Santiago Barreda
Assistant Professor, Department of Linguistics, UC Davis

Main Research Vignettes CV phonTools Statistics Links

 

What is the relationship between the average pitch (f0) and formant frequencies produced by different speakers?


The information presented here was collected from two large vowel datasets, Peterson and Barney (1952) and Hillenbrand et al. (1995). This data comprises productions from 215 unique speakers, 76 adult females, 78 adult males, and 61 children of either sex.


Adult males (blue) tend to produce vowels with lower formant frequencies (FFs) and lower f0s than adult females (red) and children of either sex (green). Horizontal lines indicate group means.


Each point in the figure below represents a single speaker (239 in total). As seen in the figure below, the average f0 and average FF produced by a speaker are positively correlated.



A model predicting mean FFs from mean f0, fit to the data from the aforementioned 239 speakers indicates that:

1) Mean f0 explains about 70% of the variance in mean FFs (and vice versa).

2) Mean f0 can predict mean FFs to within an expected error of about 6%.

3) A difference between two speakers of one octave of in mean f0 is expected to be associated with a difference of about 25% in mean FFs.