Incorporating microphone arrays have been and will continue to be a common trend for voice communication and voice control applications.  Having multiple air-conduction (AC) microphones allows for implementers to include advanced signal processing techniques such as, beamforming and noise cancellation for speech enhancement. However, with the growing market of wearables and hearables, incorporating a bone conduction (BC) microphone into product designs allows for a greater opportunity for speech enhancement.

The major advantage of bone conduction microphones is they do not capture ambient environmental noise.  If you are in a noisy environment with a television in the background or a restaurant/bar with multiple people talking in the background, or even outside in the wind, a BC will not sense this audio.  Since BC mics capture the vibration of the human skull, only the speech of the person wearing the BC mic will be present in the resulting audio signal.

The major disadvantage of bone conduction microphones is high frequency content is not picked up by the BC mic due to transmission loss through the skull.  A well placed and good fitting microphone will have an upper frequency limit around 2000 to 2500 Hz, but it is not uncommon for BC mics to only capture up to 750 to 1000 Hz.  To overcome this limitation, it is possible to utilize bandwidth extension and other sound synthesis techniques, but the resulting speech sounds robotic and unnatural. The common approach is to combine a BC mic signal with an AC mic, but utilizing multiple AC microphones is best.

Utilizing multiple air-conduction microphones with a bone conduction microphones combines the best of both worlds.  The SNR improvement of beamforming at low frequencies is minimal due to the low spatial resolution, but above 1000 Hz the directional noise cancellation features of beamforming start to bare their fruit. Therefore, combining the BC mic with a good SNR below 1000 Hz and the beamformed AC mics above 1000 Hz results in a noise robust speech signal.  In addition, the BC mic provides reliable voice activity detection (VAD) information that can be used by the beamforming and other traditional speech enhancement algorithms.

More Information



VOCAL’s optimized software is available for the following platforms. Please contact us for specific speech enhancement supported platforms.

ProcessorsOperating Systems
  • Texas Instruments – C6000 (TMS320C62x, TMS320C64x, TMS320C645x, TMS320C66x, TMS320C67x), DaVinci, OMAP, C5000 (TMS320C54x, TMS320C55x)
  • Analog Devices – Blackfin, ADSP-21xx, TigerSHARC, SHARC
  • PowerPC
  • MIPS – MIPS32, MIPS64, MIPS4Kc
  • ARM – ARM7, ARM9, ARM9E, ARM10E, ARM11, StrongARM, ARM Cortex-A8, Cortex-M1
  • Intel / AMD – x86, x64 (both 32 and 64 bit modes)
  • Google Android
  • Apple iOS / iPhone / iPad & MacOS
  • Unix,  Linux, μClinux, BSD
  • Microsoft Windows ACM / RTC / CE / Mobile
  • Symbian
  • eCOS / eCOSPro
  • Wind River VxWorks
  • Green Hills Integrity
  • Micrium μCOS