1
00:00:00,060 --> 00:00:02,400
The following content is
provided under a Creative

2
00:00:02,400 --> 00:00:03,790
Commons license.

3
00:00:03,790 --> 00:00:06,030
Your support will help
MIT OpenCourseWare

4
00:00:06,030 --> 00:00:10,120
continue to offer high-quality
educational resources for free.

5
00:00:10,120 --> 00:00:12,660
To make a donation or to
view additional materials

6
00:00:12,660 --> 00:00:16,590
from hundreds of MIT courses,
visit MIT OpenCourseWare

7
00:00:16,590 --> 00:00:17,966
at ocw.mit.edu.

8
00:00:20,860 --> 00:00:22,110
PROFESSOR: So the cochlea is--

9
00:00:22,110 --> 00:00:23,318
I'm just going to close this.

10
00:00:23,318 --> 00:00:23,850
Is that OK?

11
00:00:26,986 --> 00:00:29,940
The cochlea is long and skinny--

12
00:00:29,940 --> 00:00:31,220
and maybe not?

13
00:00:36,840 --> 00:00:39,305
Maybe not.

14
00:00:39,305 --> 00:00:40,680
And the different
parts of it are

15
00:00:40,680 --> 00:00:44,350
sensitive to
different frequencies.

16
00:00:44,350 --> 00:00:45,360
Ringing any bells?

17
00:00:45,360 --> 00:00:46,100
Wow.

18
00:00:46,100 --> 00:00:46,800
Thanks.

19
00:00:46,800 --> 00:00:49,750
AUDIENCE: That's where the
higher [INAUDIBLE] I think,

20
00:00:49,750 --> 00:00:54,539
Abby said yesterday that
for really low tones,

21
00:00:54,539 --> 00:01:01,844
it's basically an increase
in the frequency [INAUDIBLE]

22
00:01:01,844 --> 00:01:05,741
but that at a certain point if
that becomes impractical then--

23
00:01:05,741 --> 00:01:09,540
and then certain
areas are [INAUDIBLE]

24
00:01:09,540 --> 00:01:13,960
PROFESSOR: So I think
mainly for the frequencies

25
00:01:13,960 --> 00:01:14,810
that speech are in--

26
00:01:14,810 --> 00:01:18,080
I'm here to talk about
language, by the way--

27
00:01:18,080 --> 00:01:20,480
the frequencies
of speech are more

28
00:01:20,480 --> 00:01:23,780
like the different
frequencies of the cochlea.

29
00:01:23,780 --> 00:01:26,090
Different parts are sensitive
to different frequencies.

30
00:01:26,090 --> 00:01:28,920
So I'm just going to go
with that story for now.

31
00:01:28,920 --> 00:01:32,345
But I'm sure Abby knows
more about this than I do.

32
00:01:32,345 --> 00:01:34,310
The different frequencies
of the cochlea

33
00:01:34,310 --> 00:01:37,050
are sensitive to different
frequencies and sound.

34
00:01:37,050 --> 00:01:39,545
Do you guys know
much about waves?

35
00:01:43,580 --> 00:01:47,810
Basically, any signal,
like any sound signal

36
00:01:47,810 --> 00:01:54,290
can be analyzed as a composite
of different frequencies

37
00:01:54,290 --> 00:01:55,490
of waves.

38
00:01:55,490 --> 00:02:01,430
So it's going to be
maybe a lot of 400 Hertz

39
00:02:01,430 --> 00:02:05,540
and a little bit of 600
Hertz or what have you,

40
00:02:05,540 --> 00:02:09,229
and you can basically
entirely understand

41
00:02:09,229 --> 00:02:12,480
the sound, more or less,
based on this analysis.

42
00:02:12,480 --> 00:02:13,880
So that's what the cochlea does.

43
00:02:13,880 --> 00:02:15,200
Do you understand this?

44
00:02:15,200 --> 00:02:19,700
So cochlea will tell
you, at any given moment,

45
00:02:19,700 --> 00:02:25,430
how present each of
the frequencies are.

46
00:02:25,430 --> 00:02:27,650
Nod if you understand
what I'm saying.

47
00:02:27,650 --> 00:02:29,270
Go like this.

48
00:02:29,270 --> 00:02:33,892
Do like this, if you
don't understand.

