A range of techniques has been developed in conventional radiography for the chest, because of technical difficulties in imaging the wide variation of tissue densities. These techniques can differ significantly in tube potential selection, for scatter reduction and in application of manual or automatic exposure control. In this study a geometric chest phantom designed to simulate a chest image, has been used to study image quality on an indirect digital radiography (DR) system. Relative values for the signal-to-noise ratio (SNR), have been derived from analysis of contrast detail objects within the lung, heart and subdiaphragm regions of the phantom. Results showing the variation in SNR and detail perception with tube potential are presented for three radiographic techniques: using a grid, an air gap and no scatter reduction. SNR measurements provided a useful objective measure for assessment of image quality for optimising DR systems.