When using microfocus x-ray sources for x-ray imaging (x-ray projection microscopy or microradiography), the measured intensities are influenced by the non-monochromaticity of the incident x-ray beam. This affects the transmitted signal in the image pixels and consequently brings about errors in quantitative measurements by x-ray absorption analysis or in tomographic reconstruction. A model developed to predict the spectral distribution of x-rays generated by electron bombardment on a metallic target has been modified to be adapted to transmission targets. It is used here, in combination with a semi-empirical analytical expression for mass attenuation coefficients, to calculate the overall energy transmitted to an x-ray imaging system and thus to simulate any x-ray projection experiment. In particular, spectral purity, which is a measure of the degree of monochromaticity, can be easily evaluated. In this paper simulations are used to evaluate the effect of each part of an imaging system on the spectral purity. Here we are concerned with the effect of target and/or filter thicknesses and the influence on the purity of the selective absorption efficiency of phosphor screens with x-ray energy. As an example, three targets and two kinds of phosphor screen widely in use in x-ray CCD cameras have been considered.