Effect of masking noise on cognitive performance and annoyance in open plan offices.

Even for low noise levels, employees working in open-plan offices declare they are disturbed by different sound sources and especially by intelligible speech. This paper presents two experiments that aim at studying the effects of sound sources on task performance and cognitive load. In the first experiment, thirty-five individuals were assigned a serial memory task in four noise conditions: printers, intelligible speech, non-intelligible speech and phone rings. Noise annoyance was assessed by measuring task performance and cognitive workload (NASA-TLX). Although there was no significant difference between intelligible speech, unintelligible speech and phone ring tone signals in terms of annoyance, the performance in a memory task decreased in the presence of intelligible speech. In the second experiment, the relevance of a new intelligibility index was evaluated. This index was derived from eSII as defined by Rhebergen et al. (2006) to deal with fluctuating noise as speech babble. Fifty-five subjects performed a serial memory task in four STIt (time varying Speech Transmission Index calculated for every 12.5 ms in dynamic environment) conditions (from 0.38 to 0.69) while STI values were kept constant (around 0.36). No significant difference appeared between the decreases in performance observed with multiple levels of intelligibility. However, a significant difference between two groups of individuals appeared in performance measurements. The group of better performers seemed to be less affected by the high level of intelligibility condition than the other ones.