A method for ``time/frequency'' analysis is shortly described, with a discussion of its limitations in time and frequency resolution, and in sensitivity to noise in the signal. This is then applied to observational data. The data used are the luminosity measurements of the IPHIR experiment, and the velocity measurements performed at the Observatorio del Teide. The application to these two contemporary data sets confirms the reliability of the method and provides additional proof that the observed variation with time is not an effect of noise. The long and continuous data set allows to extract information on the temporal behaviour of low degree solar p-modes. The power of the modes is shown to vary strongly over a time-scale of a few days. These estimates are direct measurements from power variations, unlike those deduced from width measured in Fourier spectra. The temporal behaviour of several modes are compared, mainly showing independent behaviour, but also giving some indications of weak correlations between neighbouring modes.