Ever since our generation saw Stanley Kubrick’s 2001: A Space Odyssey, having the ability to communicate with a computer by voice has been the Holy Grail of computer interfaces. In that film , HAL, the spaceship’s computer, was not only able to understand continuous speech but was also able to read lips.
Although we are approaching the title date of the film, that level of technological sophistication is still out of reach. To be able to converse with a machine, as astronaut David Bowman did, has created a tremendous challenge for software designers, as the variances found in human speech do not readily lend themselves to the discrete input requirements of today’s computers. We have a computer that can play an expert game of chess, but having one that is able to deal with the nuances of the human voice has been a more difficult challenge.
This is not to say that voice recognition software has been unavailable. There have been several products of this type on the market for some time, however, it has only been recently that the effectiveness and price of these products have put them within the practical reach of the average consumer. Currently, for the Windows platform, there are at least three companies making products that will handle various voice related operations ranging from navigation to dictation. IBM, Dragonware, and Kurzweil make professional level applications that allow users to use their computers hands-free and to dictate textual and numeric content without using the keyboard. The primary drawback for the casual user has been the high cost of these applications. Each of them is in the $300-$700 category.
IBM was the first to release a less expensive version of their software. VoiceType: Simply Speaking is available for about $75 and contains much of the functionality of its big brother. This version of the software supports "discrete" speaking which means that the user must pause slightly between every word to allow the software to extract the individual words from the flow of speech. The system requirements are a Pentium-100, 16 MB RAM, a soundcard, speakers, and Windows 95.
A newer version of the software, ViaVoice, was just released in August and supports continuous speech. Priced at $99, it has a steeper set of hardware requirements: 150 MHz MMX Pentium, or 166 MHz Pentium, 32 MB RAM, Windows 95, a soundcard and speakers.
For this review I tested VoiceType: Simply Speaking using a Pentium-90 with 16 MB RAM.
Installation of the software was quite easy as the installer on the CD-ROM took care of everything. The software also comes with a headset-mounted microphone and an adapter that is required by some soundcards. The microphone plugs into the "mic" jack on the computer’s soundcard. The cable is quite long and I was able to stand and move around while wearing the headset. There is an "on/mute" switch on the cable that allows the user to turn off the microphone to prevent unwanted conversation or noise from being detected without having to pause the software. After installation is completed, it is necessary to a perform a microphone test. This ensures that the input level of the microphone is adequate and that the background noise level is measured and accounted for.
Once running, the software is unlikely to recognize a new user’s speech, so the next step of the process is to train the software with the user’s voice, a procedure called "enrollment." There are two levels of enrollment that the user can do. The first one takes about 30 minutes and will give a moderate level of voice recognition accuracy. In order for the software to really do an accurate job, 95% or better, it is recommended that the user complete the second level of enrollment, a process which takes about 60 to 90 minutes.
The enrollment process entails reading a series of sentences to the computer. The words are shown in a dialogue box, with a small rectangle appearing around the word to be read. As the user speaks the indicated word the box moves on to the next word. Words that are misunderstood are highlighted in red, while the understood words remain in black. At the end of each sentence the user has the opportunity to say the sentence over, correcting the misunderstood words, or move on to the next sentence. The content of the text introduces the user to the workings and features of the software. When all of the sentences are completed, the software builds a profile of the user. On the test machine this process took about 40 minutes.
After completing the first enrollment, I tested the software’s accuracy and was impressed that it was able to understand most of what I was saying. This encouraged me to complete the second enrollment. This entailed reading several more passages of the online manual followed by the complete text of Mark Twain’s Ghost Story. This took about 90 minutes to complete while the building of the user profile took about two hours on the test machine. It was worth the wait, as the accuracy seems to be at least 95%.
After spending all of that time reading to the computer I began to wonder who was training who. I found that as I used the software in subsequent dictation sessions I could increase the recognition accuracy if I pronounced the words and spoke in a pattern as I had during the enrollment sessions. The software shows the words that it "guesses" as they are spoken, sometimes incorrectly. This is not a problem because the system is not making just a word for word match, but looks at the context of the sentence and changes the text accordingly. Quite a marvel to watch as this happens.
