Tim Phipps wrote:

> Quoting Mark Greenwood <fatgerman@gmail.com>:
>> On 26 Oct 2012, at 16:04, Mark Greenwood wrote:
>>> rec.storagegroup, db=db) returns 'None'.
>>
>> Right, found what I suspect is a bug in the python bindings for 0.26
>>
>> sg.dirname+filename == /var/lib/mythtv/recordings1029_20121023230500.mpg
>>
>> So, change the call to
>>
>> sg = findfile('/'+rec.basename, rec.storagegroup, db=db)
>>
>> and the script runs. (Haven't yet checked whether it has found the breaks correctly)
>>
>> Looks like sg.dirname is stripping a trailing slash where it shouldn't. Where should I report this?
>
> I just upgrades to 0.26 to see if I get this problem and I don't. One thing I noticed though is that all my storage groups have a slash at the end. Do yours? I just created a new one in mythtv-setup and it put the slash on the end even though I didn't type it in. It could be that they are meant to have slashes at the end, a bit odd but won't kill anything.

Funnily enough the same thought just occurred to me as this email came in. My storage groups don't have a slash at the end. I'm using a default mythbuntu install, so I didn't create them myself. So it's more likely a mythbuntu bug than an mythtv bug then I guess? When I get a slack minute I'll report it to the mythbuntu guys. I'm loath to change my storage groups now because who knows what else it might break? :)

BTW once I'd made that change, your script ran and detected commercials successfully on two recordings. The CPU usage on my atom/ION backend was still too high to do flagging during recording (not that I actually tried it, but flagging an existing recording I watched 'top' while it ran and the CPU was maxed.) It's quicker than the mythcommflag_wrapper script but uses more CPU, which is not what I was expecting and I don't think it's what you were expecting either :) Must be something odd going on there. I'll keep investigating.

Cheers,

Mark

>
> Cheers,
> Tim.
>
>
> _______________________________________________
> mythtv-users mailing list
> mythtv-users@mythtv.org
> http://www.mythtv.org/mailman/listinfo/mythtv-users

_______________________________________________
mythtv-users mailing list
mythtv-users@mythtv.org
http://www.mythtv.org/mailman/listinfo/mythtv-users

On Sun, Oct 28, 2012 at 10:34 AM, Mark Greenwood <fatgerman@gmail.com> wrote:
>
> On 27 Oct 2012, at 22:15, Tim Phipps wrote:
>
>> Quoting Mark Greenwood <fatgerman@gmail.com>:
>>> On 26 Oct 2012, at 16:04, Mark Greenwood wrote:
>>>> rec.storagegroup, db=db) returns 'None'.
>>>
>>> Right, found what I suspect is a bug in the python bindings for 0.26
>>>
>>> sg.dirname+filename == /var/lib/mythtv/recordings1029_20121023230500.mpg
>>>
>>> So, change the call to
>>>
>>> sg = findfile('/'+rec.basename, rec.storagegroup, db=db)
>>>
>>> and the script runs. (Haven't yet checked whether it has found the breaks correctly)
>>>
>>> Looks like sg.dirname is stripping a trailing slash where it shouldn't. Where should I report this?
>>
>> I just upgrades to 0.26 to see if I get this problem and I don't. One thing I noticed though is that all my storage groups have a slash at the end. Do yours? I just created a new one in mythtv-setup and it put the slash on the end even though I didn't type it in. It could be that they are meant to have slashes at the end, a bit odd but won't kill anything.
>
> Funnily enough the same thought just occurred to me as this email came in. My storage groups don't have a slash at the end. I'm using a default mythbuntu install, so I didn't create them myself. So it's more likely a mythbuntu bug than an mythtv bug then I guess? When I get a slack minute I'll report it to the mythbuntu guys. I'm loath to change my storage groups now because who knows what else it might break? :)


Adding a slash at the end won't break anything!