49
00:02:33,892 --> 00:02:34,823
No?

50
00:02:34,823 --> 00:02:35,323
Yes?

51
00:02:38,966 --> 00:02:40,090
The cochlea is in your ear.

52
00:02:40,090 --> 00:02:42,898
AUDIENCE: [INAUDIBLE]

53
00:02:42,898 --> 00:02:44,890
That was nice of
you guys to tell us.

54
00:02:44,890 --> 00:02:46,000
Hey, Sara--

55
00:02:46,000 --> 00:02:49,270
PROFESSOR: --sensitive
to different frequencies.

56
00:02:49,270 --> 00:02:51,280
Also sensitive to
different frequencies

57
00:02:51,280 --> 00:02:54,820
is my handy dandy
computer program,

58
00:02:54,820 --> 00:02:57,966
which I will now show you.

59
00:02:57,966 --> 00:02:59,090
So this is what it sounds--

60
00:02:59,090 --> 00:02:59,820
COMPUTER: b.

61
00:02:59,820 --> 00:03:00,760
[LAUGHTER]

62
00:03:00,760 --> 00:03:02,343
PROFESSOR: I don't
know why it shakes.

63
00:03:05,280 --> 00:03:07,240
Is says bee.

64
00:03:07,240 --> 00:03:13,090
So we have a beh sound going
on and then an e vowel.

65
00:03:13,090 --> 00:03:19,000
So what's up here is just plain
and simple, just the wave form.

66
00:03:19,000 --> 00:03:22,840
And what's going on
down here is what

67
00:03:22,840 --> 00:03:26,770
I was talking about before
with the frequency analysis.

68
00:03:26,770 --> 00:03:31,660
So basically, along the
x-axis, we have time.

69
00:03:31,660 --> 00:03:35,620
And then on the y-axis, we have
many different frequencies.

70
00:03:35,620 --> 00:03:40,000
And then, the
darkness of the dot

71
00:03:40,000 --> 00:03:44,530
is how strong that
frequency is at that time.

72
00:03:44,530 --> 00:03:46,820
Does this make sense?

73
00:03:46,820 --> 00:03:49,370
So what you can
see, for example,

74
00:03:49,370 --> 00:03:52,414
is that during this e vowel--

75
00:03:52,414 --> 00:03:53,740
COMPUTER: e.

76
00:03:53,740 --> 00:03:58,870
PROFESSOR: During the e vowel,
we have many striations,

77
00:03:58,870 --> 00:03:59,950
like up and down stripes.

78
00:03:59,950 --> 00:04:01,090
Do you see this?

79
00:04:01,090 --> 00:04:02,470
Up and down stripes.

80
00:04:02,470 --> 00:04:07,360
And that's from, just plain
and simple, the vocal chords--

81
00:04:07,360 --> 00:04:11,870
just the sound
going up and down.

82
00:04:11,870 --> 00:04:15,071
So sound is a periodic function,
sometimes high pressure,

83
00:04:15,071 --> 00:04:16,029
sometimes low pressure.

84
00:04:16,029 --> 00:04:18,490
And that's what
that's all about.

85
00:04:18,490 --> 00:04:20,050
So does this make sense?

86
00:04:20,050 --> 00:04:21,070
So this is voice.

87
00:04:21,070 --> 00:04:23,800
And you can tell that it's voice
because of all these stations,

88
00:04:23,800 --> 00:04:24,510
especially down here.

89
00:04:24,510 --> 00:04:27,051
And this is a pitch track, which
we should probably turn off.

90
00:04:31,010 --> 00:04:34,776
So does everyone understand
what's going on here?

91
00:04:34,776 --> 00:04:35,975
You're not at all confused?

92
00:04:39,450 --> 00:04:42,270
So let me explain
in more detail.

93
00:04:42,270 --> 00:04:45,960
So let's say we were
to turn on the lights.

94
00:04:49,360 --> 00:05:08,190
Then, the vocal track
looks kind of like this.

95
00:05:17,810 --> 00:05:20,810
This would be the mouth opening.

96
00:05:20,810 --> 00:05:22,910
These are the opening parts.

97
00:05:22,910 --> 00:05:24,630
This is the inside
of your mouth.

98
00:05:24,630 --> 00:05:28,310
And then you have
here your windpipe.

