S. Amagai et al., Detection of modulation in spectral envelopes and linear-rippled noises bybudgerigars (Melopsittacus undulatus), J ACOUST SO, 105(3), 1999, pp. 2029-2035
Budgerigars were trained to discriminate complex sounds with two different
types of spectral profiles from flat-spectrum, wideband noise. In one case,
complex sounds with a sinusoidal ripple in (log) amplitude across (log) fr
equency bandwidth were generated by combining 201 logarithmically spaced to
nes covering the frequency region from 500 Hz to 10 kHz. A second type of r
ippled stimulus was generated by delaying broadband noise and adding it to
the original noise in an iterative fashion. In each case, thresholds for mo
dulation depth (i.e., peak-to-valley in dB) were measured at several differ
ent ripple frequencies (i.e., cycles/octave for logarithmic profiles) or di
fferent repetition pitches (i.e., delay for ripple noises). Budgerigars wer
e similar to humans in detecting ripple at low spatial frequencies, but wer
e considerably more sensitive than humans in detecting ripples in log rippl
e spectra at high spatial frequencies. Budgerigars were also similar to hum
ans in detecting linear ripple in broadband noise over a wide range of repe
tition pitches. Taken together, these data show that the avian auditory sys
tem is at least as good, if not better: than the human auditory system at d
etecting spectral ripples in noise despite gross anatomical differences in
both the peripheral and central auditory nervous systems. (C) 1999 Acoustic
al Society of America. [S0001-4966(99)03802-3].