We estimate photospheric velocities of Type II-P supernovae (SNe) using model spectra created with synow, and compare the results with those obtained either by more conventional techniques, such as cross-correlation, or by measuring the absorption minimum of P Cygni features. Based on a sample of 81 observed spectra of five SNe, we show that synow provides velocities that are similar to those obtained by more sophisticated non-local thermodynamic equilibrium modelling codes, but they can be derived in a less computation-intensive way. The estimated photospheric velocities () are compared to those measured from Doppler shifts of the absorption minima of the Hβ and the Fe iiλ5169 features.
Our results confirm that Fe ii velocities (vFe) have a tighter and more homogeneous correlation with the estimated photospheric velocities than those measured from Hβ, but both suffer from phase-dependent systematic deviations. The same is true for a comparison with the cross-correlation velocities. We verify and improve the relations between vFe, vHβ and vmodel in order to provide useful formulae for interpolating/extrapolating the velocity curves of Type II-P SNe to phases not covered by observations. We also discuss implications of our results for the distance measurements of Type II-P SNe and show that the application of the model velocities is preferred in the expanding photosphere method.