A Note on the Range of Estimates of F1 and F2 in German, English, and Chinese
D. N. Sherman · The Journal of the Acoustical Society of America · 1969
Estimates of the mean (m1) and second moment divided by the mean (m2 m1) of speech-power spectra for 18 bilingual speakers (German and English or Chinese and English) over averaging time of 100 and 140 sec have been obtained. Tabulated results for T = 140 sec support the conjectures of others regarding the utility of m1 and m2 m1 for the extraction of formant frequencies. There is evidence that these statistics, even for averaging times of 140 sec, are similar to those obtained by examining individual phonemes and that they may, in fact, provide information about the average position of the first and second formant frequencies.