>
> BTW once I'd made that change, your script ran and detected commercials successfully on two recordings. The CPU usage on my atom/ION backend was still too high to do flagging during recording (not that I actually tried it, but flagging an existing recording I watched 'top' while it ran and the CPU was maxed.) It's quicker than the mythcommflag_wrapper script but uses more CPU, which is not what I was expecting and I don't think it's what you were expecting either :) Must be something odd going on there. I'll keep investigating.
>
> Cheers,
>
> Mark
>
>>
>> Cheers,
>> Tim.
>>
>>
>> _______________________________________________
>> mythtv-users mailing list
>> mythtv-users@mythtv.org
>> http://www.mythtv.org/mailman/listinfo/mythtv-users
>
> _______________________________________________
> mythtv-users mailing list
> mythtv-users@mythtv.org
> http://www.mythtv.org/mailman/listinfo/mythtv-users
_______________________________________________
mythtv-users mailing list
mythtv-users@mythtv.org
http://www.mythtv.org/mailman/listinfo/mythtv-users

Quoting Mark Greenwood <fatgerman@gmail.com>:
> Funnily enough the same thought just occurred to me as this email
> came in. My storage groups don't have a slash at the end. I'm using
> a default mythbuntu install, so I didn't create them myself. So it's
> more likely a mythbuntu bug than an mythtv bug then I guess? When I
> get a slack minute I'll report it to the mythbuntu guys. I'm loath
> to change my storage groups now because who knows what else it might
> break? :)

Well, there's only one way to find out!

> BTW once I'd made that change, your script ran and detected
> commercials successfully on two recordings. The CPU usage on my
> atom/ION backend was still too high to do flagging during recording
> (not that I actually tried it, but flagging an existing recording I
> watched 'top' while it ran and the CPU was maxed.) It's quicker than
> the mythcommflag_wrapper script but uses more CPU, which is not what
> I was expecting and I don't think it's what you were expecting
> either :) Must be something odd going on there. I'll keep
> investigating.

When you run this script on an old recording the CPU is going to be
the only limit and it will run ffmpeg/avconv at 100% CPU and mausc at
whatever it takes to keep up. If you set it to start when a recording
starts it will be throttled by the incoming data rate and will use
much less CPU, plus it won't need to read the data from disk because
it will almost certainly be in the cache. The system will take less
energy and will be done quicker.

Try it out in a quiet time with a programme you don't care about. You
should see the CPU spike at the start and then drop down once things
settle into a nice steady flow.

Cheers,
Tim.


_______________________________________________
mythtv-users mailing list
mythtv-users@mythtv.org
http://www.mythtv.org/mailman/listinfo/mythtv-users

On 28/10/12 8:09 AM, Tim Phipps wrote:
> Quoting Andrew Leech <coronasensei@gmail.com>:
>> On 27/10/12 9:24 PM, Mark Greenwood wrote:
>>> On 27 Oct 2012, at 07:56, Andrew Leech wrote:
>>>> File "./mausc-wrapper", line 18, in runjob
>>>> db = MythDB()
>>>> <deleted>
>>>> as my libmyth-python
>>>> (2:0.25.2+fixes.20121002.139bd59-0ubuntu0mythbuntu1)
>>> I'm using mythbuntu 12.04.1, which has python 2.7.3. I'm no python
>
> I don't think that will help. I'm on crufty old Debian stable and it
> has python version 2.6.6. The MythDB() call is where the failure is
> rooted. Can you run mythpython and type 'db=MythDB()' into it?
>
> Cheers,
> Tim.
Hi Tim,
Yeah you're right, it's quite a reproducible failure. I had to type:
>>> from MythTV import MythDB
before db=MythDB() would resolve but that does trigger the same failure.
Looks to me like something is out of whack with my mythtv
library/bindings, probably updating mythtv and ensuring all the packages
are up to date would help before updating the entire system.
I'm always a little reticent about doing updates to mythtv, I want to be
sure I have a bit of spare time in case I have new problems to fix. I
might have to wait a bit, it's far too nice a day outside today to dive
into a myth update just now! I'll try that later.

