Formants: Extract smoothest part (constrained)...

Extracts the best matching part from the slected Formants.

Settings

Minimum F1 (Hz)
suppresses models whose average first formant frequency, f1, is below minimumF1 by a factor sqrt (minimumF1 - f1 + 1). You can use this constraint to disfavour models with a low average first formant. Sometimes due to the generally high frequency of the /a/ a lower harmonic of the fundamental frequency is taken as a candidate for the first formant. You can suppress these models with this constraint.
Maximum F1 (Hz)
suppresses models whose average first formant frequency, f1, is above maximumF1 by a factor sqrt (f1 - maximumF1 + 1).You can use this constraint to disfavour models in which the first formant is missing.
Minimum F2 (Hz)
suppresses models whose average second formant frequency, f2, is below minimumF2 by a factor sqrt (minimumF2 - f2 + 1). This constraint might be used for high front vowels that normally have a large distance between the first and second formant.
Maximum F2 (Hz)
suppresses models whose average second formant frequency, f2, is above maximumF2 by a factor sqrt (f2 - maximumF2 + 1). This factor is sometimes necessary to suppress models for high back vowels where the second formant is "missing", i.e. where the third formant is playing the role of the second.
Minimum F3 (Hz)
suppress models whose average third formant frequency, f3, is below minimumF3 by a factor sqrt (minimumF3 - f3 + 1). This constraint might sometimes be usefull to suppress low lying third formants that can occur for /a/-like vowels.

© djmw, April 24, 2014