[mythtv-users] commflag on master very unreliable any more

[Mike Perkins]

On 17/04/12 03:33, Brian J. Murrell wrote:
> On 12-04-16 09:07 PM, jk90090 wrote:
>>
>> Brian, still having this issue?
>
> Yep.  I still regularly have recordings where only a single break is
> flagged and sometimes only the opening of the break and that's it.  Yet
> I can (visually) quite clearly see very obvious markers of commercial
> breaks -- like long spans of black frames.
>
>> I know I'm coming into this thread late, but I too am having
>> commflagging troubles often enough to be troublesome.  Not all
>> recordings, but probably double-digit percentage are jumping at the
>> wrong time and/or for too large of a jump.  All my recordings from an
>> HDHR using QAM via Cable, and signal strength is very high.
>
> Most of my problem recordings are from QAM also, mostly from an HVR-950Q
> with some from a PVR-1600 when the HVR-950Q is busy.
>
> Recordings from PVR-x50 analog MPEG2 encoder cards seem much more
> reliably commflagged.
>
> But I've mentioned all of this before and nobody seemed interested.
>
These are digital signals, no? It is quite likely that there are changes in 
format at advert breaks which mean that the commflag software doesn't detect the 
breaks properly. For example: changes between 1080i/p and 720i/p, aspect ratio, 
frame rate, bit rate, whatever. The adverts themselves could easily be a 
different format to the program which precedes and follows it (and each advert 
in a break could be different, as well!).

On the other hand, your PVR-x50 signal has been *encoded* as a single stream by 
the tuner card, and therefore *must* all be at the same rate, format, etc. This 
is a lot easier for a commflag program to analyse.

I'm not sure what the answer to this problem is. Of course, one way might be to 
transcode *first*, that way you'd be simulating what a PVR-x50 card does, but at 
least the resulting stream would be all the same format.

-- 

Mike Perkins