An acoustic analysis of frontal misarticulation of /s/ in children

The Development of Voiceless Sibilant Fricatives in Putonghua-Speaking Children

“…People perceive speech in a categorical fashion, experiencing perceptual equivalence for variation within a phoneme's boundary (categorical perception;Liberman, Harris, Kinney, & Lane, 1961). When children produce two sounds in a range corresponding to a single adult sound category, transcribers tend to overlook the subtle distinction, a phenomenon termed covert contrast, which was supported by a growing body of literature that uses instrumental analysis (Baum & McNutt, 1990;F. Li, Edwards, & Beckman, 2009;Macken & Barton, 1980;Scobbie et al, 2000).…”

Acoustic Characteristics of /s/ in Adolescents

“…The latter position was taken in the current study. (1989) adults (F) 20-21 yrs in isolation Spectrograph before M = 6.8-6.9 kHz children (NR) 3 yrs (DFT) offset F = 7.4-7.8 kHz children (NR) 4 yrs 3 = 6.9-7.2 kHz children (NR) 5 yrs 4 = 6.8-7.1 kHz children (NR) 7 yrs 5 = 6.8-7.2 kHz 7 = 6.7-6.8 kHz Baum & children (NR) 6 yrs (s)V(t) syllables BLISS speech 100 ms 20 kHz 0-8.5 kHz Centroids: McNutt (1990) children (NR) 8 yrs in isolation analysis system after onset 6 = 5.9-6.4 kHz (DFT) 8 = 6.0-6.7 kHz Katz et al (1991) 10 adults (5/5) g 26-45 yrs (s)V syllables CSpeech 100 ms 20 kHz 0-10 kHz Centroids: children (5/5) 3 yrs (phrase final) (DFT, moments) before A = 6.1-6.8 kHz children (5/5) 5 yrs offset 3 = 6.2-6.6 kHz children (5/5) 8 yrs 5 = 6.3-6.6 kHz 8 = 6.1-6.4 kHz…”

“…Strevens did note that his failure to analyze above 8 kHz was a function of instrumental limitations. In addition to limiting the analysis range, at least nine of the studies made use of low-pass filtering, probably in response to concerns about aliasing (Bauer & Kent, 1987;Baum & McNutt, 1990;Behrens & Blumstein, 1988;Harmes et al, 1984;Hoole et al, 1989;Miccio et al, 1996;Nittrouer et al, 1989;Sereno et al, 1987;Tjaden & Turner, 1997). Such filtering is evident in those studies in Table 1 in which the upper limit to the analysis range is less than one half of the specified sampling rate.…”

“…Harmes et al (1984) derived their centroid values from FFT data weighted to include 40% of the spectral energy surrounding the amplitude peak. Most centroid calculations used FFT data and formulae presented by Forrest et al (1988;Avery & Liss, 1996;Baum & McNutt, 1990;Nittrouer, 1995;Tjaden & Turner, 1997). Nittrouer et al (1989) and Katz et al (1991) also used the Forrest et al formulae to derive their centroids, but based their analyses on DFT data.…”

“…Because the present concern was with the independent characterization of /s/, studies that focused primarily on the coarticulation of /s/ with surrounding segments (e.g., McGowan & Nittrouer, 1988;Soli, 1981) were not included in Table 1 unless they also reported independent data on /s/. Additionally, some studies reported data for both typical and disordered speakers (Baum & McNutt, 1990;Daniloff et al, 1980;Harmes et al, 1984;Miccio et al, 1996;Tjaden & Turner, 1997), in which case only data based on the typical (i.e., normal) speakers were included. Examination of Table 1 indicates a range of methodological differences across the studies.…”