Although the occasional malfunction of rotating polarizing elements at the beginning of the SOHO mission, which led to its stop, GOLF instrument has been performing precise and continuous measurements over the last two years. In this work, the nature of the actual GOLF measurement, two monochromatic intensities Ib+ and Ib- on the blue wing of the sodium doublet, is investigated by comparing the temporal phase of the low degree (ell <= 3) p-modes obtained from simultaneous data sets, either from SOHO (GOLF in different operating modes) and from a well known ground-based instrument (MARK-I). It is found that the signals are ``almost'' pure velocity signals whose behaviour can be explained, amongst other possible mechanisms, by a contamination of an intensity-like signal of 14% amplitude. Furthermore is it proved that the ratio (instrumental velocity) as defined in MARK-I instrument (and by extension in IRIS one) do have the same nature as the GOLF blue wing signal Xb.