If the software makes a mistake, the dictation process can be paused by saying "Stop-Dictation." Double-clicking on the incorrect word opens a small menu showing the incorrect word and a set of words that are its second guesses. A recording of the word, as spoken by the user, can be heard from the speakers. The correct word can be selected from the list or entered by use of the keyboard. On my initial outing, the software did not know the word "Missoula," instead substituting "Missouri." After "teaching" VoiceType the correct word it has used it accurately every time. The correction feature can also be used to enter non-existent or technical words.
It is possible to dictate formatting instructions as well as normal text. For example, to insert a quotation, the user says, "OPEN-QUOTE." To end the quotation say, "CLOSE-QUOTE." Other formatting commands include "ALL-CAPS," "CAPITAL-LETTER," "COMMA," "PERIOD," "NEW-LINE," "NEW-PARAGRAPH," and "SPACE-BAR." Although I have not tested it, the software does support the use of Windows commands, allowing the user to navigate through menus and perform routine desktop chores using voice rather than the keyboard or a pointing device.
The software does not come with a printed manual, reflecting a current market trend. All documentation is available using a standard Windows help file within the application which I found to be complete and easy to use.
While this software is not perfect, it is certainly a glimpse into the very near future. For someone who is a good touch-typist, it might be frustrating and less productive to have to correct the system’s errors. For someone like myself who has never learned to type (do I dare admit that?) this holds a great promise. I used the software to dictate my initial draft of this article, which I found to be quite handy, though I did several rewrites before the final version was complete, something that I do anyway. As I tend to speak in a style that is different from the way I write, some degree of further training will probably be required on my part. On the test system, I was able to dictate at a rate of about 75 words per minute without overloading the computer.
My guess is that we are 3 to 5 years away from seeing a widespread deployment of this technology. I look forward to some very practical uses of this in the classroom. For example, I would like to be able to wear a wireless microphone and have my lectures "recorded" as text that I could make immediately available via a web page. This would be a boon to students who miss a class or are home-bound for other reasons.
This technology also holds great promise to those with physical disabilities that prevent them from using the standard keyboard or other input devices. Students who have a difficult time transferring their thoughts to paper or have trouble with a keyboard could learn to dictate to the computer and eliminate the need for someone to transcribe their work.
Back at it again. This marks the beginning of my 29th year of teaching. The bad news is that I must be getting older. The good news is that I enjoy what I am doing now more than ever. Kids use to walk in to my math classes and run out when they were over. Now they run into the computer lab and you have to ask them to leave when the class is over...a nice change. Switching from mathematics to technology education was one of the better decisions that I ever made.
So much is happening with technology that it is hard to keep up with all the changes. This makes it a challenging field to be in but it also makes it an enjoyable one. One of the better ways of staying current is by attending conventions. I was able to attend NECC ’97 in Seattle this summer. Listening to Bill Gates talk about the future of Microsoft and what he envisions for education and educational software was enlightening and the "Woz’" (Steve Wozniak) was also entertaining and gave us some insight into the future of Apple (I think it still has one!). The vendor displays were overwhelming and the number of new whiz-bang computer software products makes being a technology guru for your school district in charge of making software selections somewhat mind-boggling.
Attending conventions allows you to "experience" the software, ask questions, get advice and in the end make more informed software purchasing decisions. The upcoming MEA/MFT convention in Billings on October 16th and 17th will provide some of these opportunities, albeit on a somewhat smaller scale, but I have never failed to attend one of these gatherings and not come away with some new, inspiring, or enlightening information. Yvonne Andres, our keynote speaker, was one of the featured attractions at the NECC meetings in Seattle. Be sure to attend her sessions on Thursday, and find out about ThinkQuest and the Global SchoolNet Foundation. I have organized a Think Quest team at our school and we plan on participating in the ThinkQuest program during the next school year.
The MCCE Technology educator of the year has been chosen and the winner will be announced at the convention on Thursday. This marks our first ever TOY selection and will provide an opportunity to recognize one of the outstanding technology using teachers in Montana. The winner will be the Montana nominee for the national TOY award given by ISTE.
