We developed and tested an innovative concept to assist blind and visually impaired consumers with their reading needs. The concept was to provide "access on demand" to all print materials, realized by having a remote, human reader system comprised of a fax machine (to send the printed information) and a fax computer display (to receive the printed information for reading by a sighted person). The project included exploration of novel applications of telecommunications technologies to augment the basic system. Significant support for the project was obtained from the Easter Seals Research Foundation, and subsequently an NIDRR Field Initiated Grant enabled exploration of advanced system features. We also received support from Allnet, Inc., in the provision of toll-free telephone numbers, and from the Sydney Stern Foundation.
We established a trial fax-based reading service with two sighted, mobility-impaired readers telecomputing from their homes daily from 1:00 p.m. to 9:00 p.m. in two 4-hour shifts. Twenty local and six distant blind recipients of fax machines represented a broad spectrum of reading needs. They used the system for both personal and professional reasons and have extended periods in their day during which other types of reading assistance is not at hand.
For each document read, a record was made of significant document parameters. Twenty-six original blind users participated in the early studies. An analysis of 800 Individual Transaction Records indicated average fax transmission times of 68 seconds, and reading time of 163 seconds. Much of the material faxed was handwritten or graphic -- and therefore inaccessible except by a sighted reader. Responses from users of the system indicated strong potential. We were subsequently able to assist our local Easter Seals chapter to establish and enhance their own volunteer fax reader program, transferring the project to real-world use.
Studies were performed on system cost-effectiveness. If the manager's salary was excluded (as it effectively would be in a larger system serving many users), the per-user monthly cost would be $47. This is comparable to (or less than) amounts typically paid by blind persons hiring readers for several hours each week. All users said they would be willing to pay, at some level, for the service they were receiving. (It is also, of course, likely that employers would be willing to defray these costs in lieu of hiring sighted readers.) For another comparison, it was noted that a three-year amortization of equipment and service contracts for a typical synthetic speech reading machine would be in the order of $200 per month.
We also explored non-fax media for "interactive" remote reading systems, employing compressed video technology of very low frame-rate, high-resolution video transmission over ordinary telephone lines. The bi-directional transmission of data would allow the remote reader to pan and zoom the remote camera over the document or object of interest to the user. These functions could be accomplished electronically or by voice command from the reader to the user. Benefits would include increased speed and a closer approach to the level of natural interaction possible with the sighted reader in the same room as the blind consumer. We have been investigating the augmentation of the fax reader service with the option of integrating optical character recognition (OCR) and speech output. Many faxed documents which are long and need to be read word-for-word could be submitted to the OCR. This processing would free up the reader/operator for other customers by having the OCR task done "off-line" -- storing the spoken text at the reader end for later retrieval at the customer's convenience. This possibility is proving to be one of the most promising outcomes of the project.
Overall, it is clear that the fax-based reader service concept has considerable merit due to the low cost and ready availability of fax machines, the "on-call" nature of the service (avoiding the problem of scheduling readers), and the ability to handle unusual types of printed material (handwriting, notes, graphics, etc). Preliminary conclusions indicate that the basic fax machine and sighted reader combination is most useful for blind consumers with high requirements for reading these types of materials, and those for whom sighted help is not readily available. The OCR augmentation has the promise of increasing the appeal of the system to a wider audience, while future telecommunications industry developments such as faster data transmission rates over telephone lines will also increase its potential. A final report on the various outcomes of the present project will be produced in 1996.
In an effort to provide an extremely inexpensive option for computer access by blind persons, we developed a simple screen review program for the blind. This system, "Batrup," was completed in 1990 as the least expensive available means of accessing a computer without sight. It was designed to operate with the "Speech Thing" software speech synthesizer, costing $80, and the only other necessary component was a software disk made available free of charge from the Smith-Kettlewell Eye Research Institute's privately funded Rehabilitation Engineering Service.
The system was beta-tested and modified as necessary to become usable as a basic IBM PC access system for most widely used software. User comments were favorable, and the availability of the system effectively removed the cost barrier to computer access for entry-level blind computer users and/or those requiring a low-cost system for home use.
We have been surveying the field of computerized music in terms of its accessibility to vision-impaired musicians. Our project manager for this task, Jay Williams, is now considered a national source of information on the various aspects of this issue. Our efforts have also given impetus to "networking" among musicians already in the field and those just beginning.
