Measurement of the 2f1–f2 cubic difference tone with the binaural masking-level difference

Stimulus-like properties of the 2f1-f2 cubic difference tone (CDT) were examined by determining whether a binaural masking-level difference (BMLD) could be obtained for a 500-Hz physical tone to one ear and a CDT of the same frequency to the other ear. Masked thresholds for the physical tone (signal) were obtained as the phase of this tone varied through 360 degrees relative to harmonically related primaries (f1 = 625 Hz, f2 = 750 Hz). A continuous low-pass (2000 Hz) diotic noise comprised the masking stimulus. The level of this noise was varied to establish thresholds in a two-interval same-different adaptive procedure with the primaries occurring in both observation intervals. In keeping with other stimulus-like properties of the CDT, the results show BMLDs of 10-13 dB. Additional BMLDs were obtained with a method of adjustment to derive phase and level estimates of the CDT as a function of primary level. These estimates are compared to cancellation estimates obtained from the same subjects. Both BMLD and cancellation estimates of CDT level show a 10-dB growth in the CDT with each of 10-dB increase in primary level. However, cancellation-level estimates are consistently above corresponding BMLD-level estimates, greater by an average of 4-14 dB, depending on the subject. BMLD and cancellation estimates of CDT phase show a near-identical dependence on primary level, decreasing by an average of 5 degrees/dB. The agreement between the two procedures argues against the notion that the phase dependence is the result of confounding peripheral interactions between the cancellation tone and the primaries in the cancellation procedure.