Thanks,
Andrew
_______________________________________________
mythtv-users mailing list
mythtv-users@mythtv.org
http://www.mythtv.org/mailman/listinfo/mythtv-users

On 26 October 2012 22:34, Michael Watson <michael@thewatsonfamily.id.au>wrote:

> On 26/10/2012 3:04 PM, Anthony Giggins wrote:
>
>>
>>
>> On 26 October 2012 09:47, Roger Siddons <dizygotheca@ntlworld.com<mailto:
>> dizygotheca@ntlworld.**com <dizygotheca@ntlworld.com>>> wrote:
>>
>> On Thu, 25 Oct 2012 23:23:22 +0100, Mark Greenwood
>> <fatgerman@gmail.com <mailto:fatgerman@gmail.com>> wrote:
>>
>> > In all the years I've used mythtv (every release from 0.22 to
>> 0.26) it
>> > has *never* detected a commercial break. I've had about 3 false
>> > detections, and that's it. I know it must work for other people
>> so my
>> > question is, for all UK users, what settings do you use to make
>> it work?
>>
>> I had the same experience with mythcommflag when I tried it (some
>> time ago).
>>
>> However, have you tried
>> http://www.mythtv.org/wiki/**Mythcommflag-wrapper<http://www.mythtv.org/wiki/Mythcommflag-wrapper>? I find the
>> silence-detect method works brilliantly most of the time.
>>
>> The wiki version is for 0.24. I remember having to modify it for
>> 0.25 (IIRC the setcutlist options moved from mythcommflag to
>> mythutil). I haven't tried it on 0.26 yet.
>>
>> silence-detect
>> works pretty well in Australia too on most channels however the default
>> commflag and the silent-detect are pretty useless on the HD Channels.
>>
>
> What settings do you use for silence detect in Australia? I had very got
> very poor results when I played with it


The defaults below are commented out

MAXCOMMBREAKSECS=400
# Max level to be considered 'silent'
SILENCE_THRESHOLD="-76"
#SILENCE_THRESHOLD="-70"
# Min duration to be considered 'silent'
SILENCE_DURATION="0.15"

I've only got this enabled for

TEN?Digital) return 0 ;;
7*) return 0 ;;
SBS*) return 0 ;;

All the other channels return no or almost no commflags

Cheers,

Anthony

From: Anthony Giggins
> In all the years I've used mythtv (every release from 0.22 to
0.26) it
> has *never* detected a commercial break. I've had about 3 false
> detections, and that's it. I know it must work for other people
so my
> question is, for all UK users, what settings do you use to make
it work?
Which channels must make a large difference. For me, the defaults have worked fine since I think 0.19.
I make sure to mark BBC as non-commercial channels (although the number of adverts they now run I might change this).
I get decent commflag detection on 3,5,11,12,21

On 28 Oct 2012, at 01:31, Tim Phipps wrote:

> Quoting Mark Greenwood <fatgerman@gmail.com>:
>> Funnily enough the same thought just occurred to me as this email came in. My storage groups don't have a slash at the end. I'm using a default mythbuntu install, so I didn't create them myself. So it's more likely a mythbuntu bug than an mythtv bug then I guess? When I get a slack minute I'll report it to the mythbuntu guys. I'm loath to change my storage groups now because who knows what else it might break? :)
>
> Well, there's only one way to find out!
>
>> BTW once I'd made that change, your script ran and detected commercials successfully on two recordings. The CPU usage on my atom/ION backend was still too high to do flagging during recording (not that I actually tried it, but flagging an existing recording I watched 'top' while it ran and the CPU was maxed.) It's quicker than the mythcommflag_wrapper script but uses more CPU, which is not what I was expecting and I don't think it's what you were expecting either :) Must be something odd going on there. I'll keep investigating.
>
> When you run this script on an old recording the CPU is going to be the only limit and it will run ffmpeg/avconv at 100% CPU and mausc at whatever it takes to keep up. If you set it to start when a recording starts it will be throttled by the incoming data rate and will use much less CPU, plus it won't need to read the data from disk because it will almost certainly be in the cache. The system will take less energy and will be done quicker.
>
> Try it out in a quiet time with a programme you don't care about. You should see the CPU spike at the start and then drop down once things settle into a nice steady flow.

You're right, it worked a treat - about 5% CPU usage from avconv, mausc didn't even register. I'll be using that from now on.

