Distortion of binoculars revisited: does the sweet spot exist?

Sixty years ago, August Sonnefeld of Zeiss reported on observations with experimental telescopes. The goal of his investigation was to determine the ideal amount of distortion applied to optical instruments that are used in combination with the human eye. His studies were inconclusive and partially contradictory. We have picked up this problem once again, adopting a modern point of view about the human imaging process, and supported by computer graphics. Based on experiments with Helmholtz checkerboards, we argue that human imaging introduces a certain amount of barrel distortion, which has to be counterbalanced through the implementation of an equally strong pincushion distortion into the binocular design. We discuss in detail how this approach is capable of eliminating the globe effect of the panning binocular and how the residual pincushion distortion affects the image when the eye is pointing off-center. Our results support the binocular designer in optimizing his instrument for its intended mode of application, and may help binocular users and astronomers better understand their tools.