Tell me what you want
A beginner's guide to Windows Speech Recognition
By Mark Carpenter
At first I was skeptical. Sitting back and talking to my computer seemed odd to me, a little too much out of a science-fiction movie. The first time I tried Windows Speech Recognition, I made sure to close the door to my office, lest everyone in the hallway think I'd lost my mind. But then a funny thing happened—it worked. I could start programs, move between open windows, and even write e‑mail, all without touching my mouse or keyboard. Once I got past thinking that I needed to talk like a computer to talk to my computer, I found that it was pretty easy and started to see what all the fuss was about.
In the past, speech recognition programs were used primarily by people who had difficulty manipulating a mouse and typing on a keyboard. With this version of Windows, speech recognition has become a more mainstream business program.
Speech Recognition is available only in English, French, Spanish, German, Japanese, Simplified Chinese, and Traditional Chinese.
What you need
Besides your voice, the most important piece of equipment that you need to get started with Windows Speech Recognition is a good microphone. While many new laptops have microphones built in, it’s better to get a separate microphone that's dedicated to speech recognition. A USB microphone can provide the best fidelity, but a mike that plugs into a mini-jack on your computer should work well enough.
Consider getting a headset microphone. Unlike a microphone that sits on your desk, a headset mike allows you to move around—and if you’re going to use Windows Speech Recognition a lot, that’s something you’ll want. Also, headset microphones are less prone to making mistakes because they stay a consistent distance from your mouth as you talk.
You’ll also want a consistent environment, without a lot of external noise that your microphone might pick up. As you use Windows Speech Recognition, you’ll train it to filter out sounds from your environment; but it’s good to start out with as quiet an environment as possible.
Now that you have your microphone, it’s time to dive in. Fortunately, setting up Speech Recognition is pretty easy. Here’s how to do it:
Open Speech Recognition Options by clicking the Start button , clicking Control Panel, clicking Ease of Access, and then clicking Speech Recognition Options.
Click Start Speech Recognition.
The first thing that Speech Recognition will ask you to do is set up your microphone. Before you start, make sure that your microphone is plugged in and positioned where you plan to use it.
Once your microphone is set up, the Set up Speech Recognition wizard will guide you through a series of questions. When you’re finished, you can choose to take a tutorial.
It's worth your time to run through the Speech Tutorial for a couple of reasons. First, the tutorial will help you understand how best to interact with your computer using Windows Speech Recognition. Also, as you run through the tutorial, the computer will begin building your speech profile, which makes Speech Recognition more accurate.
Tell the computer what you want to do
Once you’ve got Speech Recognition up and running, the first task is getting the computer to do what you want. This is surprisingly easy. In general, when you see an item on your screen that you want to interact with, say what you want the item to do, followed by the item’s name. For example, to open your Pictures folder you would say "Start," and then "Pictures."
It’s similar if you know the name of a program that you want to open or switch to. For example, you can say "Open Solitaire" or "Switch to Internet Explorer."
If you don't know the name of an item (such as a file or folder) that you want to work with, you can say "show numbers" and numbers will appear next to every item on the screen. Simply say what you want to do with an item, followed by its number. For example, you might say "double-click five" to open the item with that number.
To use Windows Speech Recognition for browsing the Internet, you need to figure out what the items in the web browser are called and say the commands you need to interact with them.
To enter the address of a website, say "Go to address," and then dictate the address of the website that you want to visit. If you've entered the address before, say "address," and then choose a previously entered address.
Tell the computer what you want to say
Computer, take a letter. Well, it’s not quite that easy, but for many people one of the pleasures of speech recognition is being able to dictate. From a simple e‑mail message to a complex document, using dictation to get your words into your computer can be faster than typing.
Most programs accept text dictation automatically. All you need to do is start the program and then begin dictating. It’s important to speak clearly, and it works best to speak in complete sentences. Keep in mind that punctuation isn’t automatically added when you dictate text, so you have to tell the computer where to add punctuation. The Speech Tutorial has a great primer on the ins and outs of dictating text. If the program you’re using doesn’t automatically accept dictation, you can enable it by selecting Enable dictation anywhere on the Speech Recognition Options page. For more information, see Turn on dictation for all programs.
No matter how good your computer is at recognizing your speech, there will still be occasional errors. Fortunately, with Windows Speech Recognition, you can fix dictation errors on the fly. When you discover a dictation error, you can say "correct," followed by the word or phrase that you want to correct. A dialog box then pops up and provides choices for the most likely corrections. If you don’t see what you want, you can just say the correction that you want to make.
If you find words that Speech Recognition consistently mishears or doesn’t recognize, you can add them to the Speech Dictionary. To open the Speech Dictionary, say "Open Speech Dictionary."
Wrapping it all up
It takes a little practice and a little patience, but once you get started it’s fun and easy to use Windows Speech Recognition.
About the author
Mark Carpenter is a writer on the Windows team at Microsoft. He spent his first six years at Microsoft working on documentation for the Macintosh product group. He's enjoyed the past two years seeing how the other half lives.
Have a comment for this columnist? Enter your feedback using the tool below. (You'll see the comment box after you click one of the buttons.) Note that although the columnist will read your feedback, personal replies are not possible due to the volume of feedback received.