ABSTRACT
Articulation theory, referred to as the articulation index (AI) model, provides a quantitative framework for characterizing the deleterious effects of background noise and/or filtering on speech reception. Articulation theory is a model designed to predict speech reception over band-limited and/or noisy channels, conditions not unlike the signal degradation caused by sensorineural hearing loss. Although developed for listeners with normal hearing, J. Fletcher suggested modifications to accommodate sensorineural hearing loss. Several studies have reported that AIs tend to overestimate speech reception scores for hearing-impaired subjects when compared to scores of normal-hearing subjects. Monotonicity has been observed for individual hearing-impaired subjects under a variety of stimulus conditions, including quiet backgrounds; backgrounds of band-limited and wide-band noise; and for high- and low-pass filtered speech. The intention of the speech-in-noise experiment was to establish a normative AI-intelligibility function for the test materials for comparison with data collected from hearing-impaired subjects.
