To: WSFAlist at KeithLynch.net
Date: Sun, 24 Dec 2006 23:42:14 -0500
Subject: [WSFA] Re: Data processing by the human brain (was: modems)
From: ronkean at juno.com
Reply-To: WSFA members <WSFAlist at KeithLynch.net>

On Sat, 23 Dec 2006 11:06:51 -0500 "Mike B." <omni at omniphile.com> writes:
> At 12/23/2006 01:13 AM, ronkean at juno.com wrote:
>
> >While on the subject of 300 bps data streams, I note that 300 bps
> happens
> >to be roughly the speed limit at which a person can read English
> text,
> >assuming that each character of text represents 7 or 8 bits.
>

> 300 baud is not 42.8 or 37.5 characters per second.  There are extra
>
> bits (or more correctly, "mark" or "space" data) sent to mark the
> start of a byte and the end of a byte.  A better approximation is
> that the cps rate is about 1/10 the baud rate, so 300 baud would do
>
> about 30 characters per second.  You can easily read at this
> rate...I've done it when using a 300 baud modem to connect to a BBS
>
> system.  A bit faster would have been nice.
>

Quite so.  I had ignored/forgotten the start and stop bits in my rough
calculation. The max character rate at 300 bps (bits per second)
asynchronous would be about 30 cps (characters per second), as you say.
Also, the actual character rate is not even determined by the bps rate;
only the maximum character rate is determined by the bps rate.  The
sender might only type one character per second, in which case the
character rate would be one character per second, even though the
asynchronous data clock is running at 300 Hz.

And I agree that a human reader who can only read as fast as 30 cps is a
relatively slow reader, the reading 'speed limit' (with comprehension)
being considerably higher, maybe 100 cps, at a guess, which would be
about 20 words per second, or maybe even as high as 200 cps.  Trained
speed readers can probably go much faster than that, but I'm skeptical
that they can do so with good comprehension.

Where I was trying to go with all of this, was to pin down an estimate of
how high a data rate the human brain can 'take in', with comprehension.
To that end, I raised the issue of data compression.

> >Since
> >English text can be data compressed by a ratio of about three to
> one,
>
> Or more...
>
> >I would conclude that a human reader can handle a true data input
> rate of
> >only about 100 bps.  That estimate is roughly consistent with the
> >estimate that conversational speech has a data content rate of
> about 50
> >bps, and that most people can probably follow rapid speech at two
> or
> >three times the normal conversational rate.
>
> Since I'm not sure if your "bps" is "bits per second" or "bytes per
>
> second", I won't comment on the estimates, but I know for sure that
>
> at least some people can handle spoken data at a rate several times
>
> what most people actually use.  Text to speech software for blind
> people often outputs at several times the usual spoken rate, and
> with
> some practice, they have no trouble following it.
>

bps = bits per second, Bps = bytes per second.

The basic unit of information is the bit, and a byte is arbitrarily
defined as 8 bits.  What I was looking for is an estimate of how many
bits per second of information flow rate can be comprehended by a human.
Recognizing that just about any method of information transfer (e.g.
text, audio, or video) includes some redundancy, I want to discount the
redundancy by putting the information through a compression algorithm to
remove most of the redundancy, and arrive at the 'real' data rate, as
distinguished from the 'raw' or 'gross' data rate. To that end, I chose
text as the originally most compressed form of natural information
transfer.  The next step would be to run the text through a text
compression process, to determine the fundamental information rate.

> >So, here's the question.  If people can accept information input
> only up
> >to a rate of about 100 bps, reading text, why is it that people
> can
> >easily 'take in' a movie, which requires a streaming data rate of
> perhaps
> >tens of millions of bits per second, for a wide screen movie
> theater
> >presentation?
>
> Different parts of the brain are involved, and vision has peripheral
>
> processors in the eyes too.  Also:
>
> "All of the books in the world contain no more information than
>   is broadcast as video in a single large American city in a single
>   year. Not all bits have equal value."
>                                          -- Carl Sagan
>

After reading Ted's comments, I now think it is hopeless to try to
measure the maximum human information intake rate by the method of
measuring how fast humans can read text, and then applying data
compression.  The same point is implied by the Sagan quote you provided.
Data compression of text is an open-ended wild card.  The entire text of
'Moby Dick' could be compressed as the Library of Congress catalog number
of that book, if the compression algorithm included the Library of
Congress in its data base.  Plain text itself is a form of compressed
data, since we usually need the rich data base of human experience to
properly interpret a block of text.

Ron Kean

.