The range of combined intersubject and cushion‐fit variation in the response of the TDH‐39 audiometer earphone in an MX41/AR cushion, reported in the literature and confirmed in these measurements, is of the order of 25 dB at 100 Hz, 6 dB at 1 kHz, and 15 dB at 5 kHz. The performance of this earphone system is further impaired by cushion‐induced physiological noise that masks low‐frequency thresholds, typically by 6 dB in normal‐hearing subjects. An experimental mounting for the TDH‐39 is described and analyzed; it is a circumaural device designed to keep the earphone‐ear acoustical system as near constant as possible with successive fittings and different subjects, and to reduce physiological noise. The performance of the TDH‐39 in this mounting, compared to performance in an MX41/AR cushion, showed a substantial reduction in cushion‐fit and intersubject response variation, and masking by physiological noise was largely eliminated in the subjects tested. The mounting has an optional cup for reducing the transmission of external noise, but this reduction is at the expense of a partial return of physiological noise masking. Average midrange sensitivity was reduced 5–6 dB. Response measurements were made with a free‐field reference method, using a probe microphone at the subject's eardrum.