Our collective contacts with some software developers have resulted in more manuals appearing as ASCII files on computer disks. These efforts, and the evolution of computer and synthesizer technology, have provided many more options and rendered this equipment more user-friendly. We have conducted a number of workshops for blind musicians and others in this field. Examples include a workshop for special education classes under the direction of Mike Cole at San Francisco State University, and similar well-attended "hands-on" workshops with other blind musicians, both at the 1993 American Council of the Blind convention in San Francisco, and at an annual music camp operated by the San Francisco Lighthouse. Jay Williams of our staff assisted several students in the "realization" of compositions that would otherwise have exceeded their present performance skills. They were encouraged to do the actual hands-on editing of their music.
We have explored a wide range of hardware and software accessibility problems relating to synthesizers, sequencers and other components of the modern studio. One example of great importance to the education of blind musicians is the aggregate of standard MIDI files that contain classical music compositions, both for solo instruments and ensembles. These provide a convenient method of understanding the division of instruments within a musical score. Such files are found on several bulletin boards. We developed a more systematic approach than presently exists so that the standard repertoire is available. As another example, in the Smith-Kettlewell Technical File, 1992-1993, Vol. 13, No. 1, we reviewed a sequencer called "Band-in-a-Box." This program could be very useful in jazz education programs because of its format and the kind of information it presents on the screen. It provides an accompaniment that closely resembles the improvisation style of a jazz combo, and most of the current screen-access programs can automatically read the names of the chord structures as they appear during the playing of the accompaniment. The program contains many diverse styles and songs and its format is such that the user can construct a new accompaniment very rapidly. With a modicum of practice, even an involved accompaniment, and hence a new song, can be incorporated into its library in a matter of minutes.
We have nearly completed a primer for the novice user of computerized music equipment based on our own experience as well as a great deal of anecdotal evidence gleaned from other blind musicians. It addresses such issues as how to "browse" through a music store when shopping for "menu-driven" equipment; how to gauge the accessibility of a prospective purchase; resources of information concerning computerized music production; and a discussion of the current state of software that converts the synthesizer-playing data into data that generates print or braille music. This primer is available through multiple media including print, ASCII disk and the World Wide Web.
We have conducted a variety of other activities related to computer accessibility. In general, we attempted to avoid duplicating the many efforts of commercial companies and other researchers to solve these problems, while targeting specific applications such as those described in the sections above which were not being addressed by others. We also offered our services as evaluators for early versions of commercial software. For example, we were approached by Berkeley Systems as potential beta testers of their software that would make Windows accessible to vision-impaired persons.
We were approached by TiNi Alloys Inc. to assist in developing an evaluation protocol for a new braille-display-based concept for computer access, in which a single line of volatile braille cells is moved up and down a physical "page" to simulate a full page or screen of braille. We assisted TiNi in developing a suitable test protocol. In the course of this work, we were able to evaluate the ALVA ABT80, one of a number of braille output computer access systems based on the Tiemens braille cell. Similar evaluations were performed on the Provox Screen Reader and the Turbobraille Grade II Braille Translator.
We also explored the use of OS/2 by blind people, as well as several products for accessing Windows (both 3.1 and 95), and engaged in the user groups and forums with people helping each other overcome the steep learning curve.
In the field of the World Wide Web and other Internet features, we have reviewed many Web sites. When one is found that is particularly inaccessible, we write to its Webmaster and advise on the problem and means of solution.
Our overall philosophy in computer and Internet accessibility is to consider the needs of, and include, persons other than the technically proficient "power users."
We undertook feasibility studies aimed towards improving computer, graphics, and print access via the development of new technology for braille and graphics printing. The concept was developed for a novel thermal printing method which offers the potential for fast, silent operation with greatly reduced numbers of moving parts.
A single-cell prototype was developed and tested, and research was performed into component characteristics under the thermal cycling conditions created by this braille printing concept. Initial results were promising, and further development will be pursued when suitable outside funding can be identified to support the significant specialized outside fabrication and development that will be required to produce a full-sized system.
We have explored various improved approaches to specific vocational instrumentation problems where the usual commercial sensory aids are not applicable. One such approach uses a commercial single-board microprocessor in conjunction with our in-house-developed speech. As a first example, a talking counter was developed for Mrs. Kathy Larson of Williston, North Dakota, who has a home business using a knitting machine.