Cheers,

Mark

>
> Cheers,
> Tim.
>
>
> _______________________________________________
> mythtv-users mailing list
> mythtv-users@mythtv.org
> http://www.mythtv.org/mailman/listinfo/mythtv-users

_______________________________________________
mythtv-users mailing list
mythtv-users@mythtv.org
http://www.mythtv.org/mailman/listinfo/mythtv-users

On 28/10/12 11:14, Gareth Glaccum wrote:
> *From:* Anthony Giggins <mailto:seven@seven.dorksville.net>
> > In all the years I've used mythtv (every release from 0.22 to
> 0.26) it
> > has *never* detected a commercial break. I've had about 3 false
> > detections, and that's it. I know it must work for other people
> so my
> > question is, for all UK users, what settings do you use to make
> it work?
> Which channels must make a large difference. For me, the defaults have
> worked fine since I think 0.19.
> I make sure to mark BBC as non-commercial channels (although the
> number of adverts they now run I might change this).
> I get decent commflag detection on 3,5,11,12,21
>
>
> _______________________________________________
> mythtv-users mailing list
> mythtv-users@mythtv.org
> http://www.mythtv.org/mailman/listinfo/mythtv-users
Gareth,

Which defaults ? Are you talking about default mythcommflag, with "all
available methods" ?

Similar to OP I've never found commercial skipping to be worth doing,
more errors than successes.

3,5,11,12,21 == ITV1,Channel5,picktv,dave,viva ??

Just re-run on ITV1 and still see terrible results :-( On a 2 hour ITV1
"Lewis" episode, algorithm picks up the rubbish before program start
perfectly, then finds 2 more adverts during the show, there are actually
about 6, both the ones it does pick up included significant amounts of
the episode itself.

Not complaining it doesn't work :-), just interested that it seems to
work for you.

I also tried mythcommflag wrapper, that tended to be over enthusiastic,
and find lots of small adverts all the way through. Maybe I need to
start watching American shows with continuous laugh tracks ;-)

On 29/10/12 12:36, John wrote:
> On 28/10/12 11:14, Gareth Glaccum wrote:
>> *From:* Anthony Giggins <mailto:seven@seven.dorksville.net>
>> > In all the years I've used mythtv (every release from 0.22 to
>> 0.26) it
>> > has *never* detected a commercial break. I've had about 3 false
>> > detections, and that's it. I know it must work for other people
>> so my
>> > question is, for all UK users, what settings do you use to make
>> it work?
>> Which channels must make a large difference. For me, the defaults
>> have worked fine since I think 0.19.
>> I make sure to mark BBC as non-commercial channels (although the
>> number of adverts they now run I might change this).
>> I get decent commflag detection on 3,5,11,12,21
>>
>>
>> _______________________________________________
>> mythtv-users mailing list
>> mythtv-users@mythtv.org
>> http://www.mythtv.org/mailman/listinfo/mythtv-users
> Gareth,
>
> Which defaults ? Are you talking about default mythcommflag, with "all
> available methods" ?
>
> Similar to OP I've never found commercial skipping to be worth doing,
> more errors than successes.
>
> 3,5,11,12,21 == ITV1,Channel5,picktv,dave,viva ??
>
> Just re-run on ITV1 and still see terrible results :-( On a 2 hour
> ITV1 "Lewis" episode, algorithm picks up the rubbish before program
> start perfectly, then finds 2 more adverts during the show, there are
> actually about 6, both the ones it does pick up included significant
> amounts of the episode itself.
>
> Not complaining it doesn't work :-), just interested that it seems to
> work for you.
>
> I also tried mythcommflag wrapper, that tended to be over
> enthusiastic, and find lots of small adverts all the way through.
> Maybe I need to start watching American shows with continuous laugh
> tracks ;-)
>
>
> _______________________________________________
> mythtv-users mailing list
> mythtv-users@mythtv.org
> http://www.mythtv.org/mailman/listinfo/mythtv-users
Just for laughs: Same episode found 127 adverts with
mythcommflag-wrapper. :-)

Sorry, I was wrong, I did change one variable...... maximum advert length. I up-ed it to 6 minutes. On 5 for instance (yes, the channels you guessed were correct), in CSI most of the adverts are 4 minutes 13 seconds exactly in length. This is longer than the default advert length. What I guess is that most of the devs might be American, where they tend to get much more frequent adverts, but shorter lengths.


