In vision, it is known that humans use summary statistics to efficiently perceive and encode the ‘gist’ of groups of features. For instance, after viewing a set of differently sized circles, people can reliably estimate the average circle size, often more accurately than they can identify an individual member of the set (Ariely, 2001). Summary statistical encoding (i.e., ensemble coding) is common in visual processing, having been demonstrated not only for low-level visual features (e.g., size and orientation) but also for high-level features such as facial expression (Haberman & Whitney, 2007). Here, we present evidence that ensemble coding is operative in audition. Specifically, participants were able to estimate the mean frequency of a set of logarithmically spaced pure tones presented in a temporal sequence, even though they performed poorly when asked to identify an individual member of the set (an identification task) or identify a tone’s position in the set (a localization task). This suggests that ensemble coding is not limited to visual processing and instead is an important, multisensory mechanism for extracting useful information from groups of objects.