The proposed “ingressive speech” method enables placement of a microphone very close to the front of the mouth without suffering from pop-noise, capturing very soft speech sounds with a good S/N ratio. It realizes ultra-small (less than 39dB(A)) voice leakage, allowing us to use voice input without annoying surrounding people in public and mobile situations as well as offices and homes.
By measuring airflow direction, SilentVoice can easily be separated from normal utterances with 98.8% accuracy; no activation words are needed. It can be used for voice-activated systems with a specially trained voice recognizer; evaluation results yield 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 whisper-like natural voice can also be used for real-time voice communication.