Measurement variability in vowel formant estimation: A simulation experiment

This paper considers sources of error in common vowel formant extraction techniques, investigating the extent to which the temporal point of measurement and software settings influence the formant values obtained. To do this, we report on the results of a vowel measurement simulation where, rather than extracting a single measurement for each vowel token, thousands of measurements are taken for each vowel with varied settings in jittered measurement locations (seeded by measurements from human analysts). Such an undertaking, we argue, yields important insight into the bounds of measurement error in vowel analysis.