On Fri, 18 Feb 2005 01:06:59 -0500 (EST), Chris Pinkham <[EMAIL PROTECTED]> wrote: > > The patch also adds some values to the edgeDiffs[] array. I have > > experienced needing a value as high as 50 to detect the logo area. > > This is in there also, I don't beleive this can cause false positives. > Are these logos transparent or opaque? I never even tried bumping it > up that high, maybe that's why I wasn't getting results as good as I > hoped when testing opaque logos. >
I have only tried it with two different logos. One of the logos (Norwegian ZTV for those interested) is a opaque logo with a transparent ring around it. The other logo (Norwegian TV3) is somewhat transparent. I usually need a high edgeDiff value for both logos. The needed value is related to the first frame behavior I mentioned on mythtv-users. If the logo is present at the beginning of a recording it's usually enough with edgeDiff = 7. If the recording is started during commercials and the logo is not present until a couple of minutes into the recording, a value of 40-100 is needed. I don't understand why, though. Regards, Petter
_______________________________________________ mythtv-dev mailing list [email protected] http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-dev
