This study investigates the airborne acoustic emissions detected by an optical microphone during the laser welding process with the goal of improving quality assurance through in situ monitoring. The acoustic emissions provide valuable insight into the keyhole formation and dynamic behavior of the welding process. A detailed analysis of the influence of process gas revealed that using compressed air introduced significant background noise, while argon gas supplied via a coaxial nozzle allowed the optical microphone to accurately record the process sound from laser-material interaction. Spectrogram analysis in the frequency domain highlighted distinct acoustic signatures corresponding to keyhole formation events. However, no direct correlation was found between the intensity of acoustic emissions and laser power, either for individual keyhole events or the average signal intensity over time. Furthermore, different FFT window sizes (1024, 2048, 4096) were tested. The results showed that the window size of 1024 did not provide adequate frequency resolution, while the window size of 4096 introduced oscillations at higher frequencies. The window size of 2048 offered the best balance, minimizing artifacts while retaining sufficient frequency resolution. These findings suggest that optical microphones with optimized signal processing can be a valuable technique for monitoring acoustic signatures of laser welding.