From: John
Sent: Monday, October 29, 2012 12:36 PM
To: mythtv-users@mythtv.org
Subject: Re: [mythtv-users] Commercial Flagging (UK, DVB-T)

On 28/10/12 11:14, Gareth Glaccum wrote:

From: Anthony Giggins
> In all the years I've used mythtv (every release from 0.22 to
0.26) it
> has *never* detected a commercial break. I've had about 3 false
> detections, and that's it. I know it must work for other people
so my
> question is, for all UK users, what settings do you use to make
it work?
Which channels must make a large difference. For me, the defaults have worked fine since I think 0.19.
I make sure to mark BBC as non-commercial channels (although the number of adverts they now run I might change this).
I get decent commflag detection on 3,5,11,12,21



_______________________________________________
mythtv-users mailing list
mythtv-users@mythtv.org
http://www.mythtv.org/mailman/listinfo/mythtv-users
Gareth,

Which defaults ? Are you talking about default mythcommflag, with "all available methods" ?

Similar to OP I've never found commercial skipping to be worth doing, more errors than successes.

3,5,11,12,21 == ITV1,Channel5,picktv,dave,viva ??

Just re-run on ITV1 and still see terrible results :-( On a 2 hour ITV1 "Lewis" episode, algorithm picks up the rubbish before program start perfectly, then finds 2 more adverts during the show, there are actually about 6, both the ones it does pick up included significant amounts of the episode itself.

Not complaining it doesn't work :-), just interested that it seems to work for you.

I also tried mythcommflag wrapper, that tended to be over enthusiastic, and find lots of small adverts all the way through. Maybe I need to start watching American shows with continuous laugh tracks ;-)



--------------------------------------------------------------------------------
_______________________________________________
mythtv-users mailing list
mythtv-users@mythtv.org
http://www.mythtv.org/mailman/listinfo/mythtv-users

On Mon, 29 Oct 2012 12:46:20 -0000, John <reidjr@lineone.net> wrote:

>
> Just re-run on ITV1 and still see terrible results :-( On a 2 hour ITV1
> "Lewis" episode, algorithm picks up the rubbish before >program start
> perfectly, then finds 2 more adverts during the show, there are actually
> about 6, both the ones it does pick up >included significant amounts of
> the episode itself.
>
> Not complaining it doesn't work :-), just interested that it seems to
> work for you.
>
> I also tried mythcommflag wrapper, that tended to be over enthusiastic,
> and find lots of small adverts all the way through. >Maybe I need to
> start watching American shows with continuous laugh tracks ;-)
>
> Just for laughs: Same episode found 127 adverts with
> mythcommflag-wrapper. :-)

Have you checked the log ?

When I was setting this up I saw this "127 errors" in the Status screen
after a few seconds: it turned out the job was failing.
Can't remember why exactly but I do remember the script having permission
issues when writing its log to my Mythbuntu system.

FWIW I don't watch Lewis but, like others, it does work for me on ITV1

On 29/10/12 20:33, Gareth Glaccum wrote:
> Sorry, I was wrong, I did change one variable...... maximum advert
> length. I up-ed it to 6 minutes. On 5 for instance (yes, the channels
> you guessed were correct), in CSI most of the adverts are 4 minutes 13
> seconds exactly in length. This is longer than the default advert
> length. What I guess is that most of the devs might be American, where
> they tend to get much more frequent adverts, but shorter lengths.
>
Garreth,

Thanks for the reply. I'm a touch confused still though.

Are you talking about the default mythcommflag working ? OR are you
talking about the mythcommflag-wrapper script working for you ?

If the default mythcommflag, where do you set maximum advert length ?

thanks again.

To be honest, I can't find the option at the moment. I did not make any mods
to the wrapper script, it would have probably been through one of the setup
pages (but at the time I did also make some direct changes to the database,
for things such as automatic sleep on a frontend if it hadn't been used for
X hours, a useful but hidden variable).
Looking through the database, there is a value of 3600 seconds for

MaximumCommercialSkip


Other values..


