Consonant confusions in white noise

Abstract

The classic [MN55] confusion matrix experiment (16 consonants, white noise masker) was repeated by using computerized procedures, similar to those of Phatak and Allen (2007). ["Consonant and vowel confusions in speech-weighted noise," J. Acoust. Soc. Am. 121, 2312-2316]. The consonant scores in white noise can be categorized in three sets: low-error set [/m/, /n/], average-error set [/p/, /t/, /k/, /s/, /[please see text]/, /d/, /g/, /z/, /Z/], and high-error set /f/theta/b/, /v/, /E/,/theta/]. The consonant confusions match those from MN55, except for the highly asymmetric voicing confusions of fricatives, biased in favor of voiced consonants. Masking noise cannot only reduce the recognition of a consonant, but also perceptually morph it into another consonant. There is a significant and systematic variability in the scores and confusion patterns of different utterances of the same consonant, which can be characterized as (a) confusion heterogeneity, where the competitors in the confusion groups of a consonant vary, and (b) threshold variability, where confusion threshold [i.e., signal-to-noise ratio (SNR) and score at which the confusion group is formed] varies. The average consonant error and errors for most of the individual consonants and consonant sets can be approximated as exponential functions of the articulation index (AI). An AI that is based on the peak-to-rms ratios of speech can explain the SNR differences across experiments.

Sources of Variability in Consonant Perception and Implications for Speech Perception Modeling

“…3 indicated that any considered difference in the stimuli produced a measurable effect. The observed perceptual variability across talkers is well established in the related literature (e.g., Phatak et al 2008); the equally large variability within talkers had not yet been demonstrated. Most remarkably, even a 100-ms time shift in the white maskingnoise waveform induced significant perceptual variability, indicating that "steadystate" masking noise should not be considered steady over time in the context of consonant cues.…”

“…Phatak and Allen (2007) measured consonant perception in speech-weighted noise and demonstrated noise-type induced perceptual differences to the Miller and Nicely (1955) data. In following studies, perceptual differences across different speech tokens of the same phonetic identity came more into focus (e.g., Phatak et al 2008).…”

Consonant recognition loss in hearing impaired listeners

“…However, NH listeners show the same confusion group in the presence of a white noise masker ͑Miller and Nicely, 1955; Phatak et al, 2008͒. This is because white noise masks higher frequencies more than speech-weighted noise at a given SNR.…”

“…It has been shown that NH listeners demonstrate a wide range of performance across different consonants presented in noise ͑Phatak and Phatak et al, 2008͒, and it is likely that HI listeners would also exhibit a variance in performance across the same consonants. A quantitative comparison of recognition performances for individual consonants is necessary to determine whether the loss of performance for HI listeners is consonant-dependent.…”

Perceptual effects of plosive feature modification

“…All the sounds were from the LDC2005S22 corpus ("Articulation Index Corpus" provided by the Linguistic Data Consortium, University of Pennsylvania). Data from Phatak et al (2008) verified that these utterances had 0% recognition error at and above þ12 dB SNR. A total of 14 different talkers were used, two of whom had spent some part of their childhood outside the US while others had early training in a language other than English (Fousek et al, 2004).…”