Sound localization ability has traditionally been studied using either a relative localization task, where thresholds to determine a difference in sound source location is approximately 1-10 degrees, or an absolute localization task, where the range of estimates of the source of a sound are 4-30 degrees. In order to directly relate these two psychophysical methods, we compared the psychometric functions from a relative localization task in a human subject to the same subject's performance on an absolute localization task using three different acoustic stimuli: Gaussian noise, 1-kHz tones, and 4-kHz tones. The results showed that the relative localization threshold was a poor indicator of the range of estimates of the same stimulus in absolute space, however, the width of the relative localization psychometric functions was well correlated with the width of the distribution of estimates made in the absolute localization task. It is concluded that the relative localization psychometric functions, but not threshold, provides a reliable estimate of absolute spatial localization ability in human subjects, and suggested that the same neuronal mechanisms can underlie the psychophysical data using both methods.