There are obvious differences in the automatic selection of kVp, anode, and filter between a digital mammography system and the equivalent film/screen system for the same thickness of PMMA absorber. To investigate the reason for these changes, a large number of images were acquired using 2, 4, 6, and 8 cm thick PMMA absorbing slabs, and various combinations of kVp, mAs, anode, and filter. The SNR and CNR were calculated using two different contrast test objects and plotted as a function of AGD. The results can be summarized as follows: 1) For any given AGD with 4, 6, and 8 cm thick PMMA absorbers, SNRRh/Rh > SNRMo/Rh > SNRMo/Mo, 2) For any given AGD: At 4 cm thickness, CNRMo/Mo > CNRtMo/Rh > CNRtRh/Rh, At 6 cm thickness, CNRRh/Rh ≈ CNRMo/Rh > CNRMo/Mo, At 8 cm thickness, CNRRh/Rh > CNRMo/Rh, 3) For any given absorber thickness and target/filter combination, CNR is approximately proportional to SNR, 4) For any given absorber thickness and target/filter combination, SNR (and hence CNR) is approximately proportional to (AGD)0.3 rather than the expected (AGD)0.5, 5) CNR measured using 0.1 mm thick aluminium foil as the contrast object is more dependent on choice of kVp than using a 1 mm deep hole in a PMMA slab.