ADVERTISEMENT

ADVERTISEMENT

The next voice you hear ...

The relationship between computers and writers is about to get a lot more personal -- instead of pecking away at a keyboard, writers and just about anyone else who uses a computer will be able to talk to their computer, and it will know how to re...

The relationship between computers and writers is about to get a lot more personal -- instead of pecking away at a keyboard, writers and just about anyone else who uses a computer will be able to talk to their computer, and it will know how to respond.
Until now, voice recognition software has been a tool for the physically disabled. But new, more sophisticated software is bringing the technology into the mainstream. NaturallySpeaking is the name of a software program now available locally from Scott Coleman, area sales rep for Freedom of Speech, a Minnesota-based software company.
"There are people out there right now using NaturallySpeaking and achieving in excess of 200 words per minute with 98 percent accuracy," Coleman said. "That's not uncommon. What it boils down to is a more powerful computer with a good sound card, good headset, clearly enunciating your words and speaking in continuous sentences. I think the biggest problem that people using NaturallySpeaking will have is that they're not correcting their voice-recognition errors. Or they're not properly positioning the microphone or they're not clearly enunciating their words. If you can speak like a reporter on TV and you have a really decent microphone and a powerful computer processor, you can achieve great results."
Using full sentences helps the computer recognize some words, because it uses sentence structure to help it decide, for example, whether it heard "ice cream" or "I scream." With full sentences it will also know which homonym to use, differentiating between "their," "they're" and "there," for instance.
"The program also has the capability of looking at previous documents that you have created," Coleman said. "Say you've written several previous stories for a newspaper. The computer will familiarize itself with your sentence structure, your writing style, your grammar, and that will improve its recognition."
Once a voice file is developed by reading through several short pages of text provided by the computer program, a person can sit down in front of the computer and talk away. The words appear on the screen. Sentences are automatically capitalized at the beginning. The program automatically inserts commas in large numbers and a hyphen in the middle of phone numbers. The program will recognise an e-mail address and use all lowercase letters and the @ sign rather than 'at.'
The professional version of the program also allows the user to set up macros to shorten tasks that are repeated frequently.
"It's very useful for companies and people who do a lot of repetitive dictation," Coleman said. "If you're doing a letter and your standard closing paragraph is 'Thank you very much for your time and consideration. Sincerely...' and your name and title and address and all that, when you get to the end of the letter you can say 'my closing statement' and, boom, it puts it all there."
Besides simple dictation, voice recognition software can take the place of all kinds of different keystroke commands. Anything you can do through any of the pull-down menus you can accomplish by voice. You can operate mouse movement by voice. You can tell the microphone to sleep and wake up or turn off completely, although, once the microphone is off it takes a keystroke to turn it on again.
Lots of different professions are finding the voice recognition software useful. Police officers are using it for dictation. Lawyers are using it for briefs. Even some airlines are considering a move to voice activated software.
The new, more powerful computers and the more sophisticated software is speeding the software on the road to mainstream use. When the technology was first developed in the 1960s, it was primarily used by people with disabilities. But computer processors at that time weren't fast enough or powerful enough to handle voice recognition software accurately or efficiently. Companies like IBM and others interested in voice recognition software decided it wasn't worth the investment.
The company that developed NaturallySpeaking kept with it, and now the more powerful computers today make continuous voice recognition possible. Freedom of Speech distributes all kinds of assisted technology for hearing impaired, vision impaired and mobility impaired.
"There are many, many programs out there designed for assisting people with disabilities," Coleman said. "Up until now that was their primary function. Then NaturallySpeaking came out and it's for everybody, not just for people with disabilities. It's really gone mainstream."
The professional version of NaturallySpeaking will run $695, which includes a high-quality microphone mounted on a headset. Initial training once the program is installed costs extra. After that, users get lifetime technical support.

What To Read Next
The system crashed earlier this month, grounding flights across the U.S.