We investigated the practical impact of the representation of the cardiac rhythm--the cardiotachogram--on two elementary spectral indexes in heart rate variability analysis: the low-frequency (0.07-0.14 Hz) and high-frequency (0.14-0.40 Hz) powers. Five commonly used tachograms (inter-beat interval function/series, counts, instantaneous heart rate function/series) were compared. Measurements were done on seven volunteers in the supine and standing positions. Ratios, and their deviations from 100%, of alternative power values were calculated. Mean low-frequency and high-frequency ratio deviations ranged from 0 to 5% and from 6 to 37%. The spectrum of counts yielded on average more (15-37%) high-frequency power. Spectra were incomparable without normalization of the tachogram with respect to heart rate. In conclusion, (i) the choice of a particular spectrum may lead to differing conclusions on the vagal contribution to heart rate variability and (ii) inconclusive results from studies using different tachogram variants can partly be due to the omission of normalization.