Abstract: Systems and techniques are provided for processing audio data. For example, a process can include determining audio context information corresponding to a multi-band bone conduction microphone (BCM), wherein the audio context information is indicative of at least one of noise information or voice information. A process can include generating a control signal indicative of a resonance configuration for one or more resonators of a plurality of resonators included in the multi-band BCM, wherein the resonance configuration is based on the audio context information and corresponds to one or more frequency response adjustments. A process can include transmitting the control signal to the multi-band BCM, wherein the control signal is configured to cause the multi-band BCM to generate a BCM output signal using the resonance configuration.

Abstract: Disclosed are techniques for audio processing handover between audio devices. In an aspect, a method of audio processing handover between audio devices comprises selecting a first audio device to be a primary audio input device and deactivating the second audio device; performing a voice processing algorithm on the first audio device; detecting that the quality of the first audio signal does not satisfy a quality threshold, and in response, activating the second audio device and measuring the quality of the second audio signal. If the quality of the second audio signal is better than the quality of the first audio signal, select the second audio device to be the primary audio input device, deactivate the first audio device, and perform the voice processing algorithm on the second audio device. Otherwise, deactivate the second audio device for some amount of time before checking again.