Samples of waveforms of the harmonic (H) and inharmonic (I) stimuli after passing through a simulated auditory filter centered at 1000 Hz. The value of F0 was 100 Hz and the bandpass filter was centered at 1100 Hz . In the two top panels, the samples are from H tones. The lower-left and lower-right panels show samples from I tones with and 25 Hz, respectively. The times between peaks in the TFS close to adjacent envelope maxima are indicated. Note that the exact envelope shape and modulation depth depend on the random selection of starting phases, and vary from one tone to the next.
Results of experiment 1, showing the value of the detectability index required to discriminate a frequency shift, , equal to 0.5F0, plotted on a square-root axis, as a function of the harmonic number, , on which the bandpass filter was centered. Each panel shows results for one F0. Open symbols indicate values significantly above zero, while filled symbols indicate values not significantly above zero. Error bars indicate standard deviations across subjects.
Results of experiment 2, showing F0 DLs as a function of the harmonic number, , on which the bandpass filter was centered. Each panel shows results for one F0. Open and filled symbols show F0 DLs for cosine- and random-phase stimuli, respectively. Error bars indicate standard deviations across subjects.
Comparison of results for experiments 1 and 2. All scores were converted to Weber fractions for time-interval discrimination (see text for details). Weber fractions when only TFS cues were available (experiment 1) are shown by open squares. Weber fractions when both TFS and envelope cues were available (experiment 2) are shown by open circles (random-phase stimuli) and filled circles (cosine-phase stimuli). Weber fractions are plotted as a function of the harmonic number, , on which the bandpass filter was centered. Each panel shows results for one F0. Up-pointing arrows in the panel for indicate cases where performance was too poor for the Weber fraction to be determined.
Subject identifiers (1, 2, 3, and 4) of subjects who were unable to complete the adaptive procedure for each F0 and each .
Maximum differences in excitation level in decibels between the excitation patterns evoked by the H and I tones when the value of for the I tone was equal to the mean value required for threshold or to 0.5F0 (see text for details). The effect of the TEN background was included in the calculations.
Article metrics loading...
Full text loading...