Today, we almost take voice recognition technology for granted. For example, people with an iPhone may rely on Siri to get questions answered and access a search engines. Some cars even use voice speech recognition to access some features.
Some people may have forgotten how frustrated they were by early voice recognition software used on phone answering systems. The development of these systems took a huge investment of money, brains, and talent before they even got that far. Now, we may use voice recognition to send text messages, access Internet search engines, and more.
Voice Recognition in the 1950s and 1960s
The first developments in voice recognition actually occurred way back in the middle of the 20th century before computers were things that average people had in their homes. Because it is so difficult to account for variations in accents and speech patterns, the first systems could only recognize numbers. At the 1962 World’s Fair, IBM showcased something called a “shoebox” that could understand 16 English words. This might all sound very primitive, but remember that these engineers and scientists were also working with much more limited computer technology than almost anybody can afford to buy today.
Speech Recognition Reaches the Toddler Stage in the 1970s
The 1970’s may have been the decade when speech recognition got better than what scientists tell us is the vocabulary of a smart dog. This is mostly due to efforts of the US Department of Defense and the money they plowed into projects. The DARPA program ran from 1971 to 1976, and it was one of the largest programs in the history of this technology.The program sparked Harpy at Carnegie Mellon, and this system could understand over 1,000 words. This is supposed to be the vocabulary of an average toddler, and that is what scientists say that intelligent dogs may be able to understand.
The 1980’s Begins Predictive Text
Do you have a spouse or best friend who can finishes sentences for you? The 1980’s is the decade when predictive text came into being. That may be your friend or foe on your own phone’s text messaging system.
The vocabulary of speech recognition systems also jumped to several thousand words. Still, there were not really household applications even though there were business and technology applications. There was a Julie doll in the 1980’s that used early speech recognition though.
The 1990’s: Speech Recognition Goes Mainstream
The 19990’s might have seen the birth of those frustrating phone answering systems that never really seemed to understand anything. Dragon launched its first program, called Dragon Dictates, in 1990. This first application cost $9,000 and did not work that well. It took seven more years for a much improved version called Dragon Naturally Speaking arrived on the market, but it was still not at the level of the software by the same name you can buy today much cheaper.
The 2000’s: Google Breaks the Stalemate
During most of the first decade of the 21st century, progress actually seemed stalled. Speech recognition software seemed stuck at about 80 percent accuracy. This is sort of ironic, but the Google Voice Search App for the iPhone actually started competition and progress again. But before you give to much credit to the IOS Siri, remember that Google checked in first.