Recall disruption produced by noise-vocoded speech: a study of the irrelevant sound effect

Abstract

The Irrelevant Sound Effect (ISE) is the finding that serial recall performance is impaired under complex auditory backgrounds such as speech as compared to white noise or silence (Colle & Welsh, 1976). Much of the current research investigates the role of changing-state complexity of the background stimuli in ISE (e.g., Jones & Macken, 1993). This study investigated whether speech-specific qualities of the irrelevant background have an effect on the ISE. This was done using noise-vocoded speech, an acoustic transformation that removes many of the acoustic properties of speech while preserving the speech intensity profile. Experiment 1 compared serial recall accuracy resulting from white noise and noise-vocoded speech backgrounds and found that noisevocoded speech is more disruptive. Noise-vocoded speech preserves the intensity profile of nature speech with a number of amplitude channels; each channel matches the average intensity for the corresponding channel in natural speech. Experiment 2 systematically varied the resolution of noise-vocoded speech by adjusting the number of these channels. These results show that ISE varies based on the number of channels in noise-vocoded speech, but this change in disruption is not consistent across channel conditions. Results demonstrate that changing state complexity alone is not a sufficient explanation of ISE.