Screaming for a subvocal microphone

When will we benefit from voice recognition without even using our voice?

My Samsung phone has two voice assistants — S-Voice and Google — and in the privacy of my own home, I use them often for a finite set of tasks, saying things like:

“Wake me up in twenty minutes.”

“What time is it in Moscow?”

“American bulldog.”

“Text Jim, I’ll be there in five minutes.”

But this time-saving feature gets zero action when I’m riding the bus, in line at the grocery store, or sitting at a restaurant. It’s noisy, personal, and awkward. So I asked Google — privately —

“Can I buy a subvocal microphone?”

And in more words than one, Google told me “No”.


Things that I can buy include a throat microphone, which is basically a mic that amplifies the vibrations in your throat. The upshot is that you can be heard better in noisy conditions, but you can’t exactly just whisper and expect the mic to hear you well enough for voice recognition to work its magic. Also, throat microphones are conspicuous, large, and ugly. They make an over-ear Bluetooth headset look glamorous by comparison.

Products out of my reach include a slick adaptive technology called The Audeo, which actually picks up neurological signals as they are being transmitted to your brain. Besides gaming and chatting with Google, it has real promise for people with disabling conditions like Parkinson’s. The price for a development kit is too much for me: $2,000. Also, because it’s not market-ready, it’s almost definitely not ready to hook up to your iPad at Starbucks.


If it’s not ready now, then when? While I don’t think these developments are light years away from the market, the major research NASA was doing in this field had its funding terminated along with a larger program, compellingly titled Extension of the Human Senses. In a 2013 interview, NASA scientist Chuck Jorgenson explains their past progress, saying they developed the capability to understand hundreds of words subvocally.

Unfortunately, he also explains that words spoken subvocally require special computer interpretation and have unique challenges. For example, “applosive” noises that are made with the mouth, not the throat, are difficult to translate, such as “P” sound, which you actually make with your lips. For that reason, we will sooner be able to subvocalize sentences like “find Italian food” than “find Bernardinos on St. Peter Street.”


If we could use subvocal speech dictation today, would you be ready? How well do you think vocally?

For my part, although I regularly dictate short thoughts to a computer, I usually fail to speak them coherently once they grow longer and more complex. And then editing text vocally requires far more skill and effort than I currently possess.

So until I’m able to learn how to control a computer with my voice, controlling one subvocally may be a moot point.


Alternately, the stigma of speaking to your computer in public could simply disappear. It’s a terrifying prospect, but before mobile phones became prevalent, most people wouldn’t have foreseen all the boring, intrusive, and embarrassing conversations we routinely overhear today.

In any case, I plan to wait — quietly — for my subvocal mic.

 

  • david lee

    I’m just wondering what’s Audeo going to do with thir technology? It seems they are not willing to put their product in the market yet. I really need this kind of mic for my speech analysis system

  • Roberto Voce

    This is the perfect example of one of those things that people can live perfectly well without but that will be an “absolute need” once invented and working. Like television or mobile phone.
    Actually I hope this will happen more sooner than later. 😉

  • clarkbattle