roundtable: Re: $witched digital?

roundtable: Re: $witched digital?

Re: $witched digital?

Michael Chui (mchui@cs.indiana.edu)
Thu, 03 Mar 1994 14:23:03 -0500 

Message-Id: <9403031922.AA28756@a.cni.org>
To: roundtable@cni.org
Subject: Re: $witched digital? 
In-Reply-To: Your message of "Wed, 02 Mar 1994 12:29:07 EST."
             <Pine.3.85.9403012227.A12541-0100000@essential> 
Date: Thu, 03 Mar 1994 14:23:03 -0500
From: Michael Chui <mchui@cs.indiana.edu>

James Love <love@essential.org> writes:
>How much bandwidith is needed for switched video into the home, 
>and what are the costs of provding that bandwidith, given current 
>technology?

     An analog cable channel requires 6 MHz.  However, as the MIT Media 
Lab folks are fond of telling us, the bandwidth required to send video 
information (at a given picture quality) is also dependent on how smart 
the devices send and receiving the signal are.  For example, if you have 
incredible amounts of memory, and you're willing to wait a *long* time, 
you could send an entire movie through a 1200 bps modem, store it on your 
gobs of disk, and play it back when it all arrives.  (This would be the 
equivalent of filling a supertanker through a straw, however. :-)

     Assuming that you are asking about real-time video (i.e. not 
store-and-play), the MPEG-1 digital video compression format is optimized 
to provide video (352 x 240 pixels at 30 frames/second) at a data rate of 
1.5 Mbits/s.  This is approximately equivalent to a T1 line, which costs 
around $200-$300/month around here.  Hardware implementations of MPEG are 
becoming more common, so an MPEG card for one's microcomputer could soon 
cost in the hundreds of dollars.

     Preliminary work on MPEG-4 (MPEG-2 is optimized for 3-10 Mbits/s,
MPEG-3 is for HDTV), which is designed for low bitrate applications
(up to 64 Kbps, 176 x 144 at 10 frames/s) has begun.  *Notice that these
are bitrates which are achievable over standard POTS or ISDN lines.*
It can be done - AT&T sells a video phone which runs over normal
phone lines, albeit with extremely poor picture quality.  It's fairly
safe to assume that it will be at least a couple of years before MPEG-4
hardware implementations are widely available.

     From a policy perspective, I think it's important that we start
to tariff flate-rate digital local loop access.  Bitrates will increase
with improving technology, and compression standards will evolve, but
it's important to encourage content providers to develop applications
by providing a conduit for these services in the local loop, even if
they are constrained by low picture quality, or store-and-play
applications.

Michael Chui
mchui@cs.indiana.edu

[CNI Home Page]