2514 20160701171706.0 ASTROimport-671 Hensberge, H. 2004 Do our spectra match the requirements for a precise analysis of SB2s? In the last decade of the 20th century, powerful techniques to resolve the spectra of multiple light sources into their components were developed. Non-physical undulations may appear in the resulting component spectra. These undulations may be triggered, at least partly, by systematic errors in input spectra. The whole data reduction process of echelle spectroscopy is critically reviewed with the purpose to encourage the user to evaluate, empirically, sources of systematic errors in spectra. Such checks should be done on extracted calibration and science spectra before normalising and combining single spectral orders. Furthermore, in cases where many spectra of an object will be combined in one analysis, a differential data reduction procedure is recommended. Spectroscopically and Spatially Resolving the Components of the Close Binary Stars 318 2004 43-51 http://esoads.eso.org/abs/2004ASPC..318...43H published in NONREF