99
00:05:28,310 --> 00:05:31,790
And somewhere below
here are some lungs.

100
00:05:31,790 --> 00:05:34,610
And somewhere down here
you have your vocal folds,

101
00:05:34,610 --> 00:05:36,570
and those are going to
vibrate create voicing.

102
00:05:36,570 --> 00:05:38,420
So anytime you
say a voice vowel,

103
00:05:38,420 --> 00:05:41,780
like b or ah or
whatever, you're going

104
00:05:41,780 --> 00:05:45,590
to have these guys vibrating.

105
00:05:45,590 --> 00:05:50,000
And they're going to create
those up and down striations.

106
00:05:50,000 --> 00:05:55,070
So the question is how do
you make different sounds?

107
00:05:55,070 --> 00:05:55,880
I'll ask you.

108
00:05:55,880 --> 00:05:58,205
How do you make
different sounds?

109
00:05:58,205 --> 00:06:00,920
AUDIENCE: Obstructing the
air flow from the windpipe

110
00:06:00,920 --> 00:06:06,609
use your mouth, different
parts of your mouth.

111
00:06:06,609 --> 00:06:07,650
PROFESSOR: Yeah, exactly.

112
00:06:07,650 --> 00:06:10,110
So basically, the difference
between different sounds

113
00:06:10,110 --> 00:06:11,985
is going to have something
to do with the way

114
00:06:11,985 --> 00:06:13,290
this whole thing is shaped.

115
00:06:13,290 --> 00:06:15,130
Does that make sense?

116
00:06:15,130 --> 00:06:18,090
So for example, if you
were to make a buh,

117
00:06:18,090 --> 00:06:20,310
you would close your lips.

118
00:06:20,310 --> 00:06:23,460
So then this would be
blocked off here, and then

119
00:06:23,460 --> 00:06:26,281
suddenly opened.

120
00:06:26,281 --> 00:06:30,150
Go like this if you understand.

121
00:06:30,150 --> 00:06:34,170
And then the way your
tongue is, is going

122
00:06:34,170 --> 00:06:36,132
to make this shape different.

123
00:06:36,132 --> 00:06:37,590
And then the way
you're glottis is,

124
00:06:37,590 --> 00:06:39,150
is going to make
the back different

125
00:06:39,150 --> 00:06:40,500
and all this sort of stuff.

126
00:06:40,500 --> 00:06:44,385
And this is how you make
the different sounds.

127
00:06:51,695 --> 00:06:53,820
So for example, the way
you make different vowels--

128
00:06:53,820 --> 00:07:00,090
if you try saying like awe, e,
awe, e, or something like that,

129
00:07:00,090 --> 00:07:03,540
or pretend like
you're doing it, then

130
00:07:03,540 --> 00:07:06,150
it's mainly your tongue moving.

131
00:07:06,150 --> 00:07:08,210
You can kind of--

132
00:07:08,210 --> 00:07:10,290
if you want to try
it, it would be OK.

133
00:07:10,290 --> 00:07:13,680
awe, e-- it's
basically your tongue

134
00:07:13,680 --> 00:07:15,390
going one way or another.

135
00:07:15,390 --> 00:07:18,630
And your tongue is going to
create some sort of obstruction

136
00:07:18,630 --> 00:07:23,070
somewhere in here for each one.

137
00:07:23,070 --> 00:07:27,300
If you think about it,
obstructing any part

138
00:07:27,300 --> 00:07:30,300
of your--of this whole pipe
is going to make different

139
00:07:30,300 --> 00:07:31,920
cavities of different sizes.

140
00:07:31,920 --> 00:07:32,950
Does that make sense?

141
00:07:32,950 --> 00:07:34,026
So if you were to have--

142
00:07:34,026 --> 00:07:35,400
this is just purely
theoretical--

143
00:07:35,400 --> 00:07:38,550
if you were to have an
obstruction like here,

144
00:07:38,550 --> 00:07:40,400
then you would have
a cavity of this size

145
00:07:40,400 --> 00:07:42,892
and a cavity of that size.

146
00:07:42,892 --> 00:07:45,100
If you were to make one here
or something like that--

147
00:07:45,100 --> 00:07:46,740
I'm not sure you could--

148
00:07:46,740 --> 00:07:52,050
but then you would have you--
you could have multiple ones.

