Monday, January 6, 2014

Getting Started With Voice Recognition Software

Our introductory article on using speech to type, Give Your Fingers a Break! An Introduction to Voice-Recognition Software, included tips to help you decide if voice software is the right tool for you. This article assumes that the answer is "yes!" – and that you'd like to learn how to get started.

Here, we will cover a few basic tips to help make sure you have a positive experience in getting your computer to understand your spoken words and successfully translate them into written words.

Woman using microphone on beach

General guidance

First, I highly recommend that you look online for a popular, active voice-recognition software discussion forum, such as this one, and "listen in" to the advice these groups give regarding their experience with various soundcards, microphones, voice-recognition software programs, and techniques. If you are having difficulty, these discussion groups are also a great place to ask questions and get help, often much better than you will get from the software's manufacturer.

Second, to make sure you don't get frustrated, set realistic expectations. If you envision yourself stretching back in your computer chair with your eyes closed as you dictate, you may be disappointed.  Yes, voice recognition software is amazing technology but it is far from that idealistic (and still futuristic) vision of enabling you to talk to your computer and having it fully and accurately comprehend as well as a person can.  So, while it may generally do a good job, it's best if you watch the screen to see what it's typing after you speak each phrase.

It’s worth noting that I am dictating this article with my voice software, and more noteworthy that it’s translating my words so far with 100 percent accuracy.   Yet, while it's possible to generate an entire article such as this one without a voice recognition software error, the software makes enough mistakes in its day-to-day interpreting of my words that I've got a steady stream of new blog entries for my voice recognition errors blog to share some of the entertainingly incorrect translations it has made.  For example,
What I actually said to my computer: mountain bikers and trails  
How my voice software translated it:  mountain biker’s entrails

So you see, it's best to watch your computer screen as the software interprets your spoken words so that you can catch any mistakes and correct them before hitting "send" on an important message to your boss or a prospective customer.

The fundamentals – getting set up

In order to be able to talk to your computer and have it understand, translate, and type out your words, you'll need a few basics. We cover these in detail in the previous article, but here's the highlights.
  • A fast computer.  Just about any modern-day computer will do, but the faster it is, the better the results.
  • A good quality microphone.  No matter how good the software is, it is no better than the quality of the microphone.
  • A sound card. Your computer probably already has one, but you can get better, more consistent results with soundcards recommended specifically for voice software.
  • Voice recognition software.  This is a program that is capable of listening to incoming sounds, compare them to its built-in dictionary of language, and then type out your spoken words as text into a document on your computer.  (See brand recommendations in previous article.)
  • A quiet environment. I don't recommend trying this in a place with lots of loud talking, like a coffee house. The microphone does not know which voice to listen to, and will pick up lots of extraneous stuff and try to type it into your document.
Once you've installed your voice software, it may require you to go through a setup process and a brief training. The training helps it to understand your speech patterns, accent, etc.

Laptop with microphone

Getting started tips

To help you have a good experience, here are two basic tips to improve the voice recognition process.
  • Speak clearly and consistently. The voice recognition software is listening and learning constantly. If you speak loud one time and very soft the next, it won't translate as well. Also, the more clearly you enunciate, the better your results will be.  Since, in general speech, we tend to slur and blur words together, the software is likely to make mistakes in understanding your words, unless you can learn to speak very clearly. The more you speak in the manner of a TV newscaster, the better. To demonstrate this need to speak clearly, I've made a short demonstration video, which you can watch here.
  • Position the microphone properly and consistently. The dynamics of sound reproduction change dramatically depending on whether your voice is an inch from the microphone vs. four inches. For this reason, you get best results with a headset rather than a handheld microphone or one that is built into a notebook computer. With a headset, no matter which way you turn your head, the microphone will be consistently positioned relative to your mouth's position.

Open quote/close quote

One of the trickiest parts in learning voice software is adjusting to speak punctuation aloud, which you must do if you want punctuation in your document. The software does not intuitively know where, or whether, to put a comma or question mark, for instance. It only knows if you tell it to put one in. For example, let's say you wanted to type the following:

I asked her, "Which movie do you want to see?"
To have the computer type it exactly as you intend, with punctuation, you need to literally speak aloud the punctuation. So, what you would say to the computer in order to type that line is:

“I asked her comma open quote cap which movie do you want to see question mark close quote.”

Confusing?  It's probably easier to follow if you watch a demonstration of this, so I've created another video where you can watch how voice recognition software types when I'm speaking to the computer. The demo also shows how to get it to type punctuation, perform computer commands, and more.

Bon voyage into the seas of hands-free typing!

With these basics, you are ready to set sail. Speaking your words aloud to your computer will take some adjustment but, as you get more comfortable with it, you will be amazed at how fast and accurately you can type without ever touching your keyboard, just by using your voice.

As you begin working with voice recognition software, feel free to post your voice software questions or share your voice software experiences with others using the comment field below.

Ric Moxley
Contributing Writer

No comments:

Post a Comment