An evaluation of the role of envelope-based cues for the detection of a tone added to noise.

Abstract

Using 2IFC paradigm, d’ scores were obtained for two conditions derived from the tone-in-noise task. In one condition, observers detected a tone added to a band of noise, but the waveforms presented in the two intervals were of equal energy. Noise center frequencies and signal frequencies of 600, 1800, and 5400 Hz and noise bandwidths of 40, 120, and 360 Hz were used. In the second condition, the envelopes of the noise-alone and tone-plus-noise waveforms were extracted, and the envelopes were used to modulate sinusoids of 600, 1800, and 5400 Hz. The observers’ ability to discriminate between modulation patterns was measured. For narrow bands of noise, d’ values in the two conditions were approximately equal, indicating that for level-equalized waveforms the detectability of a tone added to noise could be accounted for by differences in the envelope patterns. For bandwidths that were larger than or equal to 20% of the center frequency, d’ scores obtained using amplitude modulated tones were superior to those obtained using the level-equalized stimuli, an advantage that may reflect differences in the power spectra of the modulated tones. [Supported by the NIH.]