149
00:07:52,050 --> 00:07:55,149
If you were to do
this with your lips,

150
00:07:55,149 --> 00:07:56,940
you would make this
one a little bit bigger

151
00:07:56,940 --> 00:08:01,940
by a few centimeters,
two centimeters or so.

152
00:08:01,940 --> 00:08:04,170
Does that make sense?

153
00:08:04,170 --> 00:08:06,130
So what do you guys
know about resonance?

154
00:08:10,370 --> 00:08:10,870
Yeah?

155
00:08:10,870 --> 00:08:15,438
AUDIENCE: When two
waves have the similar--

156
00:08:15,438 --> 00:08:18,140
hold the same
frequency, they can

157
00:08:18,140 --> 00:08:19,472
have a fight with each other.

158
00:08:19,472 --> 00:08:20,180
PROFESSOR: Right.

159
00:08:20,180 --> 00:08:26,020
So basically anything has what's
called a fundamental frequency.

160
00:08:26,020 --> 00:08:29,420
And this is the
frequency of a wave that

161
00:08:29,420 --> 00:08:35,840
would have a whole number
of wavelengths with respect

162
00:08:35,840 --> 00:08:37,100
to the size of the thing.

163
00:08:37,100 --> 00:08:38,809
Something like that.

164
00:08:38,809 --> 00:08:41,570
So for instance, the
cavities in your vocal cord,

165
00:08:41,570 --> 00:08:44,570
like the vocal tract,
the various parts

166
00:08:44,570 --> 00:08:46,910
of your vocal tract that may
or may not be blocked off,

167
00:08:46,910 --> 00:08:50,276
are going to have different
resonant frequencies.

168
00:08:50,276 --> 00:08:51,470
Are you following?

169
00:08:51,470 --> 00:08:54,290
So it depends on the
size of the thing.

170
00:08:54,290 --> 00:08:57,480
So the bigger part
of your vocal tract

171
00:08:57,480 --> 00:09:00,401
is, the lower the
resonant frequency.

172
00:09:00,401 --> 00:09:02,150
And then smaller parts
of your vocal tract

173
00:09:02,150 --> 00:09:05,790
are going to have higher
resonant frequencies.

174
00:09:05,790 --> 00:09:07,790
So if you're sending
air through and you're

175
00:09:07,790 --> 00:09:10,370
making sound, what you're
basically going to get

176
00:09:10,370 --> 00:09:12,740
is that certain
frequencies are going

177
00:09:12,740 --> 00:09:16,220
to resonate and be really loud.

178
00:09:16,220 --> 00:09:18,140
And that's what we
see in this picture.

179
00:09:18,140 --> 00:09:21,013
Can you still see it,
even though the lights?

180
00:09:21,013 --> 00:09:21,655
AUDIENCE: Yes.

181
00:09:21,655 --> 00:09:22,280
PROFESSOR: Yes?

182
00:09:22,280 --> 00:09:22,779
OK.

183
00:09:22,779 --> 00:09:23,895
Good.

184
00:09:23,895 --> 00:09:25,520
So that's what you
see in this picture.

185
00:09:28,320 --> 00:09:31,840
So for b-- or for the e vowel--

186
00:09:31,840 --> 00:09:33,800
COMPUTER: e.

187
00:09:33,800 --> 00:09:38,180
PROFESSOR: We have this
is the frequency of one

188
00:09:38,180 --> 00:09:42,290
of the big cavities
in the vocal tract.

189
00:09:42,290 --> 00:09:47,090
And this is going to be the
frequency of another one,

190
00:09:47,090 --> 00:09:49,400
and this is going
to be another one.

191
00:09:49,400 --> 00:09:51,560
So this is going to
be the biggest one.

192
00:09:51,560 --> 00:09:54,200
This is the next biggest
one, et cetera, et cetera.

193
00:09:54,200 --> 00:09:55,519
These are called formants.

194
00:09:55,519 --> 00:09:56,310
I'll write it down.

195
00:10:07,900 --> 00:10:10,730
So does that make sense?

196
00:10:10,730 --> 00:10:13,810
So basically,
changing your mouth,

197
00:10:13,810 --> 00:10:16,060
changing the shape of your
mouth and your vocal tract,

198
00:10:16,060 --> 00:10:22,900
is going to directly change the
frequencies of the formants.

