Oscillometric noninvasive blood pressure (NIBP) devices determine pressure by analysing the oscillometric waveform using empirical algorithms. Many algorithms analyse the waveform by calculating the systolic and diastolic characteristic ratios, which are the amplitudes of the oscillometric pulses in the cuff at, respectively, the systolic and diastolic pressures, divided by the peak pulse amplitude. A database of oscillometric waveforms was used to study the influences of the characteristic ratios on the differences between auscultatory and oscillometric measurements. Two hundred and forty-three oscillometric waveforms and simultaneous auscultatory blood pressures were recorded from 124 patients at cuff deflation rates of 2-3 mmHg/s. A simulator regenerated the waveforms, which were presented to two NIBP devices, the Omron HEM-907 [OMRON Europe B.V. (OMCE), Hoofddorp, The Netherlands] and the GE ProCare 400 (GE Healthcare, Tampa, Florida, USA). For each waveform, the paired systolic and paired diastolic pressure differences between device measurements and auscultatory reference pressures were calculated. The systolic and diastolic characteristic ratios, corresponding to the reference auscultatory pressures of each oscillometric waveform stored in the simulator, were calculated. The paired differences between NIBP measured and auscultatory reference pressures were compared with the characteristic ratios. The mean and standard deviations of the systolic and diastolic characteristic ratios were 0.49 (0.11) and 0.72 (0.12), respectively. The systolic pressures recorded by both devices were lower (negative paired pressure difference) than the corresponding auscultatory pressures at low systolic characteristic ratios, but higher than the corresponding auscultatory pressures at high systolic pressures. Conversely, the differences between the paired diastolic pressure differences were higher at low diastolic characteristic ratios, compared with those at high diastolic characteristic ratios. The paired systolic pressure differences were within +/-5 mmHg for those waveforms with systolic characteristic ratios between 0.4 and 0.7 for the Omron and between 0.3 and 0.5 for the ProCare. The paired diastolic pressure differences were within +/-5 mmHg for those waveforms with diastolic characteristic ratios between 0.4 and 0.6 for the Omron and between 0.5 and 0.8 for the ProCare. The systolic and diastolic paired oscillometric-auscultatory pressure differences varied with their corresponding characteristic ratios. Good agreement (within 5 mmHg) between the oscillometric and auscultatory pressures occurred for oscillometric pulse amplitude envelopes with specific ranges of characteristic ratios, but the ranges were different for the two devices. Further work is required to classify the different envelope shapes, comparing them with patient conditions, to determine if a clearer understanding of the different waveform shapes would improve the accuracy of oscillometric measurements.