In the Smith-Kettlewell-developed adaptation, a magnetic switch was attached to the frame of the machine while a magnet was attached to the moving shuttle. The movement of the shuttle causes the switch to close once with each row knitted. The switch closure is detected by the counter, which is a separate unit. This counter module contains the microcontroller with its supporting memory, audio processing system, and a 16-key pad.
The counter allows the user to interrogate the number of the last row knitted, as well as to preset memories with row numbers at which some change needs to be made in the pattern. The counter beeps when reaching the row which precedes any preset memory, then announces the row count when the memorized row is reached. The row count as well as the preset memories are saved in battery-backed memory when the machine is turned off so that a project can be continued where it was left without the need for the user to remember a count.
Illustrating the flexibility of this unit, the same counter may be attached to another magnetic switch which is mounted on an exercise bicycle. The counter may be switched to count revolutions of the bicycle wheel, which are interpreted in tenths of a mile. The knitting machine and bicycle memories and other data are separately stored.
Although the market for this particular device is probably very small, this project is exemplary of adaptive devices based on existing technology which can be developed with relative ease to fit a wide variety of applications. The counter uses speech recorded with the "Digital Speech Recorder" developed by our staff for use with Talking Signs and other projects. Most of the computer hardware is contained on a commercial single-board computer. The maker of the series of single-board microcontrollers used here, "Iota Systems," asked our RERC staff to assist in beta testing of a new product line. Our plans include incorporation of this custom instrumentation approach into our privately funded Rehabilitation Engineering Service, to broaden its availability to any rehabilitation client.
We have developed a new multipurpose digital speech recording system for use in a number of talking devices for blind users. The device is intended to record speech digitally into computer memory and to program that speech data into EPROM (erasable programmable read-only memory) chips. Two main uses are intended for speech recorded in this manner:
1. This system is compatible with, and is the standard of recording for, Talking Signs. In applications where many copies of a particular sign -- "men's room" or "exit" -- are needed, the EPROM technology is easily mass copied. This system is also appropriate where field recordability is not desirable, in that the exact contents of a sign need to be permanently recorded and must not be changed by untrained personnel.
2. Devices such as the Knitting Machine Counter and Rolling Ruler use speech recorded on this device.
The need for this device, where many commercial speech digitizing systems exist, is based on its use of standard components. All other speech recording systems use specialized "speech" chips and other components often available from only one source. Since we wish all products we develop to last, and be reproducible, for a number of years, the use of components available from only one source is to be avoided wherever possible. The recorder/programmer described here, along with technology needed to play back speech recorded with it, should be reproducible into the foreseeable future with few or no alterations due to component availability changes.
Another advantage of this system is the simplicity of playback technology. Any microprocessor-controlled system requires only one extra output pin, plus a small audio amplification and filtering board to make its hardware compatible with this system.
Our laboratories have undertaken a study of difficulties experienced by blind and visually impaired people with diabetes. Two problem areas were immediately apparent, resulting in the following developments:
1. The Smith-Kettlewell Insulin Dipstick
Save for keeping a meticulous log of insulin use and reviewing it frequently, there was previously no good way of assessing how much insulin is left in the glass bottle commonly used for dispensing it; being a delicate protein, insulin must not be shaken to listen t
o the rattle of the contents, which would otherwise provide a rough guide to the liquid level. We identified a popular syringe manufacturer, Novolin (a division of Squibb), who makes cartridge-pen type injectors. Two different models allow the dosage to be measured by "clicks." Being cartridge dispensers, the insulin reservoir is a vial containing a rubber piston which is advanced to dispense the insulin. It was a simple matter to design a measuring stick with tactile markings suitable for locating the rubber piston, and there is no chance for the "dipstick" to contaminate the insulin. Each exposed rib on the stick accounts for 30 units of insulin still remaining; a full vial leaves all 5 ribs of the device exposed, indicating that 150 units are available. Finally, one side of the stick is flat to prevent it from rolling.
Samples we have sent to potential users have elicited favorable responses. Possible sources of production are being explored.
