A study on the Raman shift and width of nitrogen and nitrogen in helium has been performed as a function of pressure and temperature by means of experiments, molecular dynamics (MD) simulations and hard fluid (HF) theory. The experiments have been performed using Raman spectroscopy in a diamond anvil cell at pressures up to 10 GPa and temperatures between 250 and 400 K. Both the experimental shift and width results of pure nitrogen link up very well with accurate measurements at lower pressures and with less accurate measurements at higher pressures. For the first time the Raman shift and width have been determined as a function of temperature at an isobar, such that a sensitive test of theoretical models can be made. The MD calculations on the linewidth along an isobar show very good agreement with experiment. The influence on the linewidth of the bondlength dependence of the site-site interaction parameters (often called the attractive contribution) appears to be small, which indicates that this has a small anisotropy. For pure N-2 the MD and the HF calculations of the repulsive contribution to the Raman shift are about the same. This shows that both ways of calculation are consistent. The experimental Raman shift of nitrogen diluted in helium appears to be much larger than that of pure nitrogen. In contrast, the linewidth is much smaller than that of pure nitrogen. HF calculations were also performed for the Raman shift of N-2, infinitely diluted in He. The results for the bondlength independent (repulsive) contribution give clearly smaller values than those of the experiment, which means that the effect of the change of the potential parameters at excitation must be positive. This implies that that part of the intermolecular potential, which is due to the overlap of the molecular charge distributions has a dependence on the bondlength, that results in a positive contribution to the Raman shift. It will be shown that for N-2 the good agreement with experiment of earlier HF calculations with an attractive contribution, based on a purely dispersive model, is due to a cancellation of errors. For nondiluted mixtures of He-N-2 under noncritical conditions the plot of experimental FWHM values as a function of the volume fraction shows a broad maximum, which is indicative for inhomogeneous broadening. This behavior is described with the help of the Knapp-Fischer model.