ABSTRACT
Stimuli Pulse trains with a 3 dB/octave rolloff were initially generated at fundamental frequencies of 131 and 262 Hz. Four spectral weighting functions were used (Figure 1). The upper cutoff frequency was set at 0.7 kHz for Spectra A and C, and 2.1 kHz for Spectra B and D. The lower cutoff was set at 0.4 kHz for Spectrum C and 1.2 kHz for Spectrum D. Stimuli were sampled at 10 kHz using 12-bit D2A converters, and low-pass filtered at 4.2 kHz using two cascaded Barr & Stroud EF3 -02 filters (each 48 dB/octave slope). All notes had 20 ms onset and offsets shaped by a raised cosine function. Stimuli were presented binaurally at 68 dB SPL through TDH-39 headphones in sound-attenuated IAC booths. Loudness cues were further controlled by randomizing stimuli intensity levels within a 5 dB range on a trial by trial basis. In addition, a random pitch variation within 5 percent of the centre frequency was introduced to reduce memory effects. This resulted in a roving pitch range of a semitone.
