Web Posted on: February 24, 1998
DESIGNING ACCESSIBLE COMPUTER SYSTEMS FOR THE BLIND WITHOUT EXPENSIVE ADAPTIVE HARDWARE
David M. Tanner - Rehab Tech Specialist
Technology Resource Center
State Services For The Blind
2200 University Avenue West
Suite 240
St. Paul MN 55114
(612) 642-0795
dtanner@ssb.state.mn.us
http://www.visi.com/~dtanner
INTRODUCTION
In the past one of the biggest problems that the blind have had when it came to education, employment, and personal enrichment was that the costs of adaptive hardware and software were so expensive that the average individual could not afford to purchase the equipment and software that could make them more independent in seeking employment, pursuing higher education, or expanding their knowledge through reading of books that were only available in print. In addition, for a variety of reasons, very few blind persons were able to present a convincing enough case to agencies or other sponsoring organizations to gain their support in purchasing these expensive adaptive devices and software.
This paper deals with recent developments in the computer technology used by the general public, and developments in technologies in the adaptive technology field that have decreased the price of adaptive computer systems drastically, and take a look at what is now available to bring these prices down. We will also discuss trends that will appear to be heading toward even less expensive systems in the future, and look at some of the considerations that must be made when configuring systems with this new technology.
COMPUTER HARDWARE
Just a couple of years ago it was necessary to spend at least $2,000 or more to purchase a computer system that could support the special technologies that would be necessary for a blind person to be able to effectively use a computer system to do productive work. In 1995 the cost of most computer systems running fast processors, large hard drives, and plenty of room for adding expansion cards would cost $2,000 or more. Bottom of the line systems that came in at prices under $2,000 usually did not have fast enough processor speeds, or were to limited on expansion capabilities to be usable.
By comparison, today a large number of computer producers such as Compaq, Hewlett-Packard, Digital, and others have released a full line of computer systems that sport the latest in computer technology for prices in a range from $800 to $1,500, and have all the capacity needed by most blind computer users for education, employment or personal enrichment. These systems now contain multi media soundcards, modems, large hard drives, fast processors, and enough memory to handle the demands of the operating system, the computer application software, and the adaptive software.
SPEECH OUTPUT HARDWARE
Since the early days of access to computer access technology the most commonly used adaptive technology has been the use of speech software in combination with a hardware speech synthesizer. Most of the hardware speech synthesizers either required a expansion slot inside the computer or were attached to a serial port. The synthesizer often made configuring the system more difficult, and added $400 to $1,500 to the price of setting up a system. Add to this price the cost of the speech access software; which could run $500 to $2,500, and just adding speech access to a system might add a staggering $3,000 to the price of the system.
Today the use of multi media soundcards by Windows multi media software and Windows itself has made the soundcard a standard item included in almost any new computer that a person might buy off the shelf at the local discount computer store. And, the development of standards for producing sounds and speech output through these multi media soundcards is becoming much more sophisticated and effective at economical prices.
These developments have meant that instead of purchasing a hardware speech synthesizer that may cost from $1,200 to $2,500 to obtain very human sounding speech it is now possible to purchase a computer system that already contains a multi media soundcard and a software synthesizer package that costs in the range of $100 to $250 to obtain that same human sounding speech quality. In fact, software speech synthesizers will soon be shipping that actually give digitized human speech output in this same price range. For a list of some of the currently available software synthesizers and availability, see the last page of this paper under SOFTWARE SYNTHESIZER RESOURCES. Additionally, it should be noted that Microsoft is currently developing a basic speech software with software synthesizer output for use in the next version of Windows NT, and at least two companies have developed speech software packages for Windows '95 that include their own software synthesizers in the speech access package. See SPEECH ACCESS PACKAGES With Speech Engines on the final page of this paper.
Other vendors of specially designed software packages have also included software synthesizers in their packages. Examples are Kurzweil1000 from Kurzweil Educational Systems Inc. This ocr software incorporates as a part of the package the Flextalk software synthesizer developed by AT&T. PW Webspeak from the Productivity Works Inc. includes the software synthesizer SoftVoice in its' talking web browser. And, Zoomtext, a large print software for Windows incorporates their own software synthesizer into their software to allow the user to have both large print and speech output with no special hardware required.
