[mythtv] Re: [mythtv-commits] mythtv commits

Risto Treksler risto at elkhornbanff.ca
Wed Apr 20 23:25:45 UTC 2005



mythtv at cvs.mythtv.org said:

> ----------------------------------------------------------------------------
> Changes committed by cpinkham on Wed Apr 20 04:44:05 2005
> * Commmercial Flagging changes:

> - Rearrange some debugging statements.
you changed

"block %1 with length %2, which would put "
"comm block length under minimum of "
"%3 seconds.")

but this is actually correct for that part of the code
trust me, i tested it, there's a reason i worded it that way.

that is part of the bugfix for flagging tiny commercials
that were getting closed off if the length was less than max length
without a checking against min length 

iirc, it has to refer to the length of the previous block, not the next. 

>   - Revert the change (temporarily for now) that made strict mode require
>     blank frames to be totally blank so frames with logos on them but
>     otherwise blank would not be detected as so in strict mode.

:((
I know you said temporarily, 
but requiring blank frames to be totally blank works tons better (imho)


I know, "better" is subjuective
Here's some objective stats from 50 recordings on my drive (long):

--for the 80% of shows that don't leave the logo up--
    -it (0.18 with strict mode) detects 100% correctly 99% of the time!!!

--for the 10% shows that leave it up at the start of a commercial--
    -it still detects "100%" correctly 99% of the time
     since the end is all that matters for a skip
       -it just won't auto skip 
       -you have to hit 'z'(once or twice)

--for the other 10% of shows that leave it up at the end-- 
    -it detects 100% correctly only 5% of the time
    -it comes back too early 90% of the time which is good
     (because of the scenechange and logo detection routines)

the over all rate of false positives is --1%--
ie 10% of the time in 10% of shows


On the other hand,
Not requiring blank frames to be totally blank gets lots of false positives

My observed stats:
--for the 80% of shows that don't leave the logo up--
    -it (current cvs) detects 100% correctly only 5% of the time!!!!!
    -it detects 90% correctly 95% of the time

--for the 20% shows that leave it up 
    -it still detects "100%" correctly only 5% of the time!!!
    -it detects 80% correctly 95% of the time

rate of false positives is harder to estimate,
i'd say 20% of the time in 90% of shows, so --18%--

====
ie according to my empirical observations
(it's pretty sad how much time i spent on this - easily 8 hours)
it's 18 times worse to allow logos on "blank frames"
====

There has to be another way to deal with the few shows that leave the logo up.
Breaking comm detection for 90% of the shows, 
because 10% leave the logo up at the end of the commercial, 
doesn't make sense to me.


temporarily, maybe this should have its own setting
instead of depending on the strict setting

something like
    blankCanHaveLogo =
              gContext->GetNumSetting("BlankCanHaveLogo", 0);

BUT
===
there's got to be a way to do a check for blank frames that have logos on them
to see if the 60th frame before or after it, also has a logo on it, otherwise
don't mark it as blank.

you see what i mean?

that way only blank frames that really border on commercials get marked as blank.






More information about the mythtv-dev mailing list