199
00:10:22,900 --> 00:10:26,530
This program is basically doing
exactly what the cochlea does.

200
00:10:26,530 --> 00:10:29,230
So this program takes all of
the frequencies and figures

201
00:10:29,230 --> 00:10:33,100
out which are the strongest
and which are the weakest.

202
00:10:33,100 --> 00:10:38,560
It basically analyzes
the different magnitudes

203
00:10:38,560 --> 00:10:40,420
of each of the different
frequencies that

204
00:10:40,420 --> 00:10:41,710
are all available.

205
00:10:45,390 --> 00:10:49,450
So basically, that seems to be
what the speech signal is all

206
00:10:49,450 --> 00:10:50,860
about.

207
00:10:50,860 --> 00:10:53,960
So when you're producing
the speech signal,

208
00:10:53,960 --> 00:10:57,700
you're going to be changing
the shape of the vocal tract,

209
00:10:57,700 --> 00:11:00,029
and that's going to be
changing what the formants are

210
00:11:00,029 --> 00:11:00,820
going to look like.

211
00:11:00,820 --> 00:11:04,390
And your cochlea is mapping
out the formants for you.

212
00:11:04,390 --> 00:11:06,304
That's what it's doing.

213
00:11:06,304 --> 00:11:07,220
Does this makes sense?

214
00:11:11,020 --> 00:11:13,570
Then, we can see what
is the difference

215
00:11:13,570 --> 00:11:15,790
between different vowels.

216
00:11:15,790 --> 00:11:17,390
I'll just play the whole thing.

217
00:11:17,390 --> 00:11:23,170
COMPUTER: Dee, dee, bah, dah.

218
00:11:23,170 --> 00:11:27,040
PROFESSOR: So let's just look
at the difference between--

219
00:11:27,040 --> 00:11:29,020
let's see, this is
dee and this is bah.

220
00:11:34,920 --> 00:11:39,060
What are the differences
between the formants

221
00:11:39,060 --> 00:11:42,030
of the dee and bah here?

222
00:11:56,770 --> 00:11:59,628
AUDIENCE: There's a
very big one which has

223
00:11:59,628 --> 00:12:04,100
a low frequency [INAUDIBLE].

224
00:12:04,100 --> 00:12:05,066
PROFESSOR: For dee?

225
00:12:05,066 --> 00:12:06,770
AUDIENCE: Yes.

226
00:12:06,770 --> 00:12:08,610
PROFESSOR: So the formant one--

227
00:12:08,610 --> 00:12:09,560
this is formant one.

228
00:12:09,560 --> 00:12:13,410
It's the bottom one--

229
00:12:13,410 --> 00:12:17,715
is lower in dee than in bah.

230
00:12:17,715 --> 00:12:19,760
Are people seeing this?

231
00:12:19,760 --> 00:12:24,380
This one is lower
frequency than this one.

232
00:12:24,380 --> 00:12:27,609
Maybe if we were to show the--

233
00:12:27,609 --> 00:12:29,319
AUDIENCE: [INAUDIBLE]

234
00:12:29,319 --> 00:12:30,860
PROFESSOR: So the
red dots are trying

235
00:12:30,860 --> 00:12:32,070
to track performance here.

236
00:12:32,070 --> 00:12:34,920
So this one is
lower than this one.

237
00:12:34,920 --> 00:12:36,962
Do you see that?

238
00:12:36,962 --> 00:12:37,920
How did the formant do?

239
00:12:46,884 --> 00:12:51,366
AUDIENCE: In the awe
sounds, it's one or two

240
00:12:51,366 --> 00:12:54,670
lower [INAUDIBLE]

241
00:12:54,670 --> 00:13:00,380
PROFESSOR: So formant two
is very high in e sound

242
00:13:00,380 --> 00:13:03,092
and very low in the awe sound.

243
00:13:03,092 --> 00:13:04,550
So this is basically
the difference

244
00:13:04,550 --> 00:13:09,935
between what's called
a high vowel, like e,

245
00:13:09,935 --> 00:13:14,330
and a low vowel, like awe, is
the differences in formants one

246
00:13:14,330 --> 00:13:15,020
and two.

