Temporal stability of long-term measures of fundamental frequency

We investigated long-term mean, median, and base value of the voice fundamental frequency (F0) to estimate how long it takes their variability to stabilize. That information can be useful in the development of F0 contour normalization and forensic applications. Change point analysis was used to locate changes in underlying variance in the mean, median, and base value time-series. In one experiment, stabilization points were calculated in recordings of the same text spoken in 26 languages. Average stabilization points are 5 s for base value and 10 s for mean and median. Variance after the stabilization point was reduced around 40 times for mean and median and more than 100 times for the base value. In another experiment, four speakers read two different texts each. Stabilization points for the same speaker across the texts do not exactly coincide as would be ideally expected. Average point dislocation is 2.5 s for the base value, 3.4 for the median, and 9.5 for the mean. After stabilization, differences in the three measures obtained from the two texts are 2% on average across speakers. Present results show that stabilization points in long-term measures of F0 occur earlier than suggested in the previous literature.