On the number of auditory filter outputs needed to understand speech: Further evidence for auditory channel independence

Abstract

The number of auditory filter outputs required to identify phonemes was estimated in two experiments. Stimuli were divided into 30 contiguous equivalent rectangular bandwidths (ERBN) spanning 80 to 7563 Hz. Normal-hearing listeners were presented with limited numbers of bands having frequency locations determined randomly from trial to trial to provide a general view, i.e., irrespective of specific band location, of the number of 1-ERBN-wide speech bands needed to identify phonemes. The first experiment demonstrated that 20 such bands are required to accurately identify vowels, and 16 are required to identify consonants. In the second experiment, speech-shaped noise or time-reversed speech was introduced to the non-speech bands at various signal-to-noise ratios. Considerably elevated noise levels were necessary to substantially affect phoneme recognition, confirming a high degree of channel independence in the auditory system. The independence observed between auditory filter outputs supports current views of speech recognition in noise in which listeners extract and combine pieces of information randomly distributed both in time and frequency. These findings also suggest that the ability to partition incoming sounds into a large number of narrow bands, an ability often lost in cases of hearing impairment or cochlear implantation, is critical for speech recognition in noise.

Use of a compound approach to derive auditory-filter-wide frequency-importance functions for vowels and consonants

“…Evidence has accumulated to suggest that the normal auditory system takes advantage of momentary improvements in local signal-to-noise ratio (SNR) to extract speech from noise (Brungart et al, 2006;Li and Loizou, 2007;Apoux and Healy, 2009). More specifically, it is believed that the auditory system processes primarily the output of the auditory filters that contain a relatively undistorted view of local target signal properties (i.e., a favorable SNR) to create an interpretable representation of the signal of interest.…”

“…A first factor is the overall number of available auditory-filter outputs containing a relatively undistorted view of local target signal properties. This quantity was recently assessed in a study by Apoux and Healy (2009). To provide an estimate of the number of auditory-filter outputs necessary to understand speech, the authors developed a technique in which the stimuli are divided into 30 contiguous equivalent rectangular bandwidths (ERB N ; Glasberg and Moore, 1990) spanning 80-7563 Hz.…”

“…The listeners are then presented with limited numbers of bands having frequency locations determined randomly from trial to trial. Apoux and Healy (2009) showed that phoneme intelligibility increases with increasing number of 1-ERB N wide bands and that NH listeners require 20 bands to accurately identify vowels and 16 to identify consonants.…”

“…As a consequence, the intelligibility that can be achieved with a fixed number of auditory-filter outputs may vary with the center frequency of these auditory filters. While they acknowledged the likely influence of band location, Apoux and Healy (2009) did not investigate this factor. Instead, the authors randomized the spectral location of the bands to provide a general view, i.e., irrespective of specific band location, of the number of 1-ERB N -wide speech bands needed to identify vowels and consonants.…”

“…The primary purpose of the present study was to assess the importance or weight of each auditory-filter output as operationally defined in Apoux and Healy (2009). However, it was not possible to use their technique because it does not appear well suited for such assessment.…”