🚀 go-pugleaf

In article <1992Nov12.162330.24658@newstand.syr.edu>
dwjurkat@rodan.acs.syr.edu (Jurkat) writes:
>Here is my idea.  Take a commercial, digitize the beginning and end of
>it and stick it in a computer and also include the length of the
>commercial.  Then hook the computer to the VCR and when the VCR is used

[...] Stuff about comparing digitized samples of video deleted

>the most important parts of the signal (like maybe just a few lines of
>any given frame should be enough).

>My question:  Is this feasable, what is the minium amount of information
>needed from a frame or set of frames, is it possible to at least nix the
>most anoying commercials?

Feasable, maybe.  Practical, no.

Not getting into the discussion of how much CPU/memory bandwidth is
required to digitize video in "real-time" to any decent level of
information that could be used to compare against another signal, 
I forsee no way of comparing this information against even the same
commercial played off a video tape for instance.  It is easy enough
to compare them but the difficulty is in trying to decide if they
are the same.   There is so much noise, and other junk that gets
digitized along with it that would be different each time (not to
mention, just not starting the sampling at the exactly same spot
on the waveform). There are many other problems with this idea
as well.

>By the way I was wondering why do commercials always seem significantly
>louder then the program you are watching and can this(if true) be used?

Commercials are always louder than the programs they are stuck
in.   It is what the advertisers pay for.  This is intentional.

						.....Rich
   					rsnider@xrtll