Phased Array - Principles and Applications of Microphone and Acoustic Arrays

In the field of acoustics, it has always been important to precisely control the propagation and reception of sound. Acoustic phased array technology achieves revolutionary control over the spatial characteristics of sound waves by arranging multiple microphones or speakers in a specific geometric structure, complemented by sophisticated signal processing algorithms. This allows the system to "hear more accurately" and sound to "be transmitted more accurately". In this blog, let's explore the basic principles, key technologies (such as beamforming), and the widespread use of microphone arrays and speaker arrays.

What is a phased array?

An acoustic array is a system consisting of multiple acoustic sensors (microphones or speakers) arranged in a specific geometric configuration. This structural design allows the array to perform complex functions that are unachievable with a single sensor.

The main advantage of an acoustic array is its spatial processing capability. By coordinating the work of multiple sensors, it provides precise control or analysis of the sound propagation direction and coverage area. Acoustic grilles can focus sound in specific directions for transmission/reception, amplify target signals, suppress noise interference, and control the direction and range of sound propagation.

Based on the type of sensors used and their main function, acoustic grilles are mainly divided into two categories:

Microphone grille: Focuses on receiving, collecting, and analyzing sound.

Speaker grille: Focuses on emitting, reproducing, and controlling sound.

What is a Microphone Array?

A microphone array draws inspiration from the binaural principle that humans use to localize sound sources based on time differences. It achieves powerful spatial acoustic processing by synchronously sampling sound signals with multiple microphones and applying advanced signal processing techniques (primarily beamforming). The main functions of a microphone array are mainly manifested in the following three aspects:

Sound source localization

This function aims to determine the exact spatial coordinates of a sound event. Since sound travels at a finite speed, sound waves from the same source reach microphones at different positions in the array at slightly different times. This difference is called the Time Difference of Arrival (TDOA) or "time delay".

The basic idea of ​​beamforming is to apply adjustable "artificial time delays" to compensate for each microphone channel. By adjusting these compensation values, the system attempts to align (make in-phase) signals coming from an assumed direction across all microphone channels. When the signals from all channels are equalized and summed, the total output power for that direction is maximized.

The system scans different points in space to find a combination of "artificial time delays" that maximizes the output power. Based on this optimal combination of compensation and the known geometric structure of the microphone array, the actual spatial position of the sound source causing this time difference can be calculated. This process essentially uses the time delay information for spatial inversion.

Key application scenarios:

1. Intelligent traffic enforcement: Whistle detection systems accurately locate vehicles violating horn usage regulations; illegal vehicle detection systems track the location of the source of growling exhaust noise.

2. Industrial equipment monitoring: Real-time localization of abnormal noise points (e.g. bearings, gearboxes, pipes) in factories for predictive maintenance and fault diagnosis (e.g. detecting bearing wear, specific frequency noise from gas leaks).

3. Environmental noise monitoring: Noise monitoring systems in communities or cities quickly identify and locate sources of nuisance noise (e.g. construction noise, noise from entertainment venues), improving enforcement efficiency.

Directional Audio Capture

This feature aims to enhance the audio signal from a specific target direction while suppressing interfering noise and ambient sounds from other directions, thereby improving the signal-to-noise ratio (SNR) of the target audio.

Directional capture also relies on beamforming technology. The system pre-determines (or dynamically tracks) the direction of the target audio source and calculates the optimal "artificial time delay" compensation values ​​for that direction.

After applying these compensation values, audio signals from the target direction are aligned across the microphone channels and significantly enhanced by in-phase summation. Audio signals from non-target directions (including interfering noise and diffuse ambient sound) that cannot be aligned by this specific set of delay compensations are subject to varying degrees of suppression or attenuation by summation. This significantly improves the clarity and intelligibility of speech or audio from the target direction, achieving "directional audio focusing".

Typical application scenarios:

1. Far-field voice interaction: Smart speakers, smart TVs, and video conferencing systems can clearly capture a user's voice commands or speech from across a room (typically several meters), reducing the influence of ambient noise.

2. High-definition conference recording and speaker separation: In conference rooms, systems can target specific speakers (e.g., chairperson, current speaker) or form independent pickup beams for speakers in different locations, providing "speaker separation" recording for clearer, more traceable meeting minutes.

3. Professional outdoor recording: Effectively suppresses background noise such as wind and road noise in noisy outdoor environments (e.g., news reporting locations, wildlife viewing) to clearly capture sound from specific target objects (e.g., interviewees, specific animals).

4. Security Surveillance: Works with cameras to specifically capture specific sounds (e.g. abnormal cries for help, broken glass) in the area being monitored, improving surveillance efficiency.

High Definition Far-Field Capture and Reverberation Suppression

The main problem for far-field capture is reverberation interference. Reverberation suppression technology aims to eliminate or reduce reverberant sound caused by room reflections while preserving and enhancing direct sound to improve clarity when capturing at long range.

