That’s partly because we’ve spent tons of time making sure our IVR understands everything you say, and partly because speech-recognition technology requires a huge amount of virtual space and power to run that other tech fields just don’t have access to.
Somehow, though, we eventually figured out how to downsize these systems enough to give your computer a voice. Now personal computer speech-rec apps are nearly as good as what we run on our IVR systems.
So, encouraged by this success, we decided to put speech-rec into our smallest consumer computers yet – our smartphones. Much like the last transition, though, that came with some serious sacrifices in quality.
I can’t tell you the number of times I’ve asked my phone’s early generation voice control to call David Jones and I’ve ended up emailing Doctor Bones. These comprehension troubles have led many people to be wary of the whole world of speech-rec, when really the trouble lies in the phone itself, not the speech-rec (as our industry proves everyday).
You see, the issue with running speech-rec on your phone comes down to processing power. When you make a smaller computer, there’s less power to run these super-powerful programs and your whole system starts to crash.
Thankfully our smartphone’s programming has gone from bad to better, and will hopefully even soon begin to catch up with our best IVR systems to date, and it looks like that upgrade is going to come from a difference in approach.
According to Ars Technica, the team developed a prototype silicone chip, called In Silico Vox, which already recognizes and decodes 1,000 English words, and will hopefully know many more within a year.
If this chip does prove to be viable, it could be used in everything from remote controls to cars, making speech-rec a whole lot simpler for all of us.