[UACCESS-L] Historical perspective on tonal output of print:

["David Poehlman" <poehlman1@comcast.net>]

I find this a fascinating look and wonder what might be achievable at
current technological levels.

What Happened To the Tonal Output?

The VA has done much research on audible-output
direct-translation reading aids. Most of them
translated character shapes into tone patterns.
The work was never finished; here is what happened.

The last of these instruments to have been evaluated
was the Stereotoner. Several people including Mr.
Lauer use it to date. The results were: (1) The
best users could perform on a par with the best
Optacon users. 2) Fewer people in the population
could learn it than could learn the Optacon. 3)
A few people could make better use of the
Stereotoner than the Optacon.

After the Stereotoner was evaluated, we found
that its optical system was somewhat superior
to that of the Optacon, but its photo-sensitive
elements and its vertical resolution were very
inferior. We found that a 20-tone code is better
than the Stereotoner's 10-tone code. We discovered
this by testing prototypes which were built to our
specifications. They add the tonal output to the
Optacon. We christened the new combination
Optaudicon. The fidelity with which they rendered
the printed character shapes in audio form was
very much improved over the Stereotoner. The
testing we did led us to suspect that the tonal
code may be of equal value to the tactile code of
the Optacon. This calls the Stereotoner evaluation
into question. The work done also suggests, but
does not prove, that the tonal output would be a
worthwhile add-on to the Optacon and to a combination
direct-translation and OCR machine.

We found that some people who can learn to use a
direct-translation machine cannot use the tactile
output, and some do better with the audible output
presented as tone patterns. The evidence also
suggests that many users experience a complementary
effect with the bimodal approach; that is, using
both outputs. For example, the tactile code has a
wider window for tracking; it's better for vertical
resolution; but the finger loses sensitivity (feels
numb) more quickly than does the ear. The tonal code
is better for reading dense horizontal data and
descenders, etc. Hearing is not as susceptible to
such loss of sensitivity or negative adaptation as
is touch. The additional cost of an audible output
would be low--perhaps $150.

We failed to garner support to finish the work. The
reasons would fill a chapter. To begin with, we
lacked Ph.D. degrees and computers with which to
compete in the research environment. We therefore
turned from our main endeavor to implementing
computer access which we ourselves badly needed.
That's how we became involved with companies like
RDC. Further research is needed to test the value
of the tonal code and the bimodal approach. Such
details as were published can be found in our 1980
article on reading machines; the rest is in
unpublished documents. At this point, however, we
are free to be a resource to accomplish this work,
either inside or outside the VA.

Excerpted from the source URL:
http://personalpages.tds.net/~ti51/new72-73.htm