F0 Processing and the Seperation of Competing Speech Signals by Listeners With Normal Hearing and With Hearing Loss

Abstract

Normal-hearing and hearing-impaired listeners were tested to determine F0 difference limens for synthetic tokens of 5 steady-state vowels. The same stimuli were then used in a concurrent-vowel labeling task with the F0 difference between concurrent vowels ranging between 0 and 4 semitones. Finally, speech recognition was tested for synthetic sentences in the presence of a competing synthetic voice with the same, a higher, or a lower F0. Normal-hearing listeners and hearing-impaired listeners with small F0-discrimination (deltaF0) thresholds showed improvements in vowel labeling when there were differences in F0 between vowels on the concurrent-vowel task. Impaired listeners with high deltaF0 thresholds did not benefit from F0 differences between vowels. At the group level, normal-hearing listeners benefited more than hearing-impaired listeners from F0 differences between competing signals on both the concurrent-vowel and sentence tasks. However, for individual listeners, deltaF0 thresholds and improvements in concurrent-vowel labeling based on F0 differences were only weakly associated with F0-based improvements in performance on the sentence task. For both the concurrent-vowel and sentence tasks, there was evidence that the ability to benefit from F0 differences between competing signals decreases with age.

Computational Model Predictions of Cues for Concurrent Vowel Identification

“…Previous studies of concurrent vowel identification focused mainly on the contribution of F0 difference and formant difference cues at one vowel level (e.g., Summers and Leek, 1998;Chintanpalli and Heinz, 2013). The current study examined the role of F0 and formant frequency difference cues and vowel level (25-85 dB SPL) on the identification of concurrent vowels.…”

“…Vowel duration was 400 ms, including 15-ms raised-cosine rise and fall ramps. Table 1 includes the formant frequencies and bandwidths for each vowel, which were the same as those used in earlier studies of concurrent vowel identification (e.g., Assmann and Summerfield 1994;Summers and Leek 1998;Chintanpalli and Heinz 2013). Figure 1 shows the envelope spectrum for each vowel, computed using linear predictive coding.…”

“…The perceptual and modeling framework used in the current study can also be extended to investigate why older adults with normal and impaired hearing have reduced concurrent vowel identification across F0 differences and reduced F0 benefit (e.g., Vongpaisal and Pichora-Fuller 2007;Arehart et al 1997;Summers and Leek 1998). Physiological and anatomical changes in the cochlea and the decline in number of AN fibers due to increased age and hearing loss (e.g., Makary et al 2011;Schmiedt et al 1996) could affect phase locking of AN fibers to formants and F0s of concurrent vowels.…”

“…Differences in fundamental frequency (F0) among talkers are one important cue for talker segregation. For example, when listening to two vowels simultaneously, identification of both vowels improves as the F0 difference between the two vowels increases (e.g., Scheffers 1983;Assmann and Summerfield 1990;Arehart et al 1997;Summers and Leek 1998;Vongpaisal and Pichora-Fuller 2007;Chintanpalli and Heinz 2013). Although vowel segregation clearly benefits from F0 differences, the ability to segregate and identify two simultaneous vowels with identical F0s suggests that listeners use multiple cues for identification, including differences in formants between vowels (see Micheyl and Oxenham 2010 for review).…”

Effect of fundamental-frequency and sentence-onset differences on speech-identification performance of young and older adults in a competing-talker background

“…A review of the literature on the effects of DF0 can be summarized briefly as follows. First, DF0 significantly benefited identification performance whether the competing speech signals are steady-state synthesized vowel pairs Arehart et alet al, 1997Arehart et alet al, , 2005Assmann andSummerfield, 1990, 1994;Chalikia and Bregman, 1989;Darwin, 1993, 1994;de Cheveign e, 1997;Meddis and Hewitt, 1992;Rossi-Katz and Arehart, 2005;Stubbs and Summerfield, 1988;Assmann, 1989, 1991;Summers and Leek, 1998;Vongpaisal and PichoraFuller, 2007), nonsense syllables (Vestergaard et al, 2009), sentence pairs without natural F0 variation (Assmann, 1999;Bird and Darwin, 1998;Brokx and Nooteboom, 1982), or with natural F0 variation preserved (Assmann, 1999;Darwin et al, 2003;Oxenham and Simonson, 2009;Summers and Leek, 1998). However, the pattern of improvement was observed to be gradual over a greater range of DF0s [up to DF0 of 8 or 9 semitones (STs)] for sentence-identification whereas the vowel-identification performance at DF0 of !2 ST reached asymptote apparently due to the pattern of beating between double vowels (Culling and Darwin, 1994).…”