Aucbvax.1371
fa.human-nets
utzoo!duke!mhtsa!eagle!ucbvax!DERWAY@MIT-ML
Tue May 19 19:22:55 1981
HUMAN-NETS Digest V3 #101
HUMAN-NETS AM Digest Sunday, 17 May 1981 Volume 3 : Issue 101
Today's Topics:
FYI - Call for Proposals for NCC '82,
Computers and the Handicapped - Color Blindness,
Query Replies - 3D Displays & Voice Grade Line Bandwidth
----------------------------------------------------------------------
Date: 14 May 1981 (Thursday) 1714-EDT
From: MORGAN at WHARTON-10 (Howard Morgan)
Subject: NCC '82
For those of you who read human-nets, please submit proposals for
sessions, or papers, directly to MORGAN@WHARTON. As technical program
chairman, I will examine submissions and proposals, and route them to
appropriate program committee members. Any of you who would like to
help with NCC in other ways, please let me know via message.
Howard Morgan
------------------------------
Date: 13 May 1981 1155-EDT
From: GEOF at MIT-XX
Subject: Re: HUMAN-NETS Digest V3 #98
RE: Colorblindness.
It is unfortunate that we use red and green for no/yes stop/go
signals, since red/green confusion is the most common of all. The
problem, however, is well known, especially to designers of road
signs. For example, most traffic lights are mounted vertically, with
red at the top. If you can see the socket (a problem at night), you
can tell the color from the context. One set of horizontal lights I
saw uses a diamond for yellow, a square for red, and a circle for
green. An added clue is that there are always two red lights lit.
Similarly, in Quebec there are two sets of road- signs: those that
mean you must do X and those that mean you may NOT do X (i.e., no left
turn can be "no left turn" or "straight and right only"). The signs
were originally identical, except that the Must's were circled in
green and the Don't's in red. The colorblindness problem presented
itself, and was solved by introducing a red diagonal into the Don't
signs.
If a colorblind person can't understand a public computer
display, the fault is clearly with the designer of the software.
There are no end of additional clues which may be devised. Of course,
it is not always evident just what the clues mean (who says that a
square means red and a circle green). But for PUBLIC displays, people
who need the clues can find out for themselves rather quickly, given
that the clues are simple and consistent.
In a private display, like the VLSI plots mentioned, the
problem is more difficult, and it is not clear that it need be solved
in every case (sorry to all you color blind out there). For example,
I don't think there is a way around fancy plotting problems, other
than choosing pencils of radically different luminences (you might
also try flourescent and metallic colors). The real solution is to
use a color graphics printer, where the lines can be textured. Xerox
software uses texture to display similar plots on a black and white
dover.
------------------------------
Date: 7 May 1981 1206-EDT
From: SWG at MIT-XX
Subject: stereo display
By coincidence, I just read "Going to the Roundies" in the March
Scientific American's "Science and the Citizen". Excerpts:
Investigators at the Cinema and Photographic Research Institute
[C.P.R.I.] in Moscow are at work on a holographic motion-picture
system for public theaters. In 1976 the group exhibited a 45-second
holographic movie in which a young woman holding a bouquet of flowers
in front of her face seems to walk out of the screen toward the
audience. Each viewer, by moving his head horizontally or vertically,
could look around the bouquet and see the woman's face. The effect
was startling, although the registration of the image from frame to
frame was not perfect and the color was unrealistic (the image was
yellow). The movie and the way it was made are described by Tung H.
Jeong of Lake Forest College in \Optical Spectra/.
In the years since the 45-second movie was shown the techniques of
holographic cinematography have improved considerably. The recent
developments are discussed by two members of the Moscow group, Victor
G. Komar and O. I. Ioshin, in \SMPTE Journal/, a publication of the
Society of Motion Picture and Television Engineers. . . .
The theoretical and technological difficulties in extending holography
to motion pictures are substantial. The [C.P.R.I.] is one of only a
few institutions where the work has been attempted. Surprisingly, the
most challenging problems may lie not in making holographic motion
pictures but in projecting them to a large audience. The projection
of the hologram onto an ordinary movie screen would destroy the
reconstructive property of the image. It is necessary for the
observer to intercept not a projection of the light rays emerging from
the hologram but the rays themselves.
One hypothetical method of presentation, explained by Jeong, would
employ a mirror rather than a diffusely reflective screen as the
projection surface. For example, the mirror might be an elliptical
one with two foci. The viewer would sit at one focus and the
holographic image would be formed at the other focus. [I assume the
image degrades gradually as the viewer moves away from the focus.]
The drawback to this method is that it would work only for an audience
of one. For a larger audience the screen would have to be equivalent
to a superposition of as many mirrors as there are viewers, with the
image at a common focus and each seat at the second focus of one of
the mirrors.
It is impossible, however, to combine mirrors in this way. The
investigators in Moscow, according to Jeong, overcame the problem by
employing as the screen a specially constructed hologram that has the
same multiple-focusing property as superposed mirrors would have. The
hologram is made by exposing a photographic plate the size of the
screen [!] to beams diverging from each seat in the theater and from
the point where the image will be formed. When light rays from the
projector impinge on the holographic screen, they come to a focus at
each seat. The movie of the woman and the bouquet was shown with a
screen that could accommodate an audience of four. The investigators
plan to construct a seven-foot-wide holographic screen for an audience
of between 200 and 400 viewers. [No doubt the size of the screen
limits not the size of the image but rather its brightness,
resolution, etc.]
