The case of the missing 6 dB.

In 1933 data were reported which indicated that thresholds of hearing for frequencies apparently depended upon whether the source was an earphone (MAP) or a loudspeaker (MAF). A decade or so later the same type of discrepancy appeared when loudness balances were made at 100 Hz between an earphone source and a loudspeaker source. In both cases approximately 6 dB more sound pressure level at the eardrum was required when the earphone was the source than when the loudspeaker was the source. Later research added credence fo this paradox, namely, the ear should act as a pressure operated device, and there should be no difference between MAP and MAF; yet a difference seemed to exist. Research reported in abstract form and orally by the author in 1962 and 1963 showed that (a) the difference at threshold was due to physiological noise generated in the ear canal by the earphone-cushion-head combination (and could be eliminated with a special earphone-coupling system), and (b) the suprathreshold differences obtained with loudness balancing were due to a number of subtle procedural and experimental techniques (techniques which could be modified so as to avoid all of the problems of past experimenters). This research is reported here for the first time in full detail. A total of 15 different subjects participated in eight experimental comparisons using three to nine subjects each, with sufficient replications so that most subjects' MAF-MAP and/or loudness differences were determined within 1 or 2 dB at the 95% confidence level. It was often possible to replicate previous results using previous methods, but with the modified methods reported here the average difference across experiments was less than 0.2 dB, and no subject in any experiment exhibited more than a 1.8-dB difference averaged across trials. The case of the missing 6 dB should be considered closed.