[mythtv] rolling up my sleeves: a few questions about commflagging

[Brian J. Murrell]

So, I want to dig into why commflagging is less than accurate on my
installation here and I am hoping I can ask a few questions to get me
started.

The good news is that looking at the full output from the "blank"
method, mythcommflag is finding all of the blank frames correctly.  For
my installation here it could even probably stand to have it's threshold
of what's "black" turned down, but nonetheless, it does seem to be
finding all of the blank frames correctly.

So now it's just down to the algorithms used to determine which blank
frames denote the start and end of commercial breaks.

So first question... are the d2_* flaggers what are used when one
selects Setup->Video->General->Enable experimental speedup of commercial
detection?

Are those still experimental?  They seem like they have been around a
long time.  d2_blank seemed to do a lot better than blank at coalescing
the blanks into the full commercial break than blank did.

Second question given the output from either of the {d2_,}blank
flaggers, where in the code are those maps coalesced into what actually
goes into recordedmarkup?  I want to analyse those algorithms and see
why the are making mistakes.  Is the same algorithm used for both
d2_blank and blank or do they each have their own algorithms for
figuring out which blank frames are the start and ends of the group of
commercials in a single break?  Given the above, I imagine they are
different.

Cheers,
b.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 473 bytes
Desc: This is a digitally signed message part
URL: <http://www.mythtv.org/pipermail/mythtv-dev/attachments/20140124/659b33b3/attachment.sig>