2. Glucometer Adaptation to Assist Loading with a Blood Sample
There is a compound problem that makes glucometers hard to use: the blind user has no adequate means to tell when a blood droplet is formed, nor when it is heavy enough to fall from the puncture site. Positioning the puncture site directly over the window and keeping it there while kneading the finger is very difficult, especially when there is no distinguishing feature adjacent to the window in the test strip to aid in locating it. The finger must never touch the test strip; this would smear blood on the active area, changing its rate of absorption and consequently affecting the reading.
The Lifescan One-Touch II is a very popular glucometer among blind consumers, and was chosen for our investigation. The RERC has experimented with various materials in designing an adapter to funnel blood to the test-strip window, and with jigs and fixtures which could serve as guides for proper alignment. A collaborating machinist, James Clawson, developed a prototype called "the Slipper." It accepts the One-Touch II and provides a platform to which various devices can be attached for experimentation.
As an immediate remedy that can be installed by individuals, we have published a paper describing a method of aligning and attaching Dymo-Tape tactile markings. The user takes the punctured finger between thumb and finger of the opposite hand and stands the thumb and finger on the marks either side of the window. Centering the guiding hand with the long marks sets the longitudinal position, and locating the differing marks nearest the edges of the strip establishes the proper lateral position.
As a result of an earlier project, we published the design of an audible modification to the Zircon Stud Sensor (adding a voltage-controlled oscillator and a speaker). Communication was made with Zircon, suggesting that they produce such a modified instrument -- not only for visually impaired users, but because sighted Smith-Kettlewell staff members felt that the additional feedback was helpful. The appearance of a commercial audible model has now come to fruition.
Looking identical to the original Zircon unit, the new instrument is called the Master Mechanic "Professional Stud Finder with Sound." Its audible output consists of a Mallory Sonalert arranged to sound when the top LED of the column of four (the same visual display as the original instrument) is energized.
The audible indicator being a "go/no-go" tone which is either on or off, it can be argued that it is not as sensitive an instrument as the RERC-modified one. However, offering the new Master Mechanic device to those who have tried the VCO output has led to favorable comments about its performance. The Sonalert has been shown to be an inexpensive addition; the Master Mechanic unit costs $20.
This device is intended to fill the need for the equivalent to a long steel tape measure. Technical problems make the direct adaptation of standard tape measures impractical, yet blind people need a way to measure rooms, furniture, fences, shelves, wire/rope, etc.
Commercial ultrasonic "tape measures" with visual displays are available; these devices are impractical for use by the visually impaired for several reasons. Their resolution is quite poor and they must be "aimed" at a target. There is also the common difficulty involving electrical access to display information where no digital output signal is provided.
The device we have developed, dubbed the "Rooller," includes a small box containing a microcontroller and other hardware very similar to that used in the Knitting Machine Counter (see above). The sensor is a so-called "optical shaft encoder" mounted on a wheel which protrudes through the bottom of the box. This encoder produces computer-readable signals showing the direction of rotation of the wheel and the distance traveled. The unit will, for example, be placed at one end of anything to be measured, and rolled in a straight line. If a known piece of material is to be measured to determine its length, the box is merely rolled to the "other end" of the material.
The push of a button tells the Rooller to speak the distance covered. If a specific distance is to be measured along material to be cut to length, that distance may be programmed into a memory from the built-in 16-key pad. As the Rooller is moved closer to the desired measurement, a variable pitch tone will give the user feedback as the desired measurement is neared. When the exact length is reached, the tone will stop and the measurement will be spoken. Should the box be moved too far, the tone will resume and continue rising in pitch. Thus the user will have immediate feedback about his direction of movement and whether the measurement has been reached or passed.
Other features planned are the ability to calculate areas and perimeters from length/width measurements, readout in English and metric units, and storage of several measurements in long-term memory. The unit will be battery operated and should measure about 4-1/2 by 6-1/2 by 2 inches. It is hoped to achieve accuracy within 1/16-inch at distances up to at least 100 feet.
The prototype Rooller is now in the program debugging and development phase, and practical testing is proceeding successfully.
A recurring problem in the field of job access for blind individuals is the fact that talking instruments -- and especially talking electrical multimeters -- have been in and out of production at various times. This has primarily been due to the fact that the various speech chip sets upon which such instruments are based periodically cease being manufactured and are replaced by more sophisticated chip sets requiring different interfacing and controls. A second problem is the fact that commercially available meters intended for blind users have offered either speech output or tonal output -- each of which has limited application in practice.