JobQueueCommFlagCommand

AggressiveCommDetect

AutoCommercialFlag

AutoCommflagWhileRecording
(although it does work, but I rarely use catch-up, so it isn't worth the
cycles for me)

AutoTranscodeBeforeAutoCommflag

CommercialSkipMethod

CommSkipAllBlanks

Hope it helps, not having adverts is very good.
Gareth



From: John [mailto:reidjr@lineone.net]
Sent: 30 October 2012 15:58
To: Gareth Glaccum
Cc: mythtv-users@mythtv.org
Subject: Re: [mythtv-users] Commercial Flagging (UK, DVB-T)

On 29/10/12 20:33, Gareth Glaccum wrote:
Sorry, I was wrong, I did change one variable...... maximum advert length. I
up-ed it to 6 minutes. On 5 for instance (yes, the channels you guessed were
correct), in CSI most of the adverts are 4 minutes 13 seconds exactly in
length. This is longer than the default advert length. What I guess is that
most of the devs might be American, where they tend to get much more
frequent adverts, but shorter lengths.


Garreth,

Thanks for the reply. I'm a touch confused still though.

Are you talking about the default mythcommflag working ? OR are you talking
about the mythcommflag-wrapper script working for you ?

If the default mythcommflag, where do you set maximum advert length ?

thanks again.

On 28/10/12 3:37 PM, Andrew Leech wrote:
> On 28/10/12 8:09 AM, Tim Phipps wrote:
>> Quoting Andrew Leech <coronasensei@gmail.com>:
>>> On 27/10/12 9:24 PM, Mark Greenwood wrote:
>>>> On 27 Oct 2012, at 07:56, Andrew Leech wrote:
>>>>> File "./mausc-wrapper", line 18, in runjob
>>>>> db = MythDB()
>>>>> <deleted>
>>>>> as my libmyth-python
>>>>> (2:0.25.2+fixes.20121002.139bd59-0ubuntu0mythbuntu1)
>>>> I'm using mythbuntu 12.04.1, which has python 2.7.3. I'm no python
>>
>> I don't think that will help. I'm on crufty old Debian stable and it
>> has python version 2.6.6. The MythDB() call is where the failure is
>> rooted. Can you run mythpython and type 'db=MythDB()' into it?
>>
>> Cheers,
>> Tim.
> Hi Tim,
> Yeah you're right, it's quite a reproducible failure. I had to type:
> >>> from MythTV import MythDB
> before db=MythDB() would resolve but that does trigger the same failure.
> Looks to me like something is out of whack with my mythtv
> library/bindings, probably updating mythtv and ensuring all the
> packages are up to date would help before updating the entire system.
> I'm always a little reticent about doing updates to mythtv, I want to
> be sure I have a bit of spare time in case I have new problems to fix.
> I might have to wait a bit, it's far too nice a day outside today to
> dive into a myth update just now! I'll try that later.
>
> Thanks,
> Andrew
Follow up, it had nothing to do with python problems, it turns out it's
because I had an empty ~/.mythtv/config.xml . I did have an
/etc/mythtv/config.xml but apparently that's no use.
So score one for nasty cryptic error messages!

I'll follow up later with some commercial flagging results.
Cheers,
Andrew
_______________________________________________
mythtv-users mailing list
mythtv-users@mythtv.org
http://www.mythtv.org/mailman/listinfo/mythtv-users

On Mon, Nov 5, 2012 at 12:49 PM, Andrew Leech <coronasensei@gmail.com>wrote:

> On 28/10/12 3:37 PM, Andrew Leech wrote:
>
>> On 28/10/12 8:09 AM, Tim Phipps wrote:
>>
>>> Quoting Andrew Leech <coronasensei@gmail.com>:
>>>
>>>> On 27/10/12 9:24 PM, Mark Greenwood wrote:
>>>>
>>>>> On 27 Oct 2012, at 07:56, Andrew Leech wrote:
>>>>>
>>>>>> File "./mausc-wrapper", line 18, in runjob
>>>>>> db = MythDB()
>>>>>> <deleted>
>>>>>> as my libmyth-python (2:0.25.2+fixes.20121002.**
>>>>>> 139bd59-0ubuntu0mythbuntu1)
>>>>>>
>>>>> I'm using mythbuntu 12.04.1, which has python 2.7.3. I'm no python
>>>>>
>>>>
>>> I don't think that will help. I'm on crufty old Debian stable and it has
>>> python version 2.6.6. The MythDB() call is where the failure is rooted. Can
>>> you run mythpython and type 'db=MythDB()' into it?
>>>
>>> Cheers,
>>> Tim.
>>>
>> Hi Tim,
>> Yeah you're right, it's quite a reproducible failure. I had to type:
>> >>> from MythTV import MythDB
>> before db=MythDB() would resolve but that does trigger the same failure.
>> Looks to me like something is out of whack with my mythtv
>> library/bindings, probably updating mythtv and ensuring all the packages
>> are up to date would help before updating the entire system.
>> I'm always a little reticent about doing updates to mythtv, I want to be
>> sure I have a bit of spare time in case I have new problems to fix. I might
>> have to wait a bit, it's far too nice a day outside today to dive into a
>> myth update just now! I'll try that later.
>>
>> Thanks,
>> Andrew
>>
> Follow up, it had nothing to do with python problems, it turns out it's
> because I had an empty ~/.mythtv/config.xml . I did have an
> /etc/mythtv/config.xml but apparently that's no use.
> So score one for nasty cryptic error messages!
>