In closing, I would like to urge all of our membership to consider a more active role in our organization by running for one of the MCCE officer or board positions. Be sure to attend the membership meeting on Friday morning at Skyview High School in Billings.
Hope your are off to a great start this school year!!
Bob
While Netscape runs file right out the the box, a virtual box that is, there are a few tricks that the casual user might not know are available. Some of these are just interesting and others are downright useful in the classroom.
Some of these tips will work on both Macintosh and Windows versions of the browser. I have tested them on both platforms running Netscape Version 3.
#1 - (Windows only) Are you tired of finding Netscape resized, a different set bookmarks loaded, and it pointed to a different homepage on start up? These setting are save in a file in the Netscape directory named netscape.ini. To lock this settings down, get Netscape set up the way that you want, exit it, then save copies of the netscape.ini and bookmark.htm files under slightly different names. If someone changes the settings, you can copy these backup files over the originals to restore the original settings. In my classroom this restoration happens everytime the computer starts up.
#2 - Are you curious about where your browser has been, especially after your students have been using it? Click on the Open button and type in "about:global" (without the quotes) and Netscape will display its history file. It shows what sites have been visited including the date and time.
#3 - Are you curious about that cache directory of yours? It sure does get full as it saves elements of web pages that you have visited. You can explore its contents by clicking on the Open button and typing in "about:cache" in the dialog box. Your cache will be converted into a hypertext document. This is handy when you realize that you forgot to bookmark that "site of all sites" that you visited a few click ago.
#4 - What about my right mouse button? Use that right mouse button in Windows (Open Apple-Click on the Mac) to open a menu box that allows you to save an image, open a link in a new browser, copy a link location (the URL), add a bookmark, or a few other options depending on your operating system. Just put the cursor over a link, the background, or an image and give it a try.
#5 - Save a few keystrokes. Did you know that in Netscape 3 you do not have to type in the "http://www." when you enter an URL in the Location or Open boxes? If the site name has a ".com" in it you can leave that off too. For example, to get to Apple’s web page, just type in "apple" and leave the rest to Netscape.
#6 - How about some shortcuts? These shortcuts are not easily found in the documentation.
| Function | Mac | Windows |
| Go to Fish Cam site | Control-Option-F | Control-Alt-F |
| Go to the White House | Control-Option-W | --- |
| Hide the scroll bars | Control-Option-H | --- |
| Refresh the browser | Control-Option-S | --- |
| Toggle the status window | --- | Control-Alt-S |
| Check connection status | Control-Option-T | Control-Alt-T |
The Children's Literature Web Guide
This site is for teachers, students, and writers.
Mid-continent Regional Educational Laboratory
Lesson Plans and Activities
A huge index of lesson plans covering every grade level and content area.
US Charter Schools
Keep up on the latest information on Charter Schools.
You’re invited to relax in the lush, CyberSpace setting of ED’s Oasis (http://www.EDsOasis.org), a web site created to help teachers use the Internet in their classrooms by making it easier to find outstanding web sites, and to know how to use the sites in a typical classroom setting.
ED’s Oasis simplifies the task of locating resources by providing links to carefully-screened and evaluated educational web sites. These sites are selected both for the high quality of curriculum-supporting information they contain as well as for the opportunities they provide students to communicate with other students and with experts, and to publish their own work.
Linking to a site is only the begining. Even veteran teachers may feel unsure about how to adapt a favorite lesson to incorporate an Internet-based activity.
With this in mind, ED’s Oasis features examples of teachers who are experienced Internet users in the Spotlight. These teachers share lessons, sample student work, tips and strategies.
Visitors to ED’s Oasis are invited to contribute to the site by sharing their own experiences, by nominating sites for review and teachers to feature in the Spotlight, and to try their hand at creative writing in the on-going contest.
All of these activities are intended to follow the model set forth by the 21st Century Teacher Initiative - to strengthen the community of educators across America, and to increase the level of teacher support for using Internet resources to increase student learning.
Terrie Gray, Director, ED’s Oasis
http://www.EDsOasis.org
tgray@pepperdine.edu
ISTE Contact Information:
Web Site: www.iste.org
E-mail: iste@oregon.uoregon.edu
Editor’s Note: The site has a new section featuring web sites that are of special interest to girls.
|