When microphones are far from the sound source, the strength of the direct audio signal is weakened, while the energy contribution of reverberation formed by multiple reflections from walls, ceilings, floors, etc., increases significantly. Strong reverberation causes blurred speech and slurred syllables, significantly reducing speech recognition accuracy and auditory clarity.

To improve the clarity of capture at long range, the key is to apply reverberation suppression technology, which aims to effectively suppress or eliminate these harmful reverberant components while preserving and enhancing the direct audio signal from the source. A widely used and highly effective solution is Multi-Channel Linear Prediction (MCLP). The basic idea of ​​this method exploits the fundamental statistical differences, especially in sparsity, between the real speech signal (mainly direct sound and early reflections) and late reverberation - real speech signals typically exhibit higher sparsity (more concentrated energy distribution) in the time-frequency domain than late reverberation. By analyzing the spatial correlation between multi-channel signals and reverberation characteristics, the MCLP method establishes a linear prediction model to estimate and separate the reverberant components, ultimately outputting significantly clearer speech signals.

Technical implementation:

1. Modeling: MCLP uses the spatial correlation between signals from multiple microphone channels and acoustic characteristics of the reverberation (e.g. room impulse response - RIR models) to establish a linear prediction model.

2. Prediction and separation: This model is used to predict the reverberant components in the microphone signal at the current time (basically using information about the past signal). The prediction is based on the specific correlation patterns that the reverberation exhibits across multiple channels.

3. Estimation and suppression: The predicted reverberant components are subtracted from the original microphone signal, resulting in an estimated, relatively clean direct audio signal (speech with suppressed reverberation).

What is a Speaker Grid?

A speaker grid technology consists of a group of speakers arranged in a specific geometric pattern (e.g., straight line, curve, plane) and working together. By independently and precisely controlling the signal (amplitude and phase/delay) for each element in the grid, it achieves active control over the sound pattern, overcoming the limitations of traditional point amplification. The main application functions of a speaker grid are mainly reflected in two aspects: directional sound reinforcement and constant sound pressure level coverage.

Directional Sound Reinforcement

Directional sound reinforcement aims to concentrate the emission of sound energy toward a specific target area or direction, reducing energy leakage and reflections to non-target areas.

A plurality of speakers forming a grid are physically equivalent to increasing the effective size (aperture) of the sound source. According to acoustic principles, a larger sound source inherently provides better directionality control (i.e. more concentrated sound energy).

To achieve more precise and flexible directionality control, sound field reconstruction techniques are used. A dedicated digital filter can be designed for each driver in the array. These filters independently control the amplitude (gain) and phase (delay) of the audio signal fed to each element.

By precisely controlling the amplitude and phase relationship of the signals for each element, the overall sound waves generated by the array can be directed to interfere constructively (enhance) and interfere destructively (cancel) in space. This precisely "steers" or "focuses" the main lobe of the sound wave (the direction of greatest energy) toward the desired target direction.

The linear array structure effectively solves the problem of longitudinal sound field homogeneity. Its basis is independent control of amplitude (loudness) and phase (delay) for each driver. By precisely adjusting the amplitude and phase, it exploits the interference of sound waves to achieve a nearly constant overall sound pressure across the longitudinal depth. Specific strategy:

1. Far-end constructive summation (rear rows): Precisely control the phase relationship of sound waves arriving at distant positions (e.g. rear rows) to make them as in-phase as possible. In-phase waves are summed constructively, significantly increasing the SPL to compensate for natural attenuation.

2. Near-end destructive cancellation (front rows): Precisely control the phase relationship of sound waves arriving at close positions (e.g. front rows) to make them partially out of phase. Out-of-phase waves are destructively canceled, reducing the SPL in that area. By carefully designing the amplitude weighting and phase delay for each element (upper elements have higher power and lower delay; lower elements have lower power and longer delay), the interference of sound waves maintains a nearly constant overall SPL over large longitudinal distances from near to far. This overcomes the limitation of the inverse square law. It is widely used in large halls (e.g. concert halls, theaters, churches, conference centers, stadiums, train stations) that require uniform coverage across the depth, providing clear, comfortable and consistent loudness to all spectators.

Conclusion

Acoustic phased array technology arranges multiple acoustic transducers (microphones or speakers) in a specific geometric structure to achieve functions that a single transducer cannot perform. Its basis is divided into two categories: microphone array and speaker array. Microphone array uses beamforming technology to process signals collected synchronously by multiple microphones. The speaker grille changes the shape of the sound field radiation by independently and precisely controlling the amplitude and phase (delay) of the signal of each element in the grille. Do you experience the problem of uneven sound field distribution? Need to achieve precise directional sound reinforcement? Contact us for customized solutions.