The State of Speech

Jim Machi : Industry Insight
Jim Machi

The State of Speech

In many circles, people say Speech Recognition has kind of "faded."  For instance, SpeechTEK is next week in New York and there are 42 exhibitors according to my last count, far down from years past.  Is this any indication of the state of speech?  Certainly it is, but maybe for different reasons than you think.  What it tells me is that speech recognition has now entered the mainstream as a kind of "expected" function.  It is a mature technology now.  For instance, it is fairly commonplace now that your car will come equipped to "talk" to you.  Much of this is in the form of navigation devices, and much of this happened because of stand-alone navigation devices.  We're all familiar by now with the "recalculating" phrase from one of these.  Love it or hate it, we all know that phrase!  And by now we're all familiar with IVRs that you "speak" to as opposed to hitting digits and interacting with via DTMF tones. 


This perceived "non-interest" in a technology is very common as technology matures.  There used to be many, many internet telephony focused shows but now that the technology is embedded in the overall solution architecture and the industry "understands" this technology now, those shows have dwindled.   The fact that there still is a SpeechTEK shows the power of this technology!  We see SpeechTEK now is also more about solutions as opposed to the enabling technology that powers them.  


And as new technology enters the fray, we see new shows and specific webinars about them since people have a desire to learn about them.  We're seeing that now with HD Voice and M2M for instance.


So what is the future of Speech?  Certainly SpeechTEK would be a great place to go learn.  But outside of that, I can see we still have work to do to get a better accuracy rate.  Compared to eons ago, it's worlds ahead today.  But I've also read that the average accuracy rate is about 80%, which can still be frustrating to some.  I'm sure we'll also start to see talking avatars.  We're all familiar with some "help" avatars on websites so I'm expecting we'll start to see them talking to us soon.   And I'm expecting that we can voice enable a lot of the social networking applications out there. 


I can also see that we can go from using "words" to using more of a natural speech, and for the system to start even understanding intonations, etc., which is something you certainly get when talking to someone (and you don't get from emails for instance!).   And beyond this, I've also read of more futuristic work regarding interactive dialogues with machines, including understanding the "user" and tailoring that understanding to how the interaction might occur.   So clearly speech is alive and well and will always have an interesting and vibrant future.


Related Articles to 'The State of Speech'
hd voice pic.jpg
HD voice.png

Featured Events