247
00:13:15,020 --> 00:13:17,861
If you look at the
spectrum of vowels--

248
00:13:17,861 --> 00:13:18,360
spectrum.

249
00:13:18,360 --> 00:13:20,750
If you look at--

250
00:13:20,750 --> 00:13:22,400
all the vowel
systems in the world,

251
00:13:22,400 --> 00:13:25,850
seem to have vowels that--

252
00:13:25,850 --> 00:13:28,110
there's sort of a
trapezoidal thing going.

253
00:13:28,110 --> 00:13:32,080
So e is over here,
and ooh is over here.

254
00:13:32,080 --> 00:13:40,640
And awe and ah is over here,
and like uh, oh and eh.

255
00:13:40,640 --> 00:13:43,370
So this is some of
the vowels of English.

256
00:13:52,530 --> 00:13:56,100
If you look at vowel
systems, languages

257
00:13:56,100 --> 00:13:59,610
can have three vowels systems
or five vowels systems

258
00:13:59,610 --> 00:14:03,090
or seven or eight
or however many,

259
00:14:03,090 --> 00:14:06,120
but they're all going to
be distributed in here.

260
00:14:06,120 --> 00:14:09,049
And if you look at actually
the formants of these,

261
00:14:09,049 --> 00:14:10,590
then the big
differences between them

262
00:14:10,590 --> 00:14:15,660
is going to be between
formants one and two.

263
00:14:15,660 --> 00:14:19,650
So formant one is going to be
highest for the low vowels,

264
00:14:19,650 --> 00:14:22,470
and lowest for the high vowels.

265
00:14:22,470 --> 00:14:25,430
And formant two is basically
going to go the other way.

266
00:14:25,430 --> 00:14:25,930
No.

267
00:14:25,930 --> 00:14:27,830
Formant two is going
to be front back.

268
00:14:27,830 --> 00:14:30,945
Formant two is going to
create this distinction.

269
00:14:33,925 --> 00:14:35,550
And this all happens
because of the way

270
00:14:35,550 --> 00:14:37,216
the tongue is when
you say these vowels.

271
00:14:37,216 --> 00:14:42,210
So when you say awe, e, you're
making different obstructions.

272
00:14:42,210 --> 00:14:46,910
You're basically dividing the
vocal tract in different ways

273
00:14:46,910 --> 00:14:49,410
and creating different
resonant frequencies.

274
00:14:49,410 --> 00:14:51,295
And that's what's going on.

275
00:14:51,295 --> 00:14:52,170
Does this make sense?

276
00:14:55,150 --> 00:14:56,690
So that's what the
vowels are about.

277
00:15:00,340 --> 00:15:06,692
I just want to tell you
maybe about consonants.

278
00:15:12,509 --> 00:15:13,300
I'll play it again.

279
00:15:13,300 --> 00:15:18,640
COMPUTER: Bee, dee, baw, daw.

280
00:15:18,640 --> 00:15:24,390
PROFESSOR: So if you look at the
wave form, this is for, what?

281
00:15:24,390 --> 00:15:24,890
Dee?

282
00:15:24,890 --> 00:15:25,500
COMPUTER: Bee.

283
00:15:25,500 --> 00:15:26,590
PROFESSOR: Bee, yeah.

284
00:15:26,590 --> 00:15:28,430
This is for bee.

285
00:15:28,430 --> 00:15:31,970
So if you divide this
up, as it works out,

286
00:15:31,970 --> 00:15:34,730
there is silence for a while.

287
00:15:34,730 --> 00:15:38,120
This is the part that happens
when your mouth is closed

288
00:15:38,120 --> 00:15:43,240
and you're saying the part
that happens before the buh.

289
00:15:43,240 --> 00:15:47,180
So the buh sound is a release.

290
00:15:47,180 --> 00:15:50,090
I don't know if you've
noticed, but if you say bee,

291
00:15:50,090 --> 00:15:52,730
there's part of it where
your mouth is closed.

292
00:15:52,730 --> 00:15:55,070
And then, when your
mouth opens, a buh sound

293
00:15:55,070 --> 00:15:58,530
happens, and then the vowel.

294
00:15:58,530 --> 00:16:01,550
So this is the part where
your mouth is closed.

295
00:16:01,550 --> 00:16:03,830
And then, this is going
to be the release.

