Having a voice assistant to help you is great, but having the details of an appointment with your doctor recited in public? Arguably not so great.
Voice assistants and voice command technology are both becoming increasingly popular. Amazon alone has already sold more than 100 million Alexa devices, clearly the technology is here to stay in our homes. In public however, where strangers might overhear personal business or sensitive information, people have been much less eager to use voice commands. That’s a problem Microsoft aims to solve with a patent on “silent” voice input technology.
Microsoft’s “silent” technology makes use of the way whispering works. When you whisper you exhale breath from your lungs with your words. Microsoft’s proposed device would instead have you inhale while whispering (give it a try, it’s even harder than it sounds.) It’s a process known as ingressive airflow, and Microsoft proposes exploiting it to enable devices to process whispered voice commands in public.
There is one drawback. In order for a device to process your inhaled whisper, your face needs to be really close to the device, really close. About two millimeters close, in fact. You might feel more comfortable issuing “silent” voice commands in public, but is privacy worth the cost of practically kissing your smartphone? Here’s Microsoft researcher Masaaki Fukumoto demonstrating the technology in action.
Whether or not you’re comfortable planting a kiss on your smart device is your business but, regardless, the technology described in Microsoft’s patent is still only a patent, for now. Although there’s nothing to say that we might not someday be whispering commands to our phones, watches, rings, TVs, and other smart devices. Even if the future doesn’t involve mouth-to-device whispering, it’s an interesting exploration of a new concept which could spark novel research ideas.
READ MORE: Microsoft patent suggests you inhale whisper to your voice assistants [Engadget]
More on Voice Recognition: A New AI-Powered App Transcribes Your Conversations in Real-Time