Watch the scary presentation of Microsoft Silent Voice at the ACM CHI Conference

Microsoft is working on a new voice input interface that will allow users to speak and record without voice leaks. The research is conducted by Microsoft Research and has been featured in ACM CHI 2018.

Called SilentVoice, it uses the other 50% of your voice by recording sounds made by inhaling rather than exhaling. This allows for a whisper-like sound that allows a microphone to record the voice without disturbing the surrounding people. In addition, the module will also filter the surrounding voice so that the user can capture a clear voice even without external interference.

SilentVoice is a new voice input interface device that penetrates the natural user interface based on speech in everyday life. The proposed method of "ingressive speech" makes it possible to place a microphone very close to the front of the mouth without causing noise noise, capturing very soft speech sounds with a good S / N ratio. ultra-small voice (less than 39dB (A)), allowing us to use voice input without disturbing the surrounding people in public and mobile situations, as well as in offices and homes. By measuring the direction of airflow, SilentVoice can easily be separated from normal statements with an accuracy of 98.8%. no activation word is needed. It can be used for voice-activated systems with a specially trained voice recognition system. the results of the evaluation give word error rates (WERs) of 1.8% (speaker-dependent condition) and 7.0% (speaker-independent condition) with a limited dictionary of 85 command sentences. A natural whisper-like voice can also be used for real-time voice communication.

You can view the complete presentation at the ACM CHI conference on human factors in computer systems.
Session: Sensing and Acoustics below:

The technology is still in the research phase but will surely help those who like to use voice commands but prefer to work without disturbing those around them.