# projectPart3prep.R

# Begin by setting the working directory to the folder where you have 
# put the Table files from Group Awesome, which have the duration measure 
# that they added to everyone's Table file.  
setwd('c:/Lx286/projectParts/termProjectPart3/TableFilesFromGroupAwesome')

# Set up a list of files in the directory by using grep() to identify the 
# file names returned by the dir() command which include the regular 
# expression ".Table" which all of the Table files include in their names. 
filenames=grep(".Table",dir(),value=T)

# Read the first of the files (which is Alexa.Table) into the variable 
# classVows, and then ...
classVows=read.table(filenames[1],as.is=T,header=T)
# Loop through the rest of the files, reading in each one in turn and 
# appending it as the next 10 rows of the classVows data frame using 
# the rbind() (for "bind by rows") command. 
for (i in 2:length(filenames)) {
	classVows=rbind(classVows, read.table(filenames[i],as.is=T,header=T)[,
		c("speaker","word","vowel","F1","F2","duration")])
		# We're explicitly specifying only these six columns, because 
		# Leo's file has an F3 column and won't match the data frame 
		# structure if we don't exclude that column by not naming it 
		# in this list of named columns to take. 
}

# Now set the working directory to the folder where you have put the 
# file with the corrected values for the F1 and F2 in <hawed> and <hod>
# that Group A provided.
setwd('c:/Lx286/projectParts/termProjectPart3')

# Write out a list of talker names for the Praat script. 
write.table(unique(classVows$speaker), file="talkers.txt", 
	row.names=F, col.names=F, quote=F)

# Read that file into the variable x and loop through its rows identifying 
# the appropriate cells of classVows to change. 
x=read.table("correctedproductiondatatable.txt", header=T, as.is=T)
for (i in 1:dim(x)[1]) {
	classVows$F1[(tolower(classVows$speaker)==x$speaker[i] & 
		classVows$word=="hawed")]=x$hawed.f1[i]
	classVows$F2[(tolower(classVows$speaker)==x$speaker[i] & 
		classVows$word=="hawed")]=x$hawed.f2[i]
	classVows$F1[(tolower(classVows$speaker)==x$speaker[i] & 
		classVows$word=="hod")]=x$hod.f1[i]
	classVows$F2[(tolower(classVows$speaker)==x$speaker[i] & 
		classVows$word=="hod")]=x$hod.f2[i]
}

# Check out the other formant values by looking first at the summary 
# stats for the first and second formants. 
summary(classVows[,c("F1","F2")])
#        F1                F2      
#  Min.   :  76.83   Min.   : 353  
#  1st Qu.: 449.83   1st Qu.:1318  
#  Median : 578.89   Median :1667  
#  Mean   : 602.65   Mean   :1750  
#  3rd Qu.: 743.00   3rd Qu.:2108  
#  Max.   :2095.41   Max.   :3261 
#
# That min of 77 Hz for F1 is too low.  The formant tracker must have 
# been picking up the fundamental frequency instead of the first formant 
# frequency.  And that min of 353 for F2 must be the first formant then. 
subset(classVows, F1 < 250)
#      speaker word vowel     F1      F2 duration
# 108    Elyse whod     u  76.83 1629.08    0.273
# 1016  Robert whod     u 234.00 1554.00    0.216
subset(classVows, F2 < 800)
#       speaker word vowel  F1  F2 duration
# 117 Stephanie heed     i 743 353    0.307

# Check out the other formant values by plotting each class member's vowel 
# space and seeing whether they make sense. 
windows(width=6, height=6, pointsize=12)
par(family="serif", oma=rep(0,4), mar=c(3,3,0.1,0.1), mgp=c(1.5,0.8,0), 
	type="s", ask=T)

talkers=unique(classVows$speaker)

for (i in 1:length(talkers)) {
	x=subset(classVows, speaker==talkers[i])
	plot(F1 ~ F2, data=x, xlim=c(3300,350), ylim=c(2100,75),pch=x$vowel)
	mtext(talkers[i],side=1,line=-1)
}