296
00:16:03,830 --> 00:16:05,413
I don't know if you
can hear anything.

297
00:16:07,460 --> 00:16:09,920
That's the consonant
itself-- the buh.

298
00:16:09,920 --> 00:16:14,500
And then it's going to open
up into the vowel itself.

299
00:16:14,500 --> 00:16:17,057
COMPUTER: Bee, bee.

300
00:16:17,057 --> 00:16:18,890
PROFESSOR: The reason
it sounds like there's

301
00:16:18,890 --> 00:16:20,640
a consonant at the
beginning is because it

302
00:16:20,640 --> 00:16:23,900
starts kind of abruptly.

303
00:16:23,900 --> 00:16:26,720
So if you were to say a
vowel with no consonant

304
00:16:26,720 --> 00:16:30,770
at the beginning at all,
it would start very quietly

305
00:16:30,770 --> 00:16:32,990
and get louder and
then get softer again.

306
00:16:32,990 --> 00:16:35,480
So you can see that
it gets softer again.

307
00:16:35,480 --> 00:16:40,240
But it starts loudly because
it's preceded by this stop--

308
00:16:40,240 --> 00:16:41,240
the buh sound.

309
00:16:41,240 --> 00:16:42,530
COMPUTER: Bee.

310
00:16:42,530 --> 00:16:45,770
PROFESSOR: So this
is all the vowel,

311
00:16:45,770 --> 00:16:48,950
and this is all the consonant.

312
00:16:48,950 --> 00:16:53,120
What is mainly the difference
in volume between the vowel

313
00:16:53,120 --> 00:16:54,188
and the consonant?

314
00:16:57,604 --> 00:16:59,556
Yeah?

315
00:16:59,556 --> 00:17:07,852
AUDIENCE: The consonant
doesn't have much volume.

316
00:17:07,852 --> 00:17:09,930
PROFESSOR: The consonant
is really quiet

317
00:17:09,930 --> 00:17:12,119
compared to the vowel.

318
00:17:12,119 --> 00:17:13,670
So the consonant is going to--

319
00:17:13,670 --> 00:17:15,533
this is what the
consonant sounds like.

320
00:17:15,533 --> 00:17:16,979
COMPUTER: [SOUND SIGNAL]

321
00:17:16,979 --> 00:17:17,770
PROFESSOR: Versus--

322
00:17:17,770 --> 00:17:18,970
COMPUTER: Bee.

323
00:17:18,970 --> 00:17:23,109
PROFESSOR: So very quiet
compared to the vowel.

324
00:17:23,109 --> 00:17:25,720
And let's maybe look at dee.

325
00:17:32,030 --> 00:17:33,905
COMPUTER: Dee, dee.

326
00:17:33,905 --> 00:17:35,280
PROFESSOR: So how
about this one?

327
00:17:40,690 --> 00:17:44,170
Can you tell this is where
the vowel is happening?

328
00:17:44,170 --> 00:17:45,250
COMPUTER: Dee.

329
00:17:45,250 --> 00:17:47,625
PROFESSOR: And this
is the consonant.

330
00:17:47,625 --> 00:17:50,535
COMPUTER: [SOUND SIGNAL]

331
00:17:52,970 --> 00:17:54,980
PROFESSOR: So what do
we think about this?

332
00:17:54,980 --> 00:17:58,002
AUDIENCE: Well, in
this one, the duh sound

333
00:17:58,002 --> 00:18:01,923
doesn't obstruct the airflow,
so it's the starting point--

334
00:18:01,923 --> 00:18:03,370
or the duh sound.

335
00:18:03,370 --> 00:18:06,165
And the starting
point of the e sound

336
00:18:06,165 --> 00:18:11,590
is at the same
volume [INAUDIBLE]

337
00:18:11,590 --> 00:18:13,590
PROFESSOR: Yeah,
I mean, it looks

338
00:18:13,590 --> 00:18:17,490
like what's happening here
is that this is the release.

339
00:18:17,490 --> 00:18:19,830
This is where the
tongue releases so you

340
00:18:19,830 --> 00:18:23,340
don't have a stop anymore--

341
00:18:23,340 --> 00:18:26,230
stop in the air flow-- you don't
have that anymore right here.

