Determination of the calcium spark amplitude distribution is of critical importance for understanding the nature of elementary calcium release events in striated muscle. In the present study we show, on general theoretical grounds, that calcium sparks, as observed in confocal line scan images, should have a nonmodal, monotonic decreasing amplitude distribution, regardless of whether the underlying events are stereotyped. To test this prediction we developed, implemented, and verified an automated computer algorithm for objective detection and measurement of calcium sparks in raw image data. When the sensitivity and reliability of the algorithm were set appropriately, we observed highly left-skewed or monotonic decreasing amplitude distributions in skeletal muscle cells and cardiomyocytes, confirming the theoretical predictions. The previously reported modal or Gaussian distributions of sparks detected by eye must therefore be the result of subjective detection bias against small amplitude events. In addition, we discuss possible situations when a modal distribution might be observed.