------------------------------
Date: 12 May 1981 1055-MDT
From: Spencer W. Thomas <THOMAS AT UTAH-20>
Subject: Stereopsis
A comment on displays which attempt to simulate 3-D by presenting a
different image to each eye (so called 'stereo pairs'). A fair
portion of the population cannot see any depth in these images
(something astounding on the order of 20-30%). Back in the early days
of graphics, Ivan Sutherland was working on a stereo display, and was
having much difficulty getting it to work. Finally, he asked a
colleague, who said 'But this is great!'. Sutherland went out and got
his eyes tested and discovered that he couldn't see depth in stereo
pairs. (Allegedly a true story.) We once came up with 7 cues to depth
in normal viewing of the 'real' world:
1. Stereopsis
2. Motion parallax
3. Perspective
4. Occlusion of further objects by nearer objects
5. Color saturation (as things get further away, they tend
towards less saturated colors due to atmospheric effects)
6. Size cues (i.e. you know how big a person should be, so if
heesh looks too small, heesh must be far away.
7. Focus.
I'm sure that if you think about it, you can come up with some more.
=S
------------------------------
Date: 11 MAY 1981 2308-PDT
From: CAULKINS at USC-ECL
Subject: voice grade lines + 9600 baud
We recently leased a line from the phone company spec'd to be DC
continuity between the two connected points. We hung some short
distance modems on the line, but were unable to push more than 4800
baud over the line (we too hoped for 9600). When we checked with the
telco wizards they said that there were loading coils in the line, and
we were lucky to get 4800. To get the loading coils removed would cost
a mere $1500. We are resigned to running at 4800.
Dave C
------------------------------
Date: 12 May 1981 01:39-EDT
From: Charles Frankston <CBF MIT-MC AT>
Subject: baud vs. bps and line driver performance issues
A baud is a signal transition. As someone else pointed out, a signal
transition can very well be used to encode more than one bit of
information (quadrature modulation and other techniques). Therefore
9600 baud may not necessarily equal 9600 bits per second. Its as
simple as that. However, the computer industry constantly misuses the
word baud. It would have been much clearer for your terminal
manufacturer to say your terminal ran at 9600 bps.
Your line driver not being able to run 10 miles at 9600 bps is hardly
surprising. First thing that comes to mind is that the line you are
leasing from the phone company may not have DC continuity (ie. it is
not just a set of wires). It used to be hard to obtain a "metallic"
circuit from the phone company, but most seem to offer it now. There
are probably equalizers (capacitors) etc. on the line. Even without
that, there are not that many line drivers that are rated for that
speed at 10 miles (in fact I can't think of a single one offhand).
There ARE some short haul modems that can probably make it, (The
difference between a short haul modem and a line driver is that a
modem maintains continuous carrier on the line. The problem a line
driver will run into at high speed is that the signal transitions
having to charge a line up from the idle state on every signal
transition first have to charge the capacitor represented by a
metallic pair of any substantial length.) but they are generally more
expensive ($800 or so vs. $200 or less) than line drivers. Other
obvious factors affecting performance would be the gauge of the wire
and how well twisted it was.
Is the readership really interested in these low level
telecommincations issues? I seem to remember a huge number of issues
back recommending McNamara's book "Technical Aspects of Data
Communications" available from Digital Press (yes, that publisher is
Digital Equipment Corporation).
------------------------------
Date: 12 May 1981 1119-PDT
Subject: Leased line and line drivers.
From: the tty of Geoffrey S. Goodfellow <GEOFF AT SRI-CSL>
Having been outfitted with a 9600 baud leased line to my house for
several months now, perhaps I can offer some help.
Firstly, you MUST make sure the line(s) you have [it takes two pair,
i.e. 4 wires] are Non-loaded and have DC continuity. You order the
lines from your local TPC mentioning this and Bell specification
#43401.
Secondly, I do not know of any regular/standard line-drivers that will
drive 9600 baud over 4 miles. However, Prentice offers a line-driver
(the one I am currently using for my hookup) called the ALD-XR (for
eXtended Range) that'll handle 9600 baud up to 6 miles. Their regular
ALD only goes 4 miles for 9600 baud.
I found all this out when Prentice sent me the regular ALD's by
mistake first and I couldn't go any faster than 2400 baud (my home
Datamedia only goes 110, 300, 1200, 2400 or 9600), and sent them back
and got the ALD-XR's.
I get a few hits on the line every now and then, with an occasional
tilda appearing in my output, but it only happens every couple of
days.
Here's Prentice's table:
Miles with Non-Loaded
26-AWG Cable
Baud ALD ALD-XR
1200 10 15
2400 8 12
4800 6 9
7200 5 7.5
9600 4 6
>From the looks of this, I'm not surprised you can't go over 2400 baud
(providing you have an ALD-XR or equalvalient).
Guess you'll just have to move in closer or spring for expen$ive and
big modems if you still want 9600.
------------------------------
End of HUMAN-NETS Digest
************************
-----------------------------------------------------------------
gopher://quux.org/ conversion by John Goerzen <
[email protected]>
of
http://communication.ucsd.edu/A-News/
This Usenet Oldnews Archive
article may be copied and distributed freely, provided:
1. There is no money collected for the text(s) of the articles.
2. The following notice remains appended to each copy:
The Usenet Oldnews Archive: Compilation Copyright (C) 1981, 1996
Bruce Jones, Henry Spencer, David Wiseman.
