Diastolic Ca leak from the sarcoplasmic reticulum (SR) of ventricular myocytes reduces the SR Ca content, stabilizing the activity of the SR Ca release channel ryanodine receptor for the next beat. SR Ca leak has been visualized globally using whole-cell fluorescence, or locally using confocal microscopy, but never both ways. When using confocal microscopy, leak is imaged as "Ca sparks," which are fluorescent objects generated by the local reaction-diffusion of released Ca and cytosolic indicator. Here, we used confocal microscopy and simultaneously measured the global ryanodine-receptor-mediated leak rate (J(leak)) and Ca sparks in intact mouse ventricular myocytes. We found that spark frequency and J(leak) are correlated, as expected if both are manifestations of a common phenomenon. However, we also found that sparks explain approximately half of J(leak). Our strategy unmasks the presence of a subresolution (i.e., nonspark) release of potential physiological relevance.