I believe your personal one is best dealt with as a link to the system wide
one.


>
> I'll follow up later with some commercial flagging results.
> Cheers,
>
> Andrew
> ______________________________**_________________
> mythtv-users mailing list
> mythtv-users@mythtv.org
> http://www.mythtv.org/mailman/**listinfo/mythtv-users<http://www.mythtv.org/mailman/listinfo/mythtv-users>
>

On 05/11/12 00:44, Nick Rout wrote:
>

>
> Follow up, it had nothing to do with python problems, it turns out
> it's because I had an empty ~/.mythtv/config.xml . I did have an
> /etc/mythtv/config.xml but apparently that's no use.
> So score one for nasty cryptic error messages!
>
>
> I believe your personal one is best dealt with as a link to the system
> wide one.
>

Not sure about that. In 0.25.3 I have my personal one set to readonly
as a precaoution, and get a complaint about it every time I
clearcutlist, mythcommflag --rebuild, or mythpreviewgen. But they still
work.

_______________________________________________
mythtv-users mailing list
mythtv-users@mythtv.org
http://www.mythtv.org/mailman/listinfo/mythtv-users

On 26/10/12 00:47, Roger Siddons wrote:
> On Thu, 25 Oct 2012 23:23:22 +0100, Mark Greenwood
> <fatgerman@gmail.com> wrote:
>
> > In all the years I've used mythtv (every release from 0.22 to 0.26) it
> > has *never* detected a commercial break. I've had about 3 false
> > detections, and that's it. I know it must work for other people so my
> > question is, for all UK users, what settings do you use to make it
> work?
>
> I had the same experience with mythcommflag when I tried it (some time
> ago).
>
> However, have you tried
> http://www.mythtv.org/wiki/Mythcommflag-wrapper ? I find the
> silence-detect method works brilliantly most of the time.
>
> The wiki version is for 0.24. I remember having to modify it for 0.25
> (IIRC the setcutlist options moved from mythcommflag to mythutil). I
> haven't tried it on 0.26 yet.
>
>
>
> _______________________________________________
> mythtv-users mailing list
> mythtv-users@mythtv.org
> http://www.mythtv.org/mailman/listinfo/mythtv-users
I have been playing around with Mythcommflag-wrapper
<http://www.mythtv.org/wiki/Mythcommflag-wrapper>. It needed some simple
changes to work on 0.26. It also lost the dependency on myth-export ( In
fact I couldn't get it to work properly with myth-export to find total
no frames). I've moved what I can to mythutil, and deleted the comments
that are no longer relevent. I added a few callsigns to the whitelist,
but you need to make sure your own callsigns match !! [ on Freeview its
ITV1 +1, on Freesat its ITV1+1 ]

I've added my edited version to
http://www.mythtv.org/wiki/Mythcommflag-wrapper. I have been testing on
0.26, but it probably works on 0.25.

My changes are superficial, and are just an update of the previous
script(s). It works for me, and on some programs its excellent, others
its not so lucky, but its a great deal better than the built in
mythcommflag for UK TV.

