Schematized formant tracks that depict measurements taken during the vowel, from the consonant-vowel boundary (dotted line) through the midvowel, and how these were used for analyses.
A comparison of F3 transition data for wadrav (likely to be retroflexed) and wagrav (likely to be bunched) as a function of experience with the cochlear implant for participant ML.
Schematized comparison of a small variability index (lower panels) with a larger variability index (upper panels) due to greater /r/ dispersion in the upper right-hand plot. The /r/ dispersion is plotted in the domain of the difference at the beginning of the vowel by the F3 transition from vowel beginning to midvowel. The vowel dispersion is plotted in the F1, F2 domain.
Proportion correct from 11-item consonant confusion matrices constructed from results of consonant identification (12 repetitions per consonant) at a comfortable listening level.
Results of discriminant analyses of speakers’ productions of /b/, /d/, and /g/.
A comparison of singleton stops (open symbols) vs tokens (closed, filled symbols) for a normal-hearing speaker with auditory feedback (left panel) and a speaker with hearing loss in the preimplant condition (right panel).
Extent of /r/ contrast measured as the difference in the F3 transition for singleton stop vs productions with the indicated stops. Results are shown for implant users in the one-year session (filled circles) and for normal-hearing speakers with their hearing masked (clear triangles) or not masked (filled triangles).
Participant demographic information, vowel and consonant percent correct identification for implant users, and normal-hearing controls. (Asterisks denote missing data.)
Effect sizes from the three ANOVAs.
Article metrics loading...
Full text loading...