How Does Voice Recognition Technology Work?
Voice recognition technology typically comes in the form of a program designed to convert spoken word into text. Voice recognition technology can be found in such applications as automated phone messages, appliance control and data entry. Voice-to-data conversion involves a number of complex processes.
Other People Are Reading
-
Analog-To-Digital Converter
-
Voice recognition technology uses an ADC (analog-to-digital converter) device, which converts analog, or continuous, voltage into a digital number. Spoken words create vibrations. The ADC translates these vibrations, or sound waves, into digital data that a computer can interpret.
Phonemes
-
Voice recognition technology divides the digital data into smaller segments. The software program matches these segments to phonemes contained in the specified language. A phoneme represents the smallest unit of sound used to produce contrasts between units of speech, or utterances.
-
Text Output
-
Voice recognition software attempts to place each phoneme in a context based on the other phonemes the software it immediately picks up or registers from the speaker. The software program accomplishes this by running the phonemes through a statistical model based on mathematical computations, called algorithms, and by comparing the phonemes with a library of words, phrases and sentences it has stored. The voice recognition program then outputs the speaker's words as text or executes the appropriate command.
-
References
Resources
- Photo Credit telephone image by Goran Bogicevic from Fotolia.com
