MIT OpenCourseWare http://ocw.mit.edu24.910 Topics in Linguistic Theory: Laboratory PhonologySpring 2007 For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.24.910 Laboratory Phonology Acoustics assignment Flemming Acoustics Assignment 2: Due 3/6 1. We are planning to digitize a recording. We need to be able to measure the first threeformants of a ‘schwa’ vowel produced by the speaker, who has a vocal tract of about15 cm in length. What is the lowest sampling rate that we could use? (Explain your answer). 2. Sound waves reach the eardrum through the ear canal (or external auditory meatus).Given that the ear canal is a tube of about 2.5cm long, open at one end, what effectwill it have on the sound wave at the eardrum? (Be as precise as possible - i.e. use a simple calculation to estimate the effect). 3. Liljencrants & Lindblom (1972) make crucial use of an estimate of the range ofvowel formant patterns that can be produced by the human vocal tract. This is shownas regions in F1×F2 and F2×F3 spaces on p.840. Note that these figures plot theformant frequencies on a non-linear (mel) scale and the axes are not arranged to mirror a vowel chart. The F1×F2 space is plotted on linear Hz scales on p.843, and the coordinates of points on the edge of the space are given on the same page. This vowel space is based on a fairly complex model of articulation and its acoustic consequences developed by Lindblom & Sundberg. The assignment here is to try to explain the shape of the vowel space using simple tube models of vowels. I don’t expect you to be able to derive the shape of the whole space, but see if you can account for key features of the space. In particular: How do you achieve minimum and maximum F1 and F2 values? What determines the values of these maxima and minima? Why does the range of possible F2 values narrow as F1 increases (i.e. why is the F1×F2 vowel space approximately triangular)? Tube models for vowels are discussed in Johnson pp. 102-107. The model in fig. 6.1 is mainly suitable for modeling low vowels. The model in fig. 6.3 is suitable for modeling high and mid vowels. Bear in mind that the area of the constriction (Ac) can be varied as well as the tube lengths. Johnson doesn’t say much about lip rounding. The effect of a constriction at the lips is equivalent to lengthening the front cavity (1-2cm is probably a reasonable amount, or you can estimate the lengthening required to produce formant values similar to those assumed by Liljencrants &