Speech. How simple it seems to most of us. Spoken communication is such a fundamental part of human life that we rarely stop and think about how amazing an activity it really is. The vocal system is a very complex part of our anatomy and producing spoken language relies on highly sophisticated, and still not fully understood, physical and mental processes. Understanding spoken language is possibly even more elaborate. Yet, as we go about our daily lives, most of us remain blissfully unaware of these complexities. It comes, indeed, naturally to almost all of us.
Because spoken communication comes so easy to us and is such an elemental part of life, it is perhaps understandable that from very early on in the history of computing, we have sought to make our computers recognise speech. Indeed, speech recognition technology has been around for decades and today speech recognition engines are available for PC users as of- the-shelf software.
Yet, even the most recent speech recognition technology remains constrained by quite sharply defined limits and is by no means able to truly understand speech in the way humans do.
In this article, I will provide an overview of the possibilities and limitations of speech recognition technology and its application potential.
Next: How it works