Head-Related Transfer Functions (HRTF) characterize the scattering of an incident wave by a listener's head. Measurements of these HRTF is a complex process: the required number of positions is generally high, and the measurement setup could induce a specific scattering (not intended). Especially in the proximity of the listener the source distance has an in influence, and the total number of measurement positions could be superior to a thousand. Measurements of dummy heads are easier than measurements of real listeners, because of the simplified control of the position. However, measurement of dummy heads is challenging, several measurements campaign were done in the literature but not directly comparable because of the variability of experimental setups. In this paper, we present the HRTF measurements of five dummy heads using the same experimental setup. In total, 794 source positions mainly in the half-sphere of positive elevations were measured. These dummy heads were all built to approximate an "average listener", and then should be similar. One of these dummy head was a rigid sphere allowing a comparison to the analytical simulation and deduce an estimation of measurement errors.