In view of both the above factors, we embarked on a project to develop a new multimeter for blind users -- the Smith-Kettlewell Talk-&-Tones Multimeter. The goal of this project was to design a meter which combines both speech and total output for maximum advantage in all job situations, and which utilizes "generic" electronic chips for its speech capability. These generic chips are the basic electronic building blocks which have been in existence for many years and will continue to be available indefinitely, solving the problem of intermittent availability of sophisticated speech chip sets.
The new design incorporates a number of novel features not found in conventional "talking" devices -- these features are based on the findings of our staff and students that blind users often have different needs and priorities in instrument use than those of sighted technicians. The speech output system of the Talk-&-Tones has two possible "shortcuts" to lengthy digit pronouncements: a toggle switch can be installed which suppresses all leading zeros; the only numbers read begin with the first non-zero figure. In order to speed up readings, the word "point" is never used for the decimal point; rather, the pitch of the speech is raised for all figures succeeding the decimal point.
Separate from the speech and in addition to it, a VCO-type readout is incorporated. This enables quick probing of a circuit to trace voltages; once the pitch of a voltage is ascertained, all circuit points that give the same pitch are known to have that voltage present. With the speech set to make recurring readings, adjustments are greatly facilitated by using the VCO signal as a qualitative indication as to how much change is being effected.
Any intermittent operation of a circuit caused by "cold solder joints" or faulty wiring are immediately apparent when listening to the VCO, and unlike with speech, the technician knows exactly when the intermittency occurred (correlating it with jiggling a wire, for example). The VCO readout is also provided with a calibrated braille dial. This dial can be set to a predetermined value so that, when the VCO indicates above this setting, its tone is "chopped" (interrupted) at a rate of 15 times per second. Thus, dangerous current values and voltage boundaries can be detected immediately, which is not possible with speech. Either of the readout systems can be used independently. If both are used, a 10dB attenuator reduces the sound level of the VCO while the instrument is talking.
The resulting instrument has been evaluated successfully, and its design made widely available to other blind technicians.
The digital multimeter is a ubiquitous tool for electricians, electronics technicians, and service/technical personnel in many industries. It measures electrical parameters such as voltage, current, and resistance, usually presenting results on a visual display. Over the years, any number of devices of this type have been made to "talk" for use by the blind. Some of these have been modified commercial meters with expensive speech boards added. Others were custom-designed units built specifically to talk. Very few of these have lasted on the market for long, and most have cost $500 or more.
Recently, Radio Shack has sold, first, an actual talking meter which is now discontinued, and now sells an inexpensive meter with a standard RS232C computer interface. The talking meter was replaced in Radio Shack's product line with the Micronta No. 21182 Digital Multimeter with computer interface. It has 32 ranges/scales, and measures voltage, current resistance, capacitance and frequency. It can also test transistors and other components. The standard RS232C computer interface is made via a cable plugged into the side of the meter and into the serial port of any IBM-compatible computer. The unit is supplied with a diskette containing example programs which show the entire contents of the meter's display on the computer screen. The unit can be controlled from the computer to make timed measurements, and results can be saved to disk or printed from the computer. Although one of the supplied demonstration programs is graphically based and therefore not usable by the blind, another program is text-based and works with a variety of access devices, both speech and braille. The demonstration programs are provided in the basic computer language, making adaptation to special needs relatively straightforward. Our staff adapted the software that came with the meter so that it presents a less cluttered screen. We also explored several other ways of making this unit more useful to blind technicians; information on these options is available by contacting the RERC office.
The availability of high-quality adapted levels for use by blind artisans has been subject to frequent interruptions, primarily due to limited product life of commercial products on which various designs were based. To make matters worse, specially designed units from Germany and England, evaluated by the RERC, were found to be inadequate -- either because of poor damping and/or insufficient tilt sensitivity.
Since recent development of "magneto-resistive semiconductors" has led to a new class of inexpensive inclinometers, we decided to design an audible electronic level from scratch. This would eliminate vulnerability caused by the product life of commercial visual instruments. The design is adaptable to a wide variety of tilt sensors, so that very little work will be needed in the future as the sensors themselves become obsolete.
