Gyrophone: Recognizing Speech From Gyroscope Signals

We show that the MEMS gyroscopes found on modern smart phones are sufficiently sensitive to measure acoustic signals in the vicinity of the phone. The resulting signals contain only very low-frequency information (< 200 Hz). Nevertheless we show, using signal processing and machine learning, that this information is sufficient to identify speaker information and even parse speech. Since iOS and Android require no special permissions to access the gyro, our results show that apps and active web content that cannot access the microphone can nevertheless eavesdrop on speech in the vicinity of the phone.


Ice Cream Memory CardSun An
1. Fritto Misto 07:452. Feeling 08:47&#160;3. Feeling II 08:54&#160;4. Afternoon 16:32&#160;5. Late Afternoon 04:20&#160;6. Rest Easy 07:41

Download free/pay whatever:

Ice Cream Memory Card
Sun An

1. Fritto Misto 07:45
2. Feeling 08:47 
3. Feeling II 08:54 
4. Afternoon 16:32 
5. Late Afternoon 04:20 
6. Rest Easy 07:41

Download free/pay whatever: