[mythtv-users] RE: mythtv-users Digest, Vol 26, Issue 42

Robert Denier denier at umr.edu
Mon Mar 7 01:15:53 UTC 2005

I have never studied mpeg in great detail although I am very familiar
with fourier series approximations.  Mpeg2 is based on some fixed
blocked size.  It may be 16x16, but I don't remember.  One thing that
should be possible is if say it is 16x16 and if the full sized video was
say 1920x1440 then decoding to 120x90 would be in fact trivial provided
you had all the underlying structure in place.  This makes the
assumption that there is the equivalent to a dc level for the entire
small block.  You might be able to similarly get larger sizes by only
decoding some of the coefficients that are used in the discrete cosine

Then again there are forward and backwards predicted frames which may
mess all that up.  Overall it should be possible to optimize the
decoding for a lower resolution than what your starting with, but it may
in fact be very non trivial.  It could take months of work.  Well likely
it would take a couple months just to really get the understanding of
mpeg2 needed to do the actual work.  Perhaps longer.  Also since mpeg2
is a licensed algorithm, there probably are legal issues involved as

If someone was going to do it I'd probably start with XVID.  It should
be somewhat documented and I don't think there would be legal issues in
creating your own decoder.  Then again this could all be an attempt to
reinvent the wheel, so considerable additional research would be needed
to make sure the problem hasn't already been solved.

Robert Denier

On Sun, 2005-03-06 at 19:17 -0500, Jonathan M. Cooper wrote:
> I get where you're coming from.  I just like to occasionally have a
> baseball game on in realtime in a PIP window while watching a recorded
> program.
> -----Original Message-----
> Date: Sun, 6 Mar 2005 13:02:55 -0800
> From: Brad Templeton <brad+myth at templetons.com>
> Subject: Re: [mythtv-users] Dual HDTV Tuner Requirements
> To: Discussion about mythtv <mythtv-users at mythtv.org>
> Message-ID: <20050306210255.GC6782 at main.templetons.com>
> Content-Type: text/plain; charset=us-ascii
> On Sun, Mar 06, 2005 at 07:59:10AM -0500, Jonathan M. Cooper wrote:
> > Thanks for your response.  Does that mean that PIP is not feasible
> with,
> > say, a P4 3.2GHz?
> I remain curious as to why people want PiP when they have MythTV.  Many
> people who use Myth stop watching live TV altogether, and I am amazed
> that
> the desire would remain to watch _two_ live TVs.
> But indeed, it might be a CPU problem since nobody has written yet a
> "clever" mpeg decoder which knows it is decoding for a tiny size and
> thus
> avoids the cpu load of decoding for full size and rescaling.
> A number of people are interested in such a decoder, since it would
> allow
> some people with older processors to watch HDTV recordings on their
> SDTVs.
> Still SDTV, but a fair bit better than NTSC recording.
> The EyeTV program for the Mac has such a clever decoder, so it is
> possible.
> _______________________________________________
> mythtv-users mailing list
> mythtv-users at mythtv.org
> http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-users

More information about the mythtv-users mailing list