The analysis of synthetic aperture radar observations over the ocean to derive the directional spectra of the waves is based upon a complex transfer function which is the sum of three terms: tilt modulation, hydrodynamic modulation, and velocity bunching effect. Both the hydrodynamic and the velocity bunching terms are still poorly known. Here we focus on the hydrodynamic part of the transfer function, from an experimental point of view. In this paper a new method is proposed to estimate the hydrodynamic modulation. The approach consists in analyzing observations obtained with an airborne real-aperture radar (called RESSAC). This radar (C band, HH polarized, broad beam of 14¿¿3¿) was used during the SEMAPHORE experiment, in two different modes. From the first mode (incidence angles from 7¿ to 21¿) the directional spectra of the long waves are deduced under the assumption that the hydrodynamic modulation can be neglected (small incidence angles) and validated against in situ measurements. From the second mode (incidence angle from 27¿ to 41¿) the amplitude and phase of the hydrodynamic modulation are deduced by combining the measured signal modulation spectrum at a mean incidence angle of 34¿ and the directional wave spectrum obtained from the first mode. The results, obtained in four different wind-wave cases of the SEMAPHORE experiment, show that the modulus of the hydrodynamic modulation is larger than that of the tilt modulation. Furthermore, we find that the modulus of the hydrodynamic transfer function is several times larger (by a factor 2--12) than the theoretical value proposed in previous works and 1.5--2.5 larger than experimental values reported in recent papers. The phase of the hydrodynamic modulation is found to be close to zero for waves propagating at an angle from the wind direction and between -20¿ and -40¿ for waves propagating along the wind direction. This indicates a significant influence of the wind-wave angle on the phase of the hydrodynamic modulation, in agreement with experimental results reported in recent papers. ¿ American Geophysical Union 1996