An aluminum Stanley level was chosen as a convenient straight edge. A small round can, cemented in a cutout of the level's beam, contains the inclinometer. Made by the Midori company, the inclinometers come in various grades and sensitivities; for approximately $100, a sensor capable of repeatedly sensing tilt within 3 minutes of tilt can be obtained -- this compared to a 30-minute accuracy available with the British product referred to above.
The audible output is that used by the Smith-Kettlewell "Dynamic Meter Reader" which was created in 1982 and used in myriad instruments since -- including pressure and vacuum gauges and the RERC's blood-pressure meter. In this carpenter's level, a voltage-controlled oscillator responds to the incline of the instrument's beam; this gives the user feedback about the direction and degree of movement as adjustment is made to the member being aligned. To one side of "level," the tone of the oscillator is "chopped" (interrupted) approximately 5 times per second so that the tone is pulsating; an opposite incline produces a smooth uninterrupted tone.
Evaluation of the prototype by blind and sighted users has revealed that the unit is easy to use and achieves accuracy exceeding that obtainable with a conventional bubble-type level designed for sighted carpenters. The level is now available through our privately funded Rehabilitation Engineering Service.
From time to time, people in the field of low vision (especially educators and large-type transcribing services) have requested that we modify or redesign the 3M "Scotch" brand EA200 tape embosser. This machine uses 3/4" adhesive-backed plastic tape (similar to Dymo tape) and embosses the tape with 1/2" high sans serif lettering. Labels made with this machine are very clear and can even be read tactually. However, index markings on the embossing wheel are indistinguishable by touch and are too small to be read by persons with low vision. The common complaint is that, while blind and low vision people can read its output, operation of the machine is inaccessible.
We therefore engaged in a series of experiments to design modifications which would make this device more accessible. One modification was to improve letter visibility by removing the sliding indexer and provide the wheel with large index lettering using an overlay. We were able to modify the machine so that the right front corner of the tape could be felt as the user positions it for making the next label. We designed a spring steel support plate which, much like the "paper table" in a typewriter, supports the tape as it travels through the machine. With the embossing wheel shifted into the alphabet position, the edge of this plate can be felt through a hole in the right side of the machine. A half-round notch was filed in the edge of this plate which allows the finger to reach the right edge of the tape. Preliminary evaluations indicated that the above methods can facilitate use of the 3M EA200 dymo labeler machine by users with impaired vision. Duplicates can be made available from our Rehabilitation Engineering Service.
Blind employees sometime encounter situations where lights need to be detected which are unusually small and/or low in intensity. Such a case was encountered by Mr. Donald Belew, a blind employee at the Lawrence Berkeley Laboratories. As part of his duties he was required to operate a 16-line telephone system in the Engineering Science laboratory. This telephone presented the specific problem that its indicator lamps are much too small and dim to be detected with modern photo transistors and photo diodes as used in conventional light probes for blind persons.
For this application, we developed a design using a cadmium-sulfide photo resistor; this type of sensor, while slow compared with modern semiconductors, is extremely sensitive and responds to shorter wavelengths of light than do silicon devices. Certain earlier instruments, notably the Science Products "Audicator," were similar in concept, but improvements including the 3cm "light shield" are found to be necessary on more modern telephones in order to resolve closely spaced lamps. The new design is being tested by Mr. Belew, whose initial evaluation is extremely positive.
We have been monitoring developments in machinery, used by high-volume vendors, which reads the "Universal Product Code" (UPC). There is much interest in designing or adapting equipment with which blind shoppers and vendors could identify store goods. The major obstacle to rendering UPC codes useful to blind persons (without access to the large table-mounted scanners used in supermarkets) is being able to locate the position of the bar code on each specific container in order to read it with an inexpensive scanner.
Experiments were conducted using audible reflectance indicators which might allow a simple device to be designed providing feedback to the user as to whether or not the UPC has been located. Variations in reflectance, detected by a single sensor, led to frequent confusion between closely spaced lines of small print and the bar code itself. A pair of sensors (with adjustable spacing) was then tried to determine whether coincident leading or trailing edges of standard UPCs could lead to more definite detection. The results of these simple approaches were not sufficiently encouraging to warrant further exploration.
