A dual-microphone array (DMA) guide covers the essential engineering principles, hardware topologies, and digital signal processing (DSP) trade-offs required to design effective two-mic voice and audio capture systems. Dual-mic arrays serve as the fundamental framework for modern consumer electronics—including smartphones, smart speakers, laptops, and wearables—balancing low component costs with high performance. Core Hardware Topologies
Engineers typically implement two distinct physical geometries when placing a pair of microphones. The spatial relationship between the sound source and the microphone axis defines the primary functional layout:
Broadside Array: Microphones are placed on a line perpendicular to the incoming target sound wave. Sound arrives at both elements simultaneously (in-phase). This configuration is ideal for laptops or wall-mounted smart panels where the user sits directly in front of the device. It provides strong side-lobe noise suppression but cannot differentiate between sound originating from the exact front versus the exact back.
Endfire Array: Microphones are aligned directly on the axis of the incoming target sound wave. Sound hits the front microphone first, introducing a physical time-of-flight delay before reaching the second. This layout is highly effective for close-talk applications like Bluetooth headsets or handheld smartphones, as it naturally rejects sounds coming from the rear or sides. Physical Design Constraints and Formulas
The primary variable an engineer must manipulate is the inter-element spacing (
), which determines the frequency threshold of the entire array. Key Variables: : Distance between the two microphone centers (m) : Speed of sound in air ( 20∘C20 raised to the composed with power C fmaxf sub max end-sub : High-frequency threshold before degradation (Hz) : Wavelength of the target frequency (m)
Acoustics Formula Blocks:The relation between sound speed, frequency, and wavelength: c=f⋅λc equals f center dot lambda
To prevent spatial aliasing—a phenomenon where the processing algorithm cannot distinguish between different arrival angles and introduces unwanted artifacts—the microphone spacing must not exceed half of the shortest target wavelength:
d≤λmin2=c2⋅fmaxd is less than or equal to the fraction with numerator lambda sub min end-sub and denominator 2 end-fraction equals the fraction with numerator c and denominator 2 center dot f sub max end-sub end-fraction Design Trade-off: Increasing the distance (
) improves the signal-to-noise ratio (SNR) and directional accuracy for low-frequency sounds. However, a larger distance pushes the spatial aliasing threshold down into the human vocal range. For example, hardware guidelines from platforms like Espressif System Design Guides suggest a practical physical spacing of 40 mm to 65 mm for voice recognition applications to keep spatial aliasing above the critical threshold. Digital Processing Techniques
Dual-mic arrays rely heavily on algorithms to combine the two raw audio streams into a single, clean track:
Leave a Reply