# Fix these values and rerun.
classVows$F1[classVows$speaker=="Brandon" & classVows$vowel=="i"]=269
classVows$F2[classVows$speaker=="Brandon" & classVows$vowel=="i"]=2025
classVows$F1[classVows$speaker=="Brandon" & classVows$vowel=="u"]=291
classVows$F2[classVows$speaker=="Brandon" & classVows$vowel=="u"]=1238
classVows$F2[classVows$speaker=="Caroline" & classVows$vowel=="e"]=2866
classVows$F1[classVows$speaker=="Carrie" & classVows$vowel=="i"]=409
classVows$F2[classVows$speaker=="Carrie" & classVows$vowel=="i"]=2901
classVows$F1[classVows$speaker=="Carrie" & classVows$vowel=="e"]=520   
classVows$F2[classVows$speaker=="Carrie" & classVows$vowel=="e"]=2747
classVows$F1[classVows$speaker=="Carrie" & classVows$vowel=="I"]=578
classVows$F2[classVows$speaker=="Carrie" & classVows$vowel=="I"]=2354
classVows$F1[classVows$speaker=="Carrie" & classVows$vowel=="E"]=779
classVows$F2[classVows$speaker=="Carrie" & classVows$vowel=="E"]=2173
classVows$F1[classVows$speaker=="Carrie" & classVows$vowel=="a"]=908
classVows$F2[classVows$speaker=="Carrie" & classVows$vowel=="a"]=1942
classVows$F1[classVows$speaker=="Carrie" & classVows$vowel=="u"]=390 
classVows$F2[classVows$speaker=="Carrie" & classVows$vowel=="u"]=1728
classVows$F1[classVows$speaker=="Carrie" & classVows$vowel=="o"]=528 
classVows$F2[classVows$speaker=="Carrie" & classVows$vowel=="o"]=1414
classVows$F1[classVows$speaker=="Carrie" & classVows$vowel=="v"]=826
classVows$F2[classVows$speaker=="Carrie" & classVows$vowel=="v"]=1538
classVows$F1[classVows$speaker=="cj" & classVows$vowel=="O"]=905
classVows$F2[classVows$speaker=="cj" & classVows$vowel=="O"]=1470
classVows$F1[classVows$speaker=="Elyse" & classVows$vowel=="u"]=406
classVows$F2[classVows$speaker=="Joe" & classVows$vowel=="u"]=1582
classVows$F1[classVows$speaker=="Leo" & classVows$vowel=="o"]=360
classVows$F2[classVows$speaker=="Leo" & classVows$vowel=="o"]=832
classVows$F1[classVows$speaker=="Leo" & classVows$vowel=="u"]=363
classVows$F2[classVows$speaker=="Leo" & classVows$vowel=="u"]=843
classVows$F1[classVows$speaker=="Mike" & classVows$vowel=="O"]=740
classVows$F2[classVows$speaker=="Mike" & classVows$vowel=="O"]=1096
classVows$F1[classVows$speaker=="Stephanie" & classVows$vowel=="i"]=400
classVows$F2[classVows$speaker=="Stephanie" & classVows$vowel=="i"]=3069
classVows$F1[classVows$speaker=="Stephanie" & classVows$vowel=="u"]=406
classVows$F2[classVows$speaker=="Stephanie" & classVows$vowel=="u"]=1410
classVows$F1[classVows$speaker=="Stephanie" & classVows$vowel=="o"]=544
classVows$F2[classVows$speaker=="Stephanie" & classVows$vowel=="o"]=1184


summary(classVows$F1)
#    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
#  234.0   426.0   564.9   583.4   716.8  1142.0
summary(classVows$F2)
#    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
#     832    1284    1619    1703    2042    3069

for (i in 1:length(talkers)) {
	x=subset(classVows, speaker==talkers[i])
	plot(F1 ~ F2, data=x, xlim=c(3070,830), ylim=c(1150,230),pch=x$vowel)
	mtext(talkers[i],side=1,line=-1)
}

write.table(classVows, file="classVowels.txt", row.names=F)