Notes from Thursday, July 27:

1) Stimuli for Hardman (forthcoming) 
Jocelyn described some of the research questions in her
dissertation, and we looked at (listened to) some of the 
sample stimuli that she had made from sentences produced 
by one of her American English talkers.  We talked about 
effects of noise on L2 intelligibility.

2) Korean adaptation of English /kw/ versus /tw/
We looked at (listened to and looked at spectrograms) of 
some of Mira's recordings of Korean /twi/ versus /ti/ and 
/kwi/ versus /ki/.  Those of us who are native speakers 
of Korean hear both /twi/ and /kwi/ as clusters before 
an /i/, whereas those of us who are native speakers of  
English hear /tSI/ (where /I/ is the central vowel) rather 
than /twi/ (although we hear a cluster for /kwi/).  Mira 
will look at Hyeon-seok Kang's dissertation, to see whether 
he describes the /tw/ simplification as further advanced 
than the /kw/ simplification.

Reference:
Hyeon-Seok Kang (1997). Phonological variation in glides 
and diphthongs of Seoul Korean: its synchrony and diachrony.
Doctoral dissertation, Ohio State University.

3) Michigander vowels

We opened the file m01er.wav together with the associated
TextGrid and extracted a formant listing at the midpoint 
with the "standard" settings under the Formant menu (i.e.,
0.025 second windows and 5 formants (=10 poles?) between 
0 and 5000 Hz), we got the following values:

Time_s   F1_Hz   F2_Hz   F3_Hz   F4_Hz
0.372835   520.393153   1486.996964   1895.902066   3385.715233

whereas when we changed the Formant settings to 0.008 seconds
and 7 formants (=14 poles?) between 0 and 7200 Hz (trying to
replicate the settings in the Hillenbrand et al. paper, we got:

Time_s   F1_Hz   F2_Hz   F3_Hz   F4_Hz
0.372835   542.347884   1444.472414   1876.866171   3334.411509

Lesson:  The values you extract depend heavily on the 
model you assume.  A methods section of a paper that 
shows formant plots that doesn't specify this is a bit 
defective.

========================================================

Preparation for class on Friday, July 28:

1) Try to extract for just one speaker from the Hillenbrand
et al. corpus, a set of values to plot in something like 
Figure 1 in the article.  (Don't spend more than 20 minutes 
on this.)

2) Play with the R script that you can download from 
http://ling.osu.edu/~mbeckman/795.10/scripts until you 
feel comfortable with it.  Try to figure out what some 
of the commands do. 

3) Think more about how to represent "the phonetic space"