MEMS microphones are tiny, low-cost and extremely versatile, with characteristics that allow use in a wide variety of sound-recording applications.
Their natural omnidirectional response makes them equally receptive to sound coming from any direction. Although this is perfect for some applications, such as capturing the sound from a moving source, it is less well suited to isolating the sound coming from a single source at a fixed location, such as recording the sound of one person speaking or singing. This means that ambient noise or unwanted sounds from other directions can easily interfere with the wanted signal.
To overcome this, two or more microphones can be arranged as a beamforming array, using digital signal processing to constructively sum wanted signals, while attenuating unwanted sounds. Basic configurations include the broadside array, which contains microphones positioned on an axis perpendicular to the waveform from the wanted source. Another is the endfire array, which arranges the microphones one behind the other, to be pointed directly at the source.
This CUI Insights blog, “An Introduction to MEMS Microphone Arrays”, gives an introduction to beamforming, discussing basic configurations and the importance of closely matched microphone sensitivity and frequency response.