[mythtv-users] Mythcommflag in Italy or Europe

[Chris Pinkham]

> On 02/16/2006 02:04 AM, Thomas Börkel wrote:
> > Michael T. Dean wrote:
> >> Besides the default strategy (blank frame), there's scene change, logo, 
> >> blank frame and scene change, and all.  And, note that you must have 
> >> "Automatically Flag Commercials" enabled to use anything other than 
> >> blank frame.
> > Why is it necessary to enable automatically flagging to be able to use 
> > all stragegies?
> >
> > I'd like to flag some older recordings with all strategies.
> >   
> 
> How's this for a useless answer?  Because the documentation for the 
> setting says so:
> 
> Commercial Skip Method
> This determines the method used by MythTV to detect when commercials 
> start and end.  You must have 'Automatically Flag Commercials' enabled 
> to use anything other than 'Blank Frame'.

And since that appears to be out of date, unfortunately it's even more
useless. :)

It used to be that you could skip commercials using blank-frame detection
even if you did not run the commercial flagging job after a recording
completed.  This code was nowhere near as accurate as the post-recording
commercial flagging and was slow because it had to search forward through
the video looking for blank frames whenever you tried to skip a commercial.
Because of these and several other reasons, this code was removed.  We also
used to detect blank frames at recording-time when using software encoding.
This worked great 3-4 years ago, but as more and more stations/channels
stopped putting in blank frames as liberally as they used to, the effectiveness
of this became rather minimal for most people.  As part of a reorganization
of the commercial flagging code, the code was moved out of libmythtv and
into mythtv/programs/mythcommflag, so the blank-frame detection at
recording-time was removed at this point requiring all users of all
detection methods to turn on the "Auto Commercial Flagging" in order to be
able to skip commercials.  The setting schedules a job in the JobQueue either
at the beginning of the recording or at the end and mythcommflag is then
called to process the recording and flag the commercials.

> I think it may have to do with the way recording works with software 
> encoders.  When using a software encoder, Myth writes (used to write?) 
> information about some of the blank frames to the database, and this 
> could be used as a "poor-man's skip list" (i.e. could be used for 
> LiveTV--or even recordings--without requiring the user to run mythcommflag).

Yep, "used to write" is correct, this is no longer done.  Blank-frame
detection by itself hardly works at all for me anymore, that is part of the
reason I started working more on the 'All' method which utilizes blank-frame,
scene-change, logo detection, letter/pillar-box detection, aspect-ratio
detection, etc.. (no audio yet and if anyone replies to this email with
a "maybe you could use method 'blah'", I'll silently ignore it. I have lots
of ideas but not enough time and the current 'All' works well enough for me
right now that I don't spend a lot of time on the flagger).


> So, it's possible that this part of the documentation is not current or 
> is just a bit misleading (especially for users of hardware encoders/DVB 
> cards)--I'll have to defer that question to Chris Pinkham or someone 

Which part of the docs is this in? (coming from someone who can count on
two hands the number of times he's actually read them since I've been using
Myth since 0.7).

-- 
Chris