\'H/V\' techniques are very attractive tools for estimating local site effect characteristics. They consist in computing the spectral ratio between the horizontal and the vertical components of a signal. For the \'receiver function\', or RF method, the signal is an earthquake, whereas it is composed of ambient noise for the NN method. The classical transfer function method, called SR here, is based on the spectral ratio of an earthquake recording between one site and a reference. The aim of this paper is to collect experimental results (44 sites) and compare values worthy of note obtained from the H/V and the SR techniques. For several typical sites, we first illustrate that, when a clear peak arises on an SR curve, it also exists, at the same frequency for RF or NN curves. But there is no correlation between the two kinds of curves except for the low frequency part (below the first peak) where a fuzzy relationship seems to exist. Finally, we show that NN and RF techniques determine, very accurately, the fundamental frequency of alluvial sites, below which there is no amplification. In most cases, they also provide a lower bound (in amplitude) and bandwidth (in frequency) estimates for peak amplification.