New technology trains computer to know voices

    Speech recognition actually trains the computer to understand a person’s voice, a computer science staff member said when she spoke on NW Campus recently.

   Judy Firmin spoke on Speech Recognition for Windows XP about the usage and advantages of the new technology now available.

   One requirement for the new technology is a high-quality headset microphone, which will relay a person’s commands to the computer.

   Other system requirements consist of a 400-megahertz or faster computer, Microsoft Windows 98 or later and Microsoft Internet Explorer 5 or later.

   Firmin gave a hands-on demonstration by showing setup procedures. She wrote several paragraphs with the new technology in Microsoft Word without hitting a key on the keyboard.

   The setup consists of the repetition of several sentences into the computer so that the computer can train itself to understand a person’s speech patterns.

   Firmin also showed the disadvantages in that the computer can misunderstand what a person says without a high-quality headset.

   With the right setup and equipment, this new feature can cut worktime in half, Firmin said.

   “This feature can get up to 95 percent accuracy of your voice commands,” she said.

   Training procedures can be accessed through a Microsoft Help Web page.