2.4. Experiment 2

Experiment 2 is divided into two parts: experiment 2A and experiment 2B. For experiment 2A, the targets were the 1st–15th harmonics of a 250-Hz fundamental frequency. Three conditions were tested in addition to the baseline and same conditions. In the noise condition, the nontarget sounds were broadband Gaussian noise (after Lutfi and Jesteadt, 2006). The noise was not low-pass filtered prior to headphones. In the low and lower conditions, the nontarget sounds were the 1st–15th harmonics of f_o = 198 and 176 Hz, respectively. Note that because the harmonic numbers are the same, higher fundamental frequencies yield stimuli with wider bandwidths. In experiment 2B, the targets were the third–seventh harmonics of a 250-Hz fundamental. In addition to the baseline condition, in the off-freq condition, the nontarget sounds were the 10th–14th harmonics of a 250-Hz fundamental, i.e., having the same fundamental frequency as the target sounds but higher harmonic numbers.

The protocols for experiments 2A and 2B were as follows. In experiment 2A, three listeners ran the LT conditions and then the HT conditions; for four listeners, the order was the opposite. In experiment 2B, three listeners ran the LT conditions and then the HT conditions; for the other three listeners, the order was the opposite. In other respects, the procedures were parallel to those described for experiment 1. Listeners' practice thresholds were sufficiently low that none were excluded from completing the experiment. On one occasion in experiment 2A and two occasions in experiment 2B, practice effects were observed in the experimental condition and repeated as described above.

The low and lower conditions of experinents 1 and 2 share the same fundamental frequencies (f_o = 198 and 176 Hz, respectively). In experiment 1, the spectral extent (cutoff frequency) is the same for the target and nontargets sounds in the low and lower conditions. In contrast, in the low and lower conditions of experiment 2, the nontarget sounds have narrower bandwidths than the targets (by 780 and 1110 Hz, respectively). Comparing the same, low, and lower conditions across experiments provides an opportunity to evaluate potential cues associated with differences in frequencies per se (experiment 1, the spectral extent is similar, but the harmonic components are not) versus changes in spectral regions associated at higher frequencies (experiment 2, same range of harmonic numbers and different bandwidths).