A comparison of three speaker-intrinsic vowel formant frequency normalization algorithms for sociophonetics

This paper evaluates a speaker-intrinsic vowel formant frequency normalization algorithm initially proposed in Watt & Fabricius (2002). We compare how well this routine, known as the S-centroid procedure, performs as a sociophonetic research tool in three ways: reducing variance in area ratios of vowel spaces (thus attempting to equalize vowel space areas); improving overlap/co-extensiveness of vowel polygons; and reproducing two-dimensional vowel configurational relationships within the vowel space, compared to formant data in raw Hertz. The study uses existing datasets of vowel formant data from RP and Aberdeen English (northeast Scotland). We conclude that, for the data examined here, the S-centroid W&F procedures performs at least as well as the two most recognized speaker-intrinsic, vowel-extrinsic, formant-intrinsic normalization methods, Lobanov's (1971) z-score procedure and Nearey's (1977/8) individual log-mean procedure (CLIHi4 in Adank 2003, CLIHi2 as tested here), and in some test cases better than Nearey1.