Speech Perception under Adverse Conditions: The Processing of Social Information

Introduction

Social information inferred about the talker (such as age, gender or native language) can affect how the listener perceives phonetic information. For example, the same ambiguous sound artificially inserted into a male and a female voice is interpreted as a different sound in the male voice than it is in the female. However, the extent of this effect and the perceptual processes that enable it are not completely understood.

Additionally, research examining this effect has primarily been conducted under "ideal" laboratory listening conditions which, however, do not greatly resemble natural listening conditions. Methodologies that simulate more realistic listening conditions include the presence of background noise (e.g. "cocktail party" sounds or white noise) or conducting the listening task while multitasking. These are often referred to “adverse conditions.”

Objective

The aim of this project is to examine if adverse listening conditions increase listeners’ reliance on cues provided by social information and thereby decrease reliance on lower level phonetic cues. Moreover, it will examine how cognitive load and noise differ in their effect on the processing of both social (high-level) and phonetic (low-level) cues.

Further experiments may investigate whether social information is processed differently in a second language or by bilinguals.

Method

Perception experiments will be conducted in which participants are presented with artificial continua and asked to classify ambiguous sounds (two-alternative forced choice tasks). These will be presented under diverse adverse listening conditions, such as with a simultaneous cognitive load task, with background noise such as multi-talker babble or white noise. Future studies will employ eye-tracking methodologies.

Application

The results of this project will aid in the understanding of how speech is processed in adverse (and therefore more realistic) listening conditions and will add to the body of psycho- and sociolinguistic research examining how social cues affect speech processing. Furthermore, the direct comparison of the effect of noise (energetic masking) and cognitive load (informational masking) on speech perception assist other psycholinguistic researchers in their benefit choice of methodology.