CONSIDERATIONS WHEN USING SOFTWARE SYNTHESIZERS
While the use of the software synthesizer and multi media soundcard has drastically dropped the cost of speech access to computer output; there are some characteristics of the various software synthesizer packages and their interfaces with the speech software to be used that will need to be considered. It will also be necessary to determine how the computer system is going to be used, and whether there will be any need to use the soundcard for other functions other than the production of speech output. For example, if the blind computer user plans on using their system to access audio broadcasts of any kind from the internet there may be problems with attempting to use the multi media soundcard as a speech synthesizer because the audio software that will allow the user to hear broadcasts from the internet will take control of the soundcard and the user will either loose speech while the audio is playing or speech may be interrupted and choppy. One of the current software synthesizers takes complete control of the soundcard and would not allow use of multi media audio products such as RealAudio for listening to broadcasts from the internet or to even hear Windows system sounds.
MULTI MEDIA AND OPTICAL CHARACTER RECOGNITION
In the past if a blind person needed access to printed materials it was commonly expected that a desktop scanner that might cost $500 to $1000 or more would need to be purchased. In addition it would be necessary to purchase a specially designed optical character recognition software designed to produce speech output that a speech synthesizer could speak. This specially designed software could cost as much as $1,000 to $3,500. But, with developments in the computer field that let a person use a desktop scanner to scan a document and fax it to another person, scan a picture and store it on a computer for later viewing; the price of the desktop scanner has dropped to an amazing low price of $99 to $300 for an excellent quality desktop scanner that can also be used by the blind person for optical character recognition. Development of off-the-shelf ocr packages for the general public has also had some effect on greatly improving the accuracy of the ocr technology, and has made it possible to purchase off-the-shelf ocr packages that work to a limited extent with speech output for less than $100.
Again, because the general public computer user doesn't want his or her computer to require them to have to dig inside their computer to install interface cards the majority of new desktop scanners now allow the user to simply plug their scanner into the printer port on their computer and saves the necessity of needing additional expansion slots inside the computer for a scanner interface card. While this type of interface has often been thought to be a slower interface, we have found that some of the newest products do not sacrifice quality of recognition or speed of retrieving the image data from the scanner. Added to this the speed of the processors in new computer systems and scanning and recognition of printed documents has never been cheaper, faster or more accurate.
VOICE INPUT SYSTEMS
While voice input is still not an exact science that is widely used by blind users; more and more sighted computer users are using these systems. As these voice input systems drop in price to a basic system cost of $100 to $250 it would appear that within the next 6 months to a year it will be very possible to add voice input to a computer system with speech output at little or no additional costs as a number of software producers such as Corell and Lotus/IBM add voice recognition to their application software. As short a time as 2 years ago these same voice recognition software packages were not nearly as accurate and had base prices of $1,000 to $3,000.
WHERE ARE WE HEADED?
The future looks bright for continuing to reduce the price of the accessible computer system for the blind. Multi media developments will soon mean that the blind user will be able to have digitally produced human speech output from their computer for prices under $100, voice input for no additional charge, and a choice of optical character recognition software packages from the off-the-shelf product for under $100 to a sophisticated package that will do everything from ocr to file type conversion, reading of faxes, reading of text in graphic file formats, and much more. The general public is demanding more and better multi media, and the side effects of these developments will continue to decrease the price of the technology, hardware or software, necessary in the computer system to produce speech, recognize text, and recognize voice input. Faster and faster processors in computer systems, and multiple processors will allow access software packages to process information faster and deliver the output to the user faster and more accurately.
SOFTWARE SYNTHESIZER RESOURCES
Keynote Gold Multi Media -- quality speech usable in Windows 3.1 and Windows '95. Available from Humanware Inc. (916) 652-7253, List Price $250.
Flextalk - MSA/SAPI synthesizer for Windows '95 and Windows NT - for information call Kurzweil Educational Systems Inc. (800) 894- 5374 Price depends on whether purchasing an end user license or developer package. User license $100.
Dectalk Access 32 - copy protection tied to the speech software being used and usually must be purchased from the same company where purchasing the speech access software. List prices range from $150 to $200. GW Micro Inc. (219) 489-3671, Henter-Joyce Inc. (800) 336-5658, Synthavoice Computers Inc. (905) 662-0565.
SPEECH ACCES PACKAGES WITH SPEECH ENGINE
Kurzweil 1000, ocr software with Flextalk software synthesizer built-in, ability to scan and read documents with large print and speech output, Kurzweil Educational Systems (800) 894-5374, List Price $1,295.
PW Webspeak - talking web browser with SoftVoice software synthesizer, The Productivity Works Inc. (609) 984-8044 $50 for browser plus $65 for SoftVoice.
Winkline Reader - Windows '95 speech software with built-in software synthesizer for basic Windows '95 access, Speech Systems For The Blind (508) 226-0447, List price $50.