I'm writing my Ph.D. thesis on the sociophonetics of Modern Greek. My
research deals with the variation of prenasalization phenomena,
which occur before voiced-stop environments, e.g.
['kabos] ~ ['kambos] (field)
['adras] ~ ['andras] (man)
Phonetically prenasalization is not categorical but has gradations.
Fujimura (1962:1874) refers to the spectral properties of nasal murmur
and accepts three important features: a) the existence of a very low first
formant around 300 Hz, which is well-separated from the upper formant
structure, b) relatively high-dumping factors of the formants and
c) high density of the formants in the frequency domain.
Lieberman (1988:226), also states that a nasal murmur which is located
around 250 Hz is sufficient for the perception of a nasal consonant,
irrespective of place of articulation.
By examining the spectrograms of prenasalized voiced-stops, I found out
that there is a prominence in spectral energy around 2800 Hz. I find very
difficult to detect and differentiate frequencies from 200 to 300 Hz
between the preceding vowel and the nasal.
I'm searching a way of quantifying the prenasalization energy, so that
I can compare different levels of prenasalization extracted from my data;
a kind of measuring the amount of prenasalization on each word.
I would be very grateful for any suggestions and advices on this matter.
Thanks in advance,
George K. Mikros
University of Athens