As an alternative approach, we decided to try the Optacon (Optical-to-Tactile Converter), as a display of acknowledged integrity, to determine whether finding the UPC bar code could be done using direct presentation of its form. Subjects were given nine food packages of various sizes and shapes and asked to find the bar codes. Three of these sequences were presented and these trials were timed. The results of this experiment were inconclusive; the task was somewhat frustrating to the subjects. Subsequently, a user survey based on information gathered therein was performed, to answer questions on usefulness and desirability of a suitable bar code labeling and reading system were it to become technically feasible.
As a result of these studies, we assembled and tested a new generation of bar-code reading equipment and software. The "Christina" by Smith-Kettlewell is a bar-code system for user-customized label information. It uses a commercial bar-code reading device to feed information to a computer. When presented with a bar code, the instrument sends an ASCII alphanumeric string of data to the serial port of an IBM PC. Smith-Kettlewell has created a data-storage program into which the user can then enter pertinent information associated with each bar code. Thereafter, every time the system is presented with a known bar code, information that has been previously entered appears on the screen. In other words, a talking, large-print, or braille computer, with the attachment of this bar-code reader, will ask the user to enter data associated with each new code. Whether recipes or nutritional information are then typed in, each time the computer sees that particular code, the user's notes will be presented.
The SP*Ace bar-code reader upon which the system is based is available for approximately $1,300 from Spectra-Physics Scanning Systems, Inc. of Eugene, Oregon, tel. 1-800-547-2507. The "Christina by Smith-Kettlewell" program is on a diskette which is available, free of charge, from our RERC. Call (415) 561-1677. As the cost of these more sophisticated bar-code reading wands comes down, we believe commercially available systems adapted for use by blind persons will become feasible.
[We wish to acknowledge Christina Baer, and her father, Dr. Tom Baer, for recognizing the flexibility of the Spectra-Physics SP*Ace bar-code reader, and for their prototype system that captured the attention of engineers interested in this project. At the time of the Baers' work, Christina was a 4th-grade student, and she submitted this project for school.]
We have conducted a series of experiments to determine the feasibility of utilizing the new generation of "stepper motors" for various kinds of tactile outputs for the blind and deaf-blind. Laser milling and new winding configurations have given rise to devices whose resolution is extremely fine -- 800 steps per revolution as compared to 48 steps. Furthermore, new integrated circuits for driving stepper motors support a so-called "half-step mode" with which the number of steps can be doubled (with the sacrifice of uniformity in "holding torque").
We are investigating these new devices for their utility in creating "gearless" mechanical clocks and durable moving-needle displays. (An instrument which is often sought after is a quiet mechanical clock for use by blind technicians in broadcasting and recording industries -- situations where talking clocks cannot be tolerated and where braille output is prohibitively expensive.) It is hoped that with the avoidance of gear trains required for more conventional motors, less expensive instruments can be realized. These instruments may incorporate microprocessor control, as described in "Specialized Job Adaptations," (Section F above) on a specific proposed voltmeter. We have published an article in the Smith-Kettlewell Technical File on this subject to determine the response from consumers regarding interest in this approach.
The Smith-Kettlewell Auditory Compass, utilizing a Hall Effect sensor, was a useful aid for blind persons; in contrast to locking-type braille compasses, this instrument provided the user with dynamic auditory feedback for determining direction and relative degree of veering from a desired path of travel.
The use of the Hall Effect sensor had drawbacks in size (9 inches long) and the cost of the sensor (about $100). We therefore redesigned the original compass using a flux-gate device. The flux-gate sensor dimensions are 1-1/2 by 1-1/2 by 3/8 inches, and its cost is about $25. It can, because of its size, be mounted in the same enclosure as the circuitry for the compass. The enclosure we have used measures 4 x 3 x 1-1/4 inches.
In using this compass, the sensor (or, in this version, the box with the sensor in it) is oriented so as to achieve a null of the auditory signal; this occurs when the sensitive axis of the box is perpendicular to the earth's magnetic field; i.e., east-west. Rotating the unit in one direction causes a "beep-beep" signal, while rotating it in the opposite direction produces a "ding-ding" sound. One end can then be arbitrarily marked with tape so as to assign this end with a particular direction (our unit "beeps" when the tape end approaches north).
We designed and fabricated a printed circuit board for the new compass, and an article describing the construction of the entire compass has been published. The cost of parts for the compass in small quantities is under $65. Several units have been distributed for evaluation by blind users. Near the end of the reporting period, we began work on a further improved design based on user feedback and availability of improved fluxgate technology.
