The continued high interest in the effects of man-made sounds on marine wildlife has resulted in increased use of playback experiments. In such experiments, the intent is often to simulate a larger more powerful source of sound, e.g., a drill ship, with a playback of a recording, but using a small portable source system such as a J-11 to play the sound back. Simulation accuracy is assumed to be met in these playbacks if the receive level of the sound at the animals being observed is the same as they would be exposed to by the full scale source. To achieve this same receive level the playback source must often be moved to within short range of the animals. The purpose of this paper is to identify and discuss those components of the actual full scale exposure that may not be accurately simulated in an experiment where sound level at the animal is the only control variable. Components examined include the temporal, spectral, and spatial properties of the noise field, as well as the relative motion of animal and source with emphasis on the highly variable nature of the noise field at short playback ranges. Recommendations are made to assist experimental designers in developing more accurate simulations.