A. Kohlrausch et al., The influence of carrier level and frequency on modulation and beat-detection thresholds for sinusoidal carriers, J ACOUST SO, 108(2), 2000, pp. 723-734
This paper is concerned with modulation and beat detection for sinusoidal c
arriers. In the first experiment, temporal modulation transfer functions (T
MTFs) were measured for carrier frequencies between 1 and 10 kHz. Modulatio
n rates covered the range from 10 Hz to about the rate equaling the critica
l bandwidth at the carrier frequency. In experiment 2, TMTFs for three carr
ier frequencies were obtained as a function of the carrier level. In the fi
nal experiment, thresholds for the detection of either the lower or the upp
er modulation sideband (beat detection) were measured for ''carrier" freque
ncies of 5 and 10 kWz, using the same range of modulation rates as in exper
iment 1. The TMTFs for carrier frequencies of 2 kHz and higher remained fla
t up to a modulation rate of about 100-130 Hz and had similar values across
carrier frequencies. For higher rates, modulation thresholds initially inc
reased and then decreased rapidly, reflecting the subjects' ability to reso
lve the sidebands spectrally. Detection thresholds generally improved with
increasing carrier level, but large variations in the exact level dependenc
e were observed, across subjects as well as across carrier frequencies. For
beat rates up to about 70 Hz (at 5 kHz) and 100 Hz (at 10 kHz), beat detec
tion thresholds were the same for the upper and the lower sidebands and wer
e about 6 dB higher than the level per sideband at the modulation-detection
threshold. At higher rates the threshold for both sidebands increased, but
the increase was larger for the lower sideband. This reflects an asymmetry
in masking with more masking towards lower frequencies. Only at rates well
beyond the maximum of the TMTF did detection for the lower sideband start
to be better than that for the upper sideband. The asymmetry at intermediat
e frequency separations can be explained by assuming that detection always
takes place in filters centered above the stimulus spectrum. The shape of t
he TMTF: and the beat-detection data reflects a limitation in resolving fas
t amplitude variations, which must occur central to the inner-ear filtering
. Its characteristic resembles that of a first-order low-pass filter with a
cutoff frequency of about 150 Hz. (C) 2000 Acoustical Society of America.
[S0001-4966(00)03407-X].