Computing for the handicapped. (Voice Input Module) Shel Talmy.
Full Text: COPYRIGHT Ziff-Davis Pub. Co. 1984
"Welcome to the world of VIM, your third hand." That is how the VIM (Voice Input Module) introduces itself when booted up on the Apple IIe. This was the beginning of a demonstration I was given at the offices of Voice Machine Communications, Inc. in Santa Ana, CA.
Arnold Balliet, one of the members of the design team for the VIM put the product through its paces. Conclusion: this could be the beginning of another revolution, perhaps as big as microcomputers themselves.
The VIM is an exciting product because it represents the leading edge of technology. Don't misunderstand, the VIM works fine today and performs well enough to make life easier for many people. The technology is still embryonic, and it is the future of VIM and other voice recognition systems that is exhilarating. The way things are going, when new and improved are the norm, yesterday's product is on a par with the Model T, and day before yesterday's is antediluvian, that ultimate system is probably not too far off.
We are rapidly approaching a time when entire households will be operated by voice alone. That means the old standby, our favorite hackneyed phrase, user-friendly, will truly be that. Which leads us to conclude that the next step will not be far behind: your computer will talk back to you. Shades of HAL from 2001--hopefully with a more benevolent attitude.
Enough crystal ball gazing, let me tell you about VIM as it exists today. Voice Machine Communications may have an edge on its competitors. The president of the company, Ron Runge, was the developer of the chip for the voice recognition system and, along with his design team, is continuing to improve it. The latest version is the SS-VIM which has doubled the memory size of the previous model. This allows space for 172 words to be programmed into the device at any one time. This amount of vocabulary is more than enough to run any existing program by voice and still have plenty of room left over. Since you can create as many vocabulary programs as you like there is no limit to the number of functions that VIM can perform.
VIM comes with several pre-programmed vocabularies for such standard programs as WordStar, VisiCalc, and Apple Basic, plus Black Jack and a maze game. The beauty of this is that the computer novice literally can sit down and start running application programs at once. There is no need to memorize dozens of control characters, or puzzle out an English translation of the manual. With VisiCalc, for example, you can move the cursor by voice to any column, make entries, split the screen, scroll up and down, and do any other spreadsheet functions. Getting Started
The VIM is connected in parallel with the keyboard, which is handy because you can use a combination of speech and keystrokes whenever it suits you. The first thing you want to run is the Apple Voice Input Module program (AVIM) which presents you with a menu that includes options to build a new vocabulary, train an existing one, or test for recognition. The "train" option, for example, loads an existing vocabulary to the VIM board.
You are then asked to repeat the words into the microphone so that the VIM will recognize your voice. The wave forms are stored on disk for retrieval. You are asked to repeat each word three times and usually this is enough for the VIM to acknowledge your speech pattern. If not, the clever little devil will ask you to repeat the words until it is satisfied. I had to repeat a couple of the words about ten times, giving credence to my mother's complaint that I mumble.
The VIM will also examine words that might sound alike, such as bad and dad, give you a numerical recognition factor between the two words it has compared and then ask for further training if the words sound too similar, (VMC claims a 98% plus recognition success rate.) When this is completed and both you and the VIM are satisfied with each other, you are ready to run programs by the sound of your voice. This training procedure takes just a few minutes.
Let's examine some of the applications. Imagine a high volume billing operation or an auto accessory company with thousands of parts in its catalogue. The operator has his or her hands free to find the items wanted while making the requisite entries by voice. It would be difficult to put a percentage on it, but the efficiency quotient would have to be substantially improved.
And what does all this have to do with the handicapped user? Read on. CASH
Cash is a word that we all know and love; the dictionary defines it as "money that a person has, especially ready money." In this case CASH is an acronym for Computer Aided System for the Handicapped. The acronym is intentional. The system is the means by which the handicapped user can go back to work and do almost anything that can be accomplished by the averase user.
The CASH board has some special functions for the physically disabled. One is the ability to reset the computer vocally instead of with two keystrokes which are beyond the capability of some severely handicapped people. The other is the capacity to make a good job of running the entire household by voice.
The Apple requires a 64K extension card. This is where the "home controls" reside. It works this way. Let's suppose you are using WordStar with the VIM to compose a document and the phone rings. You speak the word interrupt, which shifts control to your "other functions" area, where you can answer the phone with a spoken command. Don't worry about your document, WordStar will be ready to pick up where you left off when you have finished with the call.
Let's say that during the conversation, you decide that the temperature in the room is too low and you want to boost it a couple of nothces. Just excuse yourself for a moment and give the appropriate command.
The VIM interfaces with the standard BSR module and other similar units. You can turn lights and tape recorders on and off, dial a telephone, raise or lower a bed, and just about anything else that can be accomplished through the magic of electricity. Also, any standard voice synthesizer can be connected to the system for use by a visually impaired operator.
At this writing, the system is going through its final testing and should be on the market by the time you read this. Certainly, this is a great step forward for the handicapped. It provides a measure of freedom and independence that was heretofore unavailable as a complete package. Voice Design
One other function of the VIM that I haven't yet covered is its use with graphics. Drawing by voice was a real revelation, especially for me who has trouble managing a straight line with a ruler. The program I used was Delta Drawing from Spinnaker, and it was simplicity itself using vocal commands to draw sets of triangles and fill them with color at a word.
VMC's parent company is KTI Cascade, who among other things sell a line of CAD/CAM software (computer aided design and manufacture). With the VIM added to the system, I am told it is now a snap to accomplish many of the difficult functions required for technical design.
To sum it all up, VIM has solid applications, for the handicapped and commercial markets. The bad news is that the VIM will interface only with the Apple, and there are no plans to make it compatible with other systems. This is a regrettable state of affairs for the multitude of CP/M and IBM PC users. I, for example, would like to use the VIM with my Compu-Pro, as I use word processing, spreadsheet and database programs. Alas, this is not to be, at least for the present.
The standard VIM with an 80-word vocabulary, microphone and software costs $920 for the Apple II Plus and $995 for the Apple IIe.
There is no price as yet for the SS-VIM with its 172-word vocabulary. However, it is expected to be approximately 50% higher than the standard VIM. Voice Machine Communications, Inc. is located at 1000 South Grand Ave., Santa Ana, CA 92705 (714) 639-6150.
The CASH system comes with four boards. The Opto-Relay board is for hard-wired items such as a book page turner or the operation of a bed. It also includes a telephone dialer. There is a 16-channel I/O board, a controller for the BSR type module, and a real time clock. The system is expected to sell for approximately $3200. The sole distributor for the CASH system is Freedom Design, Inc. located at 1884 Eastman Ave., Ventura, CA 93003. (805) 654-8221.