(Under Rachel Illingworth)
Technical experts have incorporated speech recognition software seamlessly into smartphones. So much so, that many people don’t realize they’re using it. But in all honesty, you probably use speech recognition software on a daily basis. It gets utilized when you safely use text-to-talk features on your drive home, ask Siri a question, or even using Amazon Echo to order a pizza. But what is it, how does it work, and how will it affect us?
What is speech recognition?
A simple definition of speech recognition is the understanding and translation of spoken words into digital text by computer software.
How does speech recognition work?
First, your smartphone has to determine what sounds you’re making and what sounds are coming from your environment. Once it’s determined your voice, it parses your speech by examining the sounds you make and the context of the words you use. You can see this happening in the small delay before it appears as text on your device’s screen.
Google’s engineering director Scott Huffman explained it best. He said speech recognition has improved greatly by the creation of many computers in a network working together. He went on to say that they call it “deep neural networks, or deep learning. [His team is] now able to apply very large-scale parallel computing to interpret the sounds you make.”
One example of a stellar recognition software is Google’s Cloud Speech API. It boasts the recognition and translation of over 80 languages and their variants in real-time. The API pulls text from two sources, through streaming results through the neural network or through already recognized text stored in a file.
How does this affect us?
Although Siri came out just four years ago, and you probably had a few jokes or two on her, she has come a long way in a short time. But talking to your computer just got a whole lot easier.
As of October 2016, speech recognition matches human speech transcription with an error rate of 5.9%. This means that computers can understand human speech easily and fluently transcribe it without making many errors.
In fact, a recent article in Slate magazine was written entirely through dictation to an iPhone. The article has a few misunderstood homonyms, but goes on to state that “Google, in particular, has gotten quite good at discerning your meaning from text.”
What’s in speech recognition’s future?
Where will scientists take this achievement? Maybe it means that computers will become understanding of human speech, movement, and action. Maybe it means that one day, computers will be able to read our minds. In the end, it means computers understanding humans instead of the other way around.
Because Google is a frontrunner in the race for perfect speech-to-text software, we highly recommend the new Pixel smartphone. Stop by your neighborhood TCC store to get your hands on the latest devices that feature strong speech recognition software.