How to Use Windows Vista Speech Recognition

Voice recognition software has been around for a long time, but it’s only in the last few years that it has become accurate enough and simple enough to use with any regularity. It has also been rather expensive, with “basic” versions running around $80-100 and “premium” versions running to several hundred dollars – prompting many buyers to ask what was missing from the lower-priced versions.

If you have Windows Vista, though, you might be surprised to find that voice recognition is built in – and that it’s pretty good. While it takes some getting used to, with a little practice you’ll soon be able to use speech recognition to create and edit documents as well as to control most of the functions of your computer.

Before you can begin using speech recognition, you’ll need to spend about an hour setting it up. This involves detecting your headset or microphone, running through a tutorial, and training the software to recognize your speech patterns. To get started, complete these steps:

1. Open the Control Panel and double-click “Speech Recognition Options”. This opens the speech recognition panel, with commands for starting speech recognition, configuring your microphone, running the tutorial, training the software, and opening the speech reference card which will help you learn the commands.
2. Double-click “Set up microphone”. Follow the instructions to make sure your microphone is working with your system. Note: although in theory you can use any microphone, standard microphones that plug into your sound card tend not to have good enough sound quality for speech recognition. Instead, you should look for a microphone or headset that plugs into your USB port. I use a basic Logitech model that cost about $40.00.
3. Return to the Speech Recognition Options panel and double-click “Start Speech Recognition”. The first time you start speech recognition, it will run through the beginning setup and tutorial. The first tutorial lasts about 30 minutes and will teach you the commands you need to use with speech recognition while also training the software to recognize your voice. It helps to keep in mind when the tutorial becomes frustratingly repetitive, that it is also learning your vocal patterns.
4. From the “Speech Recognition Options”, select “Train your computer to better understand you.” You’ll then be asked to read a rather lengthy text in one of several styles. This allows the computer to add to its database of vocal samples, improving recognition and reducing errors.
5. When asked, have the program scan your “Documents” folder. Speech Recognition will add the words you commonly use to its database, and when it isn’t sure what word you meant it will recommend words to you based on how often you use words in your writing.

It’s probably best if you find a quiet, secluded area to run through the set-up. First of all, you want your voice and only your voice to register when you’re training the software. More importantly, people will give you all sorts of crazy looks when they see you talking gibberish to your computer.

Once you’re set up, you can begin to enjoy the wonders of modern speech recognition. When you ran the tutorial, you learned how to do most basic tasks, so I won’t revisit those, but here are a few general tips:

• Use speech recognition in a quiet place. If your microphone is any good at all, it will pick up all the stray noises in your vicinity and attempt to transcribe them. For some reason, my PC interprets every random sound as the word “if”.
• Turn speech recognition off when you’re thinking. You say “stop listening” to put speech recognition into “sleep” mode; it awakes when you say “start listening”. For some reason, having it waiting and ready to transcribe when I’m thinking makes me feel rushed and nervous and I end up not being able to concentrate; turning off speech recognition is a way of acknowledging to myself that I can think things through as long as I need to. Plus, leaving it on is just inviting a string of random gibberish as the program transcribes the noise you make moving around, working, and even breathing.
• Speak strongly and clearly. It helps to pretend you’re giving a speech. Use your best “Abraham Lincoln delivering the Gettysburg Address” voice.
• “Spell it”. When you introduce a new word to the software’s vocabulary, or when you use a word that sounds like a lot of other words, the program is liable to screw up. Just say “spell it’ and spell the word out, slowly and precisely.
• Retrain from time to time. As you get better at this (and it does take a while) you’ll change the way you talk — you’ll learn to speak more clearly, and you’ll become more confident thinking “on the fly”. Every once in a while, run through the tutorial and voice training, and have it scan your documents to pick up any new words. This has the added benefit of reminding you of things you’ve forgotten you could do.
• Be patient. Don’t get too discouraged when you get a lot of errors on your first (and second, and third) try. You will get better at this with practice.

There are good reasons to use speech recognition beside physical impairment that makes typing difficult or impossible. Using it well requires a level of vocal control and clarity that we don’t often practice, which helps to improve your speaking ability. It also helps learn to think on your feet — you’ll be surprised at how hard it is at first to compose meaningful sentences while speaking! It’s also a good way to move from a written draft to a typed draft; speaking your sentences aloud helps to catch awkward, unnatural phrasings that the eye tends to skip over. It’s also a good way to transcribe voice notes if you’re the kind of person that uses a digital recorder to take reminders over the course of the day.