Smartphone voice recognition software has been a work in progress since 1932, when Bell Labs researchers were trying to understand speech perception.
Since then, the technology has been developed and enhanced for better reliability and increased accuracy. In the last four years, cloud computing architectures have made voice recognition technology on the smartphone progress rapidly.
Cloud computing allows the technology to access billions of records, and it uses smart algorithms to help accurately identify a variety of speech patterns. Read on to learn more about voice recognition software history.
The First Commercially Successful Technology
Not surprisingly, the history of speech recognition software begins not long ago. It was the 1990s when the first successful introduction of speech technology came onto the scene, as the vocabulary of the average speech recognition system surpassed the capacity of the average human vocabulary.
Dragon systems was the industry leader in the field, and still remains as the primary provider of speech recognition software for computers.
The challenge of applying this software to the smartphone industry was in the amount of data required by the program to accurately interpret human voice commands.
Introduction of Siri
In 2005, Apple licensed a voice recognition software by the name of Nuance for its now widely famous digital assistant — Siri.
The technology was more of an amusing feature at that point, and it didn’t provide the sort of professional results we expect of a full-featured digital assistant. The technology didn’t really start improving until the development of deep learning models that started to come out in 2009.
Other smartphone developers were quick to pick up the technology after the introduction of Siri.
Problems With Speech Recognition
The technology for speech recognition has been around for several years, but up until the introduction of speech recognition for the smartphone there have been issues with optimal word recognition. Words used to have to be spoken very slowly for them to be processed accurately.
As the capabilities increased, processors sped up and the amount of training necessary for a program to understand you correctly have been drastically reduced. Siri and other voice recognition programs create a database of your voice structure to better understand and interpret your commands.
Now, voice training in the traditional sense is no longer required, and the system automatically adjusts based on your responses. Essentially, speech recognition technology has learned how to learn on its own.
Deep Learning and the Cloud
Deep learning technology requires the use of massive databases of information to work correctly.
When you use your smartphone, you’ll notice that it require Internet access to provide accurate results. This is because the databases and service that these applications require to run operate on cloud-based servers.
The sheer amount of data required to process the simplest of commands wouldn’t fit on a handheld device. When you ask Siri a question, the delay in response is due to the fact that Siri must access a server to process your request.
Improvement of Processing Speeds
One of the developments that helped usher in a new era of smartphone technology was the increase of mobile-based processor speeds.
Faster processing speeds and quicker Internet and cellular data connections have made it possible to drastically improve the degree of service that smartphones can provide.
Some of the biggest barriers to voice recognition software are issues with homonyms, discontinuous speech patterns and ambient noise in the environment that can interfere with any spoken commands.
The Future of Speech Recognition
As technology improves, speech recognition software is going to become more deeply integrated into the products we use daily.
Most new computers already come with speech recognition, and you can even speak commands to your car to change the volume, adjust the temperature and change course in a GPS system.
Native speech recognition software will greatly replace the need to type responses and should continue to drive business in positive and effective ways.