Audiovisual perception of voicing with age in quiet and cafe noise

Abstract

Research has shown that voicing is difficult to discern in noisy environments. While voicing may be difficult to resolve from visual cues, acoustic cues for voicing are relatively robust. This study addresses these factors with normally aging audiovisual perception. Identification responses were gathered with 19–30-year-old and 49–60-year-old adults for audiovisual (AV) CVs differing in voicing and consonant place of articulation. Materials were presented in quiet and in cafe noise (SNR=0 dB) as audio-only (A), visual-only (V), congruent AV, and incongruent AV. Results show a tendency toward use of visual information with age and noise for consonant place of articulation. Notably for voicing, incongruent AV materials that had one voiced component, regardless if it was A or V that was voiced, were consistently perceived as voiced in both age groups and regardless of noise. Only if the A and V components were both voiceless was the syllable perceived as voiceless. These findings indicate the influence of age and noise in the use of perceptual information to identify place of articulation. That voicing is robustly salient from either audio or visual information, despite the unlikely presence of strong visual cues for voicing, indicates a possible bias toward the perception of voicing.

Audio-visual identification of place of articulation and voicing in white and babble noise

“…Different phonetic attributes, in this respect POA and voicing, are associated with different vulnerabilities to noise. A prior study by Behne et al ͑2006͒ showed that for AV stimuli differing in POA, babble noise at a SNR of 0 dB led to an increase in the use of the visual modality, whereas no such shift was evident for AV stimuli that vary in voicing at this noise level. These results are supported by a firm body of research ͑e.g., Jiang et al, 2006;Miller and Nicely, 1955͒.…”

“…Surprisingly, voiced consonants resulted in more A-match responses ͑M = 45%, SE = 2.8͒ than did voiceless consonants ͑M = 23%, SE = 2.65͒. This result contrasts with previous findings ͑e.g., Behne et al, 2006;McGurk and MacDonald, 1976͒. The most remarkable observation in this respect is the size of the difference in mean A-match responses between voiced and voiceless consonants, and the voicing effect's consistency across conditions.…”

“…The current study used AV recordings that were made for a previous study ͑see Behne et al, 2006;Behne et al, 2007͒. The AV recordings of a male speaker were made in the Speech Laboratory at the Department of Psychology, NTNU.…”

“…Research has shown that the visual modality plays an important role in perception of speech ͑e.g., Green and Kuhl, 1989;Massaro, 1987;McGurk and MacDonald, 1976͒ and its contribution is emphasized in conditions where the speech signal is degraded by noise ͑e.g., Erber, 1969;MacLeod and Summerfield, 1987;Sumby and Pollack, 1954͒. Previous research on AVSP has focused on the effect of white noise ͑e.g., Dodd, 1977;Fixmer and Hawkins, 1998͒, an auditory distractor infrequently present in everyday human speech communication, while the effect of babble noise is much less explored ͑e.g., Behne et al, 2006;Markides, 1989;Wang et al, 2008͒. The phonetic attributes place of articulation ͑POA͒ and voicing differ in susceptibility to noise ͑e.g., Miller and Nicely, 1955͒. However, whereas the effect of noise level on these attributes is well established ͑e.g., Jiang et al, 2006;Miller and Nicely, 1955͒, the effect of different types of noise is not thoroughly explored.…”

“…POA's lack of auditory robustness is greatly compensated for by salient visual cues; that is, seeing the face of the speaker greatly aids perceivers in identifying POA in noise ͑e.g., Binnie et al, 1974͒. This visual benefit is not found for voicing identification ͑e.g., Behne et al, 2006;Binnie et al, 1974͒. Behne et al ͑2006͒ found that responses to incongru-ent AV syllables that varied in terms of voicing always matched the auditory component, independent of which component of the stimulus was voiced and independent of the presence of noise.…”