|Speech Recognition in Office XP
Speech recognition is not new. The ability to dictate to your computer through a microphone has been around for several years. Programs such as Dragon Dictate were first on the market. They used discrete speech — say one word, pause, say the next word, pause, and so on. In the text mode, your words appeared in your document; in the command mode, you were able to open menus and make selections. Later programs used continuous speech. Now the computer is able to recognize entire phrases or sentences without having to pause between each word.
So what is new? Microsoft Office XP includes a basic speech recognition program for free! Once you set it up in Microsoft Word, you can “talk” in any Office XP program. Some people use speech recognition because their hands are occupied and they prefer to dictate their documents. Others use it because they have difficulty manipulating the keyboard and/or mouse. Still other users find it easier to speak rather than type, especially those with written, language-based disabilities. Whatever the reason, speech recognition is sure to become more and more popular in the near future.
Before you begin using speech recognition, however, you must make sure your computer system meets the hardware requirements, listed as follows:
The first step in preparing to use speech recognition is to adjust your microphone so your voice input is neither too loud nor too quiet. When you click Tools, Speech in Word XP, the Welcome to Office Speech Recognition screen opens. Make sure you are in a relatively quiet room. (If there is too much external noise, your microphone settings and voice training may be ineffective.) Click Next.
The Microphone Wizard starts. Text and an animated picture tell you how to set up your microphone and speakers or headset. Click Next again. The Test Microphone screen appears, as in Figure 1.
Figure 1: The Test Microphone screen
As you read the sentence written in the dialog box, the volume meter will move. Raise or lower the volume control on your microphone until the meter stays in the green area, then click Next. The Test Positioning screen is for headset users, only. When you read a sentence out loud into the microphone, it is spoken back to you. If it sounds like you are blowing into the microphone, reposition the microphone until it sounds normal. Then click Finish.Training Microsoft Speech Recognition
Next you will “teach” Office to understand your voice. The Voice Training window opens automatically when you finish adjusting your headset. Click Next to fill out your speech profile. Speech recognition will perform better if the speech engine (the program that interprets your voice) knows if you are male or female and have a child’s or adult’s voice. Click the option buttons for gender and age. Then click Next.
Some instructions follow, with a Sample button that lets you hear how you should read the sentences in the next screens. Read the instructions, then click Next. Make sure the room will be quiet for the next ten minutes and the microphone is properly positioned. Then click Next.
The first training screen appears, as in Figure 2.
Figure 2: The Voice Training screen
Once you have read all of the training sentences, Speech Recognition will create a file of your speech patterns, pronunciation, and so on. As you read the text, you are given helpful hints as to how this feature should be used. The final screen informs you that you shouldn’t expect total accuracy at first. It takes additional training and use to improve accuracy. Click Finish to go on.
Upon completing training, a video opens teaching you how to use speech recognition commands and techniques. Close the screen containing the video when you want to start dictating.Dictating in Microsoft Word
The Language Bar appears on top of the screen when you are using speech recognition, as shown in Figure 3.
Figure 3: The Language Bar without text labels
The Language Bar can be displayed with or without text labels. In Figure 3 the icons are Correction, Microphone, Dictation, and Voice Command. The message “Too soft” is an example of errors you may see if you speak too quietly, too loudly, or are otherwise completely misunderstood.
Assuming you have a blank document on your screen, say “dictation.” Then say the following sentence:
“I am now using speech recognition in Microsoft Word period”
As you speak, dots in a blue background move across the screen. When the speech recognition engine figures out what you have said, the dots are replaced with text. Depending on how well you have done your training, the text should be pretty accurate. It is likely that word will not be capitalized, but the word period should be replaced with an actual period to complete the sentence. Accuracy should be pretty good because many of the words were in the training. Try speaking another sentence — you should see a drop in accuracy if words are spoken that were not used in the training sentences.
You can use the keyboard and mouse to clean up mistakes or to type, such as when you have a person’s name or address using spelling that speech recognition is unlikely to understand. Alternatively, you can say “spelling mode” and then spell each letter of the word you want.
When you want to stop dictating, say “microphone” and continue working on your document. As soon as you want to resume dictating, click the Microphone icon on the Language Bar.Making Changes and Corrections
Remember, you can use your keyboard and mouse whenever you want — even in the middle of dictating. You can make some corrections, however, orally. For example, if speech recognition completely misunderstands what you said, say “scratch that.” The most recently entered word or phrase is deleted. You can keep scratching text to back up until you reach the last correctly interpreted words.
In the sentence you previously dictated, you may have noticed that you said “period” at the end of the sentence and that speech recognition inserted the period character. You could have said “dot” to get the same result. Other punctuation characters and symbols you can enter include: comma, colon, semi-colon, question mark, exclamation point, hyphen, dollar sign, open or close parenthesis, open or close quote, and numbers (20 or less are spelled out, 21 and up are entered as digits). If you want all numbers displayed as digits, say “force num,” pause, and speak the digits.
To end one paragraph and start the next (or just to insert a blank line) say “paragraph.” To select text, say “voice command” and choose from the following options:
To move the insertion point, you can say “up,” “down,” “right,” “left,” “next line word/paragraph,” or “last line/word/paragraph.”
You can also choose from the following text correction options while in the Voice Command mode:
If you want to bold or italicize a word, for example, use the Voice Command mode to move to the word, say “select word,” and then say “bold” or “italic.” To change the font, say “font” and then say the name of the font.Other Voice Commands
The Voice Command mode is also useful for making menu choices. Just speak the name of the menu, “file” for example, and the name of the menu option you want, such as “save.” Say “more” or “expand” to see all the menu options, if they are not immediately visible. If the wrong menu option opens or if you change your mind, say “escape” to close the menu.
You can choose from the toolbar options by saying their name, as well. (If you don’t know the name of the button, hover your pointer over the icon and read the Screen Tip that appears.)
Use the online Help feature to learn about more voice commands you can use to do the following:
You can also use online Help to learn about other methods for correcting speech recognition mistakes. As an example, you can move the insertion point to an incorrectly spelled word or a grammatical error (underlined in wavy red or wavy green, respectively). Then say “right-click” and “down” and “enter” when you reach the proper choice.
The more you use speech recognition, the more you will learn about it. And, remember, you can use speech recognition in other Office XP programs, such as Excel, Office, and PowerPoint.