Illustration of the effect of bandpass shift and wavelength stretching on the example of SEDs approximating the bulge (panel a) and the disk (panels b and c) that are simulated from z = 0 (black) to z = 1.6 (dark orange). Shaded areas show normalized transmission curves of the U, B, V, Cousins R and I and the NIR H and K filters. Horizontal bars mark the wavelength separation between the [O III]5007 and Hα emission lines (blue and red dot, respectively) at a redshift of 0, 0.8 and 1.6. The SEDs, computed with PéGASE 2, refer to 1 M_⊙ and a Salpeter IMF between 0.1 and 100 M_⊙. a) Stellar SED for a 4 Gyr old stellar population that forms with an exponentially decreasing SFR with an e-folding time of 0.5 Gyr (τ0.5 model). It can be seen that, whereas the emission registered in the K filter increases with increasing z, the ObsF emission in U through R decreases by more than 1 dex (cf. Fig. 2g). b) Stellar SED for a 16 Myr old stellar population of Z_⊙/5 forming at a constant SFR (contSF model). c) As in panel b, however, with the contSF model including both stellar and nebular emission. The drop of the SED shortward of 912 Å is because of the reprocessing of the ionizing LyC radiation into nebular and dust emission (see Fioc & Rocca-Volmerange 1997, for details). For the sake of better readability, emission lines are shown only for the SED at z = 0 and z = 1.6.