My first attempt at D3js: please excuse me for any errors, mistakes or poor graphics. If you have any suggestions and tips, please let me know (especially if you know why the transitions sometimes do not happen?).
Basic phyics and auditory neuroscience tells us that we need two ears to localize sounds. A sound will arrive earlier and be louder in the ear closest to the sound source. Based on these so-called interaural timing and loudness differences, the brain can compute the sound source location. Early on, however, studies suggested that one ear could also suffice, especially for human listeners that have been deaf in one ear for a long period (monaurally deaf). In 2004, we proved that this was not the case. And while there are some monaural (head shadow and fine-spectral) cues that allow some (not all) of the monaurally deaf to guess a sound source location under certain conditions, these conditions are ambiguous at best, and are presumably almost never present in real, daily life: sounds should be presented at a fixed level, on the hearing side, and only in the horizontal plane.
The single, largest predictor for spatial hearing of the monaurally deaf is sound level, not sound source location. Basically, the monaurally deaf believe that a loud sound should be at the hearing side, as a sound source from the deaf side will be attenuated by the head. In the visualization below, I hope this is clearly illustrated with D3js (also, look at my paper for more details and explanations).
Click on the buttons to see how monaural sound localization is affected when the level of a sound changes. Each circle represents the closest speaker location of a perceived sound, the size of the circle indicates how many responses were made in the vicinity of a particular sound source, and the color indicates the average localization error. Note for example the 60 dBA condition; none of the monaural listeners localized a sound on the deaf side, while the binaural listeners correctly perceived sound source location.