This study presents an innovative beamforming method tailored for audio surveillance applications, developed through virtual simulations conducted at Liverpool John Moores University. Motivated by the growing need for advanced audio analysis techniques, this method focuses on segregating and enhancing particular sounds within acoustically cluttered environments, such as those presented in forensic scenarios, including criminal court cases. Utilising a time-delay beamforming algorithm, this study introduces a novel strategy to identify and magnify specific noises within intricate acoustic settings, addressing challenges often faced in surveillance and forensic audio examinations. The foundation of our technique lies in the strategic deployment of a robust omnidirectional microphone array, which is crucial for capturing environmental sounds. Our methodology involves the application of a MATLAB algorithm to process these sounds, followed by comprehensive evaluations to gauge the system's effectiveness in isolating targeted audio sources. The investigation examines the system's robustness against microphone array degradation, demonstrating its reliability even with compromised functionality. Simulations of real-world acoustic conditions reveal the algorithm's effectiveness in handling sound reflections and reverberations, crucial for urban acoustic landscapes. This study introduces a novel beamforming method for audio surveillance, with potential applications in broadcasting, advanced audio engineering, and wildlife conservation, highlighting its versatility. In summary, this research introduces a novel approach to audio surveillance, paving the way for numerous practical applications that could leverage improved audio isolation and analytical capabilities. Our results contribute significantly to the ongoing development of sophisticated surveillance technologies, offering valuable perspectives that could influence the future landscape of audio engineering and analysis. 

URL: https://ieeexplore.ieee.org/abstract/document/10701361