Distributed propulsion systems are developed to power a new generation of aircraft. However, it is not known yet which noise emissions these propulsion systems produce, which psychoacoustic characteristics such systems exhibit, and how the generated noise is perceived. This paper investigates how fans with fewer stator than rotor blades affect the noise perception of a distributed propulsion system intended for an urban air mobility vehicle, which is equipped with 26 low-speed ducted fans. Three fan designs with different tonal to broadband noise ratio and opposite dominant noise radiation directions are examined. An analytical process is applied to determine the noise emission, propagate the sound through the atmosphere, auralize the flyover signals, and calculate psychoacoustic metrics. A validation and comparison with A320 turbofan engines at takeoff is provided. The results indicate that the distributed propulsion system generates noise signatures with complex directional characteristics and high sharpness. By applying tonal noise reduction mechanisms at source, a significant effective perceived noise level reduction is achieved for the considered fan stages with fewer stator than rotor blades. In addition, tonality, loudness and roughness are reduced well above one noticeable difference compared to a baseline fan and similar or even lower values are achieved than with turbofans.

Novel propulsion concepts are being developed for urban air mobility (UAM). This study analyses the noise emissions of a virtual UAM vehicle equipped with 26 distributed, ducted, low-speed fans. Synthetic flyover sounds are generated using an auralization framework, which involves noise predictions, sound propagation, and the binaural audio rendering at the observer position. Since noise emissions of distributed propulsion are characterized by interference effects, this paper discusses the influence of rotational speed fluctuations of the distributed fans on the sound perception of the auralized sounds. These rotational speed fluctuations are modelled as different ranges of random and constant deviations from the nominal speed, in contrast to the baseline case with the 26 fans operating synchronously. The results of a listening experiment performed to evaluate the perceptual differences between the different configurations seem to indicate that relatively small fluctuations in the rotational speed of the fans (e.g. $\pm1\%$ with respect to the nominal rpm) already notably improve the perceived noise annoyance. The observed differences in annoyance ratings are reasonably well explained by sound metrics that consider the tonal nature of sound, such as the effective perceived noise level (EPNL), tonality, and the psychoacoustic annoyance model by Di et al.