Aardvark DailyNew Zealand's longest-running online daily news and commentary publication, now in its 23rd year. The opinion pieces presented here are not purported to be fact but reasonable effort is made to ensure accuracy.
Content copyright © 1995 - 2017 to Bruce Simpson (aka Aardvark), the logo was kindly created for Aardvark Daily by the folks at aardvark.co.uk
Please visit the sponsor!
One of the coolest bits of scifi novels and movies that I recall from my younger years was the ability for computer systems to interact using voice.
Now getting computers to talk has been fairly easy and as long ago as the early 1980s, a number of machines came with voice synthesizers. The Winbond phoneme-based synthesizer chip was standard equipment in a number of CP/M machines "back in the day" and rudimentary text-to-speech algorithms allowed you to write code which "spoke" to the user in a (usually) understandable voice.
Now you'd have thought that such a capability would have been a big selling point -- but it wasn't and although the Bondwell machine was pretty well priced and spec'd for its day, it was only ever an "also-ran" and never became a "must have".
These days of course, you speech synthesis has been hugely improved and there are now any number of websites that will take your typed words and speak them to you in a wide range of accents and intonations. Piece of cake!
But what about voice recognition?
This has always been much, much harder than voice synthesis.
Well that too has now become rather "every day" in its appearance and application.
One of the first practical applications of voice recognition that actually worked was Apple's Siri -- the mobile assistant.
The first time I saw and heard Siri at work, I was impressed... very impressed.
Up until then I'd only ever worked with voice recognition systems that needed to be "trained" to recognise a voice and the words spoken by that voice. This was a cumbersome and infuriatingly unreliable method of recognition and its poor performance explains why such technology never really took off.
Now however, we have systems that are speaker-agnostic, require no training and can even cope with differing accents, seemingly with high levels of accuracy and performance.
We have now become the scifi world predicted in all those cool novels and movies.
Of course just as Siri blazed a new trail, a raft of other devices have followed the same path and now, the next "big thing" seems to be the "home assistant" device.
These devices have really bought the future to our homes and it surprises me that we have this fantastic technology -- but no flying cars (yet).
The other place we've seen voice recognition technology trying to make our lives easier is inside cars. Quite a few manufacturers have put voice-recognition systems in their vehicles in an attempt to make it safer and easier to access key functions whilst still keeping their hands on the wheel and their eyes on the road. Sadly, it seems that these implementations are often somewhat lacking, just watch this clip of someone trying to pair a bluetooth device to their car's entertainment system.
I must admit that I've also had my own issues with those horrible voice-interactive telephone systems that the likes of Spark use. Fortunately it seems that repeatedly shouting obscenities usually gets you passed onto a "real person(tm)".
Just standing back a little and contemplating how far this voice tech has come, I can't help but wonder if, when fed into a suitably capable AI system, we're not rapidly approaching the point where the difference between man and machine becomes virtually indistinguishable. The Touring Test will become nothing but a memory perhaps?
Ah well, time to get on and do some work.
"Close the pod-bay doors Hal"
Please visit the sponsor!
Have your say in the Aardvark Forums.