John

On 25 November 2012 21:30, John <reidjr@lineone.net> wrote:

I have been playing around with
Mythcommflag-wrapper<http://www.mythtv.org/wiki/Mythcommflag-wrapper>.
> It needed some simple changes to work on 0.26. It also lost the dependency
> on myth-export ( In fact I couldn't get it to work properly with
> myth-export to find total no frames). I've moved what I can to mythutil,
> and deleted the comments that are no longer relevent. I added a few
> callsigns to the whitelist, but you need to make sure your own callsigns
> match !! [ on Freeview its ITV1 +1, on Freesat its ITV1+1 ]
>
> I've added my edited version to
> http://www.mythtv.org/wiki/Mythcommflag-wrapper. I have been testing on
> 0.26, but it probably works on 0.25.
>
> My changes are superficial, and are just an update of the previous
> script(s). It works for me, and on some programs its excellent, others its
> not so lucky, but its a great deal better than the built in mythcommflag
> for UK TV.
>
> John
>
>
John,

This is great, thanks for your work on this. ALso seems to be working on
HD stuff, I saw on the list that there was an issue with this due to the
format HD audio was in.

Not tested fully, but getting more ad breaks detected anyway (more than the
usual 1 anyway!)

Cheers

Martin

I agree, this is awesome and so much quicker but it would be great to have
this added into the official commflagger as another commflagging option.

Cheers,

Anthony


On 27 November 2012 23:58, Martin Moores <moores.martin@gmail.com> wrote:

> On 25 November 2012 21:30, John <reidjr@lineone.net> wrote:
>
> I have been playing around with Mythcommflag-wrapper<http://www.mythtv.org/wiki/Mythcommflag-wrapper>.
>> It needed some simple changes to work on 0.26. It also lost the dependency
>> on myth-export ( In fact I couldn't get it to work properly with
>> myth-export to find total no frames). I've moved what I can to mythutil,
>> and deleted the comments that are no longer relevent. I added a few
>> callsigns to the whitelist, but you need to make sure your own callsigns
>> match !! [ on Freeview its ITV1 +1, on Freesat its ITV1+1 ]
>>
>> I've added my edited version to
>> http://www.mythtv.org/wiki/Mythcommflag-wrapper. I have been testing on
>> 0.26, but it probably works on 0.25.
>>
>> My changes are superficial, and are just an update of the previous
>> script(s). It works for me, and on some programs its excellent, others its
>> not so lucky, but its a great deal better than the built in mythcommflag
>> for UK TV.
>>
>> John
>>
>>
> John,
>
> This is great, thanks for your work on this. ALso seems to be working on
> HD stuff, I saw on the list that there was an issue with this due to the
> format HD audio was in.
>
> Not tested fully, but getting more ad breaks detected anyway (more than
> the usual 1 anyway!)
>
> Cheers
>
> Martin
>
> _______________________________________________
> mythtv-users mailing list
> mythtv-users@mythtv.org
> http://www.mythtv.org/mailman/listinfo/mythtv-users
>
>

On 26/10/12 09:22, Tim Phipps wrote:

> My contribution to MythTV is here:
>
> http://www.mythtv.org/wiki/Commercial_detection_with_silence_for_UK_freeviewHD
>
>
> It's based on the one above but works on DVB-T2 as well, does not need
> to transcode the audio stream and can be run in realtime.

I tried this, it often works well, but it is not doing the same as
mythcommflag_wrapper, as it ignores the minimum time for silence.

Although min is passed as a parameter to mausc.c, it is never used.
mausc.c looks for silence in a single video frame, 1/25th of a second,
not all along 0.15 seconds (or whatever min has been set to). As soon as
there is a silent video frame, gapstart is set, so it reacts to 0.04
seconds of silence.

The result is that if there is a short silence in a programme, it is
treated as an advert break, instead of being ignored.

It is easy to drop short breaks, like this:

if (!silent && last_silent) {
if (frames < gapstart + min)
gapstart = 0
else
gapend = frames
}

but that only works on the first break (e.g. the start of an advert). We
also need code to check that the silence at the end of an ad break is > min.

Also, I wonder how we deal with recordings that start in the middle of
an advertisement, so the first silence is the end of a break? For those
of us who start recordings a few minutes before the scheduled time, that
is often the case.

David Newman
_______________________________________________
mythtv-users mailing list
mythtv-users@mythtv.org
http://www.mythtv.org/mailman/listinfo/mythtv-users