Mr. William Loughborough of our staff has been working with two blind entrepreneurs to provide the necessary technical assistance for establishment of an FM radio station in Goldendale, Washington. The radio stations (KYYT-FM, KLCK-AM) are beginning to automate their daily logs so that they can be operated by computer programs. The main advantage of this is that the blind operators will have more time to devote to community interaction and local coverage.
A key requirement is the ability to record repetitive announcements in hard disk files for on-air access via a scheduling program. Most of the software interfaces are graphical and must be modified for use by the blind announcers.
The operators use a braille embosser to make scripts of commercials and daily log sheets, making these operations simpler than the previous method involving audio tape. The technical changes require replacing tape cartridge players with hard drive files, routine station breaks with indexed files, and satellite music systems with programmable CD players.
When these steps are complete we will have a model for small town broadcasting that will be of benefit to other blind entrepreneurs wishing to establish similar businesses. A manual covering the necessary steps and technical aspects will be produced. An additional benefit is the facilitation of de-urbanization of people with disabilities and consequent enrichment of small communities with what such individuals have to offer.
Dr. Tim Cranmer, Technology Committee Chairman of the National Foundation of the Blind and a member of our Consumer Advisory Board, has often remarked that an important tool that blind people lack is a pencil that can be used to leave notes for sighted readers. There have been no "low-tech" solutions, and most proposals have involved using large machines to make pieces of paper with material printed on them. What is really needed is a portable device that can be used to write on existing documents, rather than producing yet another piece of paper with notation.
We are exploring the use of the type of print head used in ink jet printers. This technology is compact, consumes little power, and lends itself to manual operation. Traditional designs have focused on operation within the framework of mechanically driven print mechanisms. By substituting manual manipulation and decision making by the user, it may be possible to make a small device that enables one to print "by hand" on existing documents.
There may be other commercial applications of such a device in the sighted world; therefore it is anticipated that cooperation from major manufacturers will be obtainable for this project. Initially, we are collaborating with an engineer at Hewlett-Packard who is making a simple version -- akin to a ball point pen with line width control. Another collaborator at Jet Propulsion Lab is designing a one-hand braille input system. These studies are at the early feasibility phase to determine whether a more formal effort is warranted.
We regularly receive requests from a variety of sources for help in making either jobs or aspects of the environment accessible. It is impractical to list here all the cases in which our staff's advice was utilized in improving accommodations and accessibility. An illustrative example, however, was a request from Kathy Quo of Booz Allen & Hamilton, Inc., a San Francisco management consulting firm, to advise the engineers at Morrison Knudsen concerning specifications for the door-open and door-close signals to be incorporated into Caltrain's new railroad cars. We advised them that in order for these sounds to be easily located by a blind traveler, they should conform to the following specifications, based on our experience with auditory displays for the blind:
1. They should be broadband sounds, most of whose energy occurs within a band of approximately 200 Hz to 1,000 Hz.
2. The onset of the sound -- the time it takes to go from silence to its full intensity -- should be no greater than 600 microseconds. (This timing is approximately equivalent to the time it takes for sound to travel from one ear to the other.)
3. Because some commuters may be riding in these cars for as much as three or four hours, the sound quality of these signals must be such as to produce as little annoyance as is practical.
In consulting with Ms. Quo and the Morrison Knudsen engineers, a sequence of sounds was generated using the Kawai K5 and the Roland SC55 music synthesizers. A PC-based music sequencer was employed so that the order in which each sound occurred, as well as the length of time required for each sound to decay, could be varied predictably. The engineers chose several possible sound options which were then recorded onto high quality tape for evaluation and subsequent installation.
At the request of the Office of Special Education and Rehabilitative Services, we carried out a technical study under separate funding regarding "Descriptive Video Services" (DVS, a "service mark" of WGBH Boston). Our report was companion to one on "commercial viability"by COSMOS Corporation in Washington, D.C. The Smith-Kettlewell study and report covered technical questions about how DVS is accomplished, and why it is a more technically viable service now than it would have been before 1984. The report was completed in June 1990 and submitted to OSERS for review by the Congressional Office of Technology Assessment and others, to assist them in determining the future of DVS. Copies of the report are now available from the Smith-Kettlewell Institute.