MEMS Microphones & Mic Arrays

Micro-electromechanical system (MEMS) mics are extremely small mics that are used in a wide range of applications and use a pulse density modulation (PDM) stream as their digital output. MEMS mics are integral components in a wide range of products and systems, including hearing aids, mobile computing, automotive and IOT systems.

The use of multiple mics in a smart device can provide distinct benefits to the end product—for example, using multiple mic inputs for improved signal-to-noise ratio (SNR). It also complicates the testing process used to evaluate and characterize designs owing to the need for simultaneous, multichannel measurements. Additionally, engineers developing systems using beamforming mic arrays, and other DSP-based multichannel audio acquisition applications, depend on a true representation of phase alignment of the input channels to validate their algorithms and designs.

Most Pulse-Density Modulation (PDM)  microphones exhibit extremely good stability, as the MEMS device's diaphragm is etched directly from a silicon wafer. Unlike mylar, this material should not exhibit sensitivities to temperature, humidity, or age. However, this is our opinion based on anecdotal experience working with MEMS microphones. Please note that we have not conducted any methodical study of MEMS microphone calibration requirements.

When testing microphones or systems with embedded microphones (such as a hearing aid or a mobile device), it is essential that frequency response of the power amplifier, the loudspeaker and the acoustic chamber be characterized and compensated for in the stimulus signal, so that the response of the system under test can be accurately measured. This process is called “leveling the acoustic output.” 

In APx, this process requires characterizing the system response using a measurement microphone and a measurement such as Frequency Response or Acoustic Response. The measured response curve is exported, and then imported into the Output EQ and inverted in order to achieve a leveled acoustic output. 

This is best performed in the Acoustic output and Acoustic input configurations, normalizing the level with Auto Gen Level before the characterization sweep, and again after the Output EQ is applied.

An APx audio analyzer with PDM 16 is well suited to research and development of products that use microphone arrays. In addition to a variety of useful audio measurements, the system offers the ability to acquire signals from all microphones simultaneously. These waveforms can be captured as .wav files or exported to MATLAB, which can be useful for evaluating beamforming and speech processing algorithms. And the ability to measure up to 16 PDM microphones simultaneously will be of interest to MEMS microphone manufacturers.