As with any technology, there is a specific process through which the technology advances to reach a certain point at which it is considered to be highly functional. Voice recognition software is no different. Today we have Siri and Cortana, but how did we arrive at such an advanced stage of this technology.
When you take a look at the progression of voice recognition software, it is similar to watching a child grow up. There is the infancy stage, when very little is actually communicable; however, you know that something special is developing behind the limited capacity that you currently see.
Voice recognition has come from a phase in which it had a very limited vocabulary and its command system was almost non-existent to a phase in which it achieved the capacity to respond to questions quickly and accurately.
During the infancy stages of the development of voice recognition technology, the systems only had the capacity to recognize digits. This is actually a natural progression. When you consider the complexity of any particular language, it is easy to understand why the designers and engineers initially focused on digits — numbers are more universal and they are simplistic. The voice system can be traced back as far as 1952, when Bell Laboratories designed the Audrey system, which has the capacity to understand digits spoken by a single voice. In 1962, IBM created the Shoebox machine, which had the capacity to understand a whopping 16 words in English.
Over the course of the next decade a number of laboratories in countries, such as the U.S., England, Japan, and the Soviet Union (Now Russia), worked to advance the technology of voice recognition in order to usher in a more comprehensive voice recognition system.
The Exceptional Growth Spurt
As with many children, voice recognition experienced an exceptional growth spurt during the 1970s. It was substantial funding from the U.S. Department of Defense through the DAPRA Speech Understanding Research program that spawned this growth spurt. Through this program the Carnegie Mellon’s “Happy” speech understanding system was born. This system had the capacity to understand 1,011 words, which was substantially more than the 16-word capacity of the Shoebox machine.
The ’70s also marked several other major milestones in voice recognition technology, which included the founding of the first speech recognition company — Threshold Technology. Bell also introduced technology that had the capacity to understand and interpret the voices of multiple people.
Through the next couple of decades, speech recognition programs grew in capacity almost exponentially, with thousands of words being added to its steadily increasing vocabulary.
It was during this time that voice recognition began to become integrated into a number of commercial applications for specific businesses. The technology even found its way into a number of private homes through the creation of the Worlds of Wonder’s Julie doll in 1987. This was a doll that children could train to respond to their voice.
Modern Voice Recognition Software
Modern voice recognition software like that which is available with Cortana and Siri, or the programs that are used by multitudinous businesses with their customer services systems, had to overcome some serious hurdles that the previous generations were not able to overcome. One of the major hurdles was the requirement for discrete dictation.
Earlier versions of voice recognition programs required that any dictation be slow and deliberate, meaning that an individual would have to speak one word at a time, with a deliberate pause between words; however, the modern versions of these types of programs have the capacity to allow humans to speak at a normal pace, in a more conversational manner. It was in the mid-2000s that Google introduced its Voice Search app that took the voice recognition game to the next level, offering a program that had an exceptionally high level of accuracy. This was important for a number of reasons, one of which is the fact that voice recognition has the capacity to help people get past those small-sized keyboards on mobile phones that had become commonplace. Another key issue was that Google had the capacity to facilitate the processing for this service.
Through the technology developed by Google, voice recognition is available on a number of platforms, and it is highly specific. For instance, Siri and Cortana can be set to only respond to the phone owner’s voice. This technology has definitely come a long way.
Jessica Kane is a professional blogger who writes about technology and other gadgets and gizmos aplenty. She currently writes for Total Voice Tech, her go to for all professional Dragon products.