Rotor wake-stator interaction is an important element of aircraft engine noise, especially in the rear arc. For numerical predictions of this type of engine noise, an accurate rotor wake description is indispensable. However, for most CFD codes accurate calculation of the development of turbulent rotor wakes is not something natural. To assess the importance of wake modelling accuracy, the NLR lifting surface model was used for a parametric study. Many rotor wake-stator interaction calculations were made on a configuration with realistic dimensions. Systematic variations of wake depth, wake width, and axial and circumferential position of the wake origin were carried out. It was found that for 1 dB precision in the final acoustic result these parameters must have a relative accuracy varying from 6% to 12%. This is a real trial of strength for most CFD models. With the same 1 dB precision requirement, it was found that relative errors in the prediction of rotor viscous drag, which is an important factor in the wake development, are acceptable up to 23%.