Defining mechanisms for cardiac fibrillation is challenging because, in contrast to other arrhythmias, fibrillation exhibits complex non-repeatability in spatiotemporal activation but paradoxically exhibits conserved spatial gradients in rate, dominant frequency, and electrical propagation. Unlike animal models, in which fibrillation can be mapped at high spatial and temporal resolution using optical dyes or arrays of contact electrodes, mapping of cardiac fibrillation in patients is constrained practically to lower resolutions or smaller fields-of-view. In many animal models, atrial fibrillation is maintained by localized electrical rotors and focal sources. However, until recently, few studies had revealed localized sources in human fibrillation, so that the impact of mapping constraints on the ability to identify rotors or focal sources in humans was not described. Here, we determine the minimum spatial and temporal resolutions theoretically required to detect rigidly rotating spiral waves and focal sources, then extend these requirements for spiral waves in computer simulations. Finally, we apply our results to clinical data acquired during human atrial fibrillation using a novel technique termed focal impulse and rotor mapping (FIRM). Our results provide theoretical justification and clinical demonstration that FIRM meets the spatio-temporal resolution requirements to reliably identify rotors and focal sources for human atrial fibrillation. Atrial fibrillation (AF) is the most common heart rhythm disorder (cardiac arrhythmia) in the United States that may cause substantial morbidity and mortality. However, the precise mechanisms causing AF are still not well understood, partly due to difficulties in reliably mapping electrical activity during the spatio-temporal variations of AF in patients. In this article, we determine the minimal spatial and temporal resolution required to accurately map in humans the electrical rotors and focal sources shown to sustain AF in various model systems. We first test these requirements in computer simulations. We then validate that a recently developed mapping technique which employs bi-atrial multielectrode contact arrays is able to capture localized rotors and focal sources for human AF.