342
00:18:26,230 --> 00:18:28,060
And then, there's a
little bit of noise,

343
00:18:28,060 --> 00:18:31,440
which is probably just air
passing out of the mouth.

344
00:18:31,440 --> 00:18:33,390
And then the vowel
seems to start.

345
00:18:33,390 --> 00:18:35,040
But it is louder.

346
00:18:35,040 --> 00:18:38,880
The stop itself is
louder, but certainly--

347
00:18:38,880 --> 00:18:42,360
louder than the buh one,
but certainly not as loud

348
00:18:42,360 --> 00:18:43,320
as a vowel--

349
00:18:43,320 --> 00:18:47,530
or not the main part
of the vowel anyway.

350
00:18:47,530 --> 00:18:50,570
And you can see the
difference here.

351
00:18:50,570 --> 00:18:56,690
So duh part looks like this.

352
00:18:56,690 --> 00:19:00,400
So it's got the
beginnings of a formants.

353
00:19:00,400 --> 00:19:02,110
You're ready to say the vowel.

354
00:19:02,110 --> 00:19:04,860
Your mouth is shaped that way.

355
00:19:04,860 --> 00:19:07,660
But you don't really--

356
00:19:07,660 --> 00:19:10,780
they don't get really
dark until here.

357
00:19:10,780 --> 00:19:13,030
This is when you can
really see stuff going on.

358
00:19:13,030 --> 00:19:14,050
So it's also quieter.

359
00:19:16,630 --> 00:19:18,580
Basically, what
I'm trying to say

360
00:19:18,580 --> 00:19:22,610
is that consonants are
much quieter than vowels.

361
00:19:27,070 --> 00:19:30,360
So we can look at these and
see if they're the same.

362
00:19:33,160 --> 00:19:40,600
COMPUTER: Baw, baw, baw.

363
00:19:40,600 --> 00:19:43,580
PROFESSOR: I don't
know why it shakes.

364
00:19:43,580 --> 00:19:46,943
So this is sort of the same
thing going on, versus--

365
00:19:52,570 --> 00:19:54,944
COMPUTER: Daw, daw.

366
00:19:54,944 --> 00:19:57,485
PROFESSOR: So again, we see that
duh is a little bit louder--

367
00:19:57,485 --> 00:19:59,015
COMPUTER: [SOUND SIGNAL]

368
00:19:59,015 --> 00:20:00,530
PROFESSOR: --than the buh was.

369
00:20:00,530 --> 00:20:03,140
But it's still nowhere near
as loud as the main part

370
00:20:03,140 --> 00:20:05,425
of the vowel.

371
00:20:05,425 --> 00:20:08,610
AUDIENCE: What if
you drug it to that--

372
00:20:08,610 --> 00:20:11,459
can you start playing it
in the middle of the vowel?

373
00:20:11,459 --> 00:20:13,220
Would we still tend
to hear a consonant?

374
00:20:13,220 --> 00:20:14,330
PROFESSOR: You would
hear a consonant,

375
00:20:14,330 --> 00:20:15,371
but it wouldn't be a dee.

376
00:20:15,371 --> 00:20:15,975
Listen.

377
00:20:15,975 --> 00:20:20,680
COMPUTER: Awe, awe,
awe, awe, awe, awe.

378
00:20:20,680 --> 00:20:23,980
AUDIENCE: [INAUDIBLE]
It's almost a bee.

379
00:20:23,980 --> 00:20:26,450
PROFESSOR: Yeah, you could
perceive it in a few ways.

380
00:20:26,450 --> 00:20:28,660
COMPUTER: Awe, awe, awe awe.

381
00:20:33,120 --> 00:20:35,830
PROFESSOR: So it sounds
like a bee to you?

382
00:20:35,830 --> 00:20:37,810
We'll talk about that.

383
00:20:37,810 --> 00:20:39,790
So you think that
there's a consonant there

384
00:20:39,790 --> 00:20:42,510
because the vowel
started so abruptly.

385
00:20:42,510 --> 00:20:44,380
First, I didn't hit
play, and then I did.

386
00:20:49,610 --> 00:20:51,170
But that's just the only reason.

387
00:20:51,170 --> 00:20:54,440
So it doesn't have the
characteristic swell

388
00:20:54,440 --> 00:20:56,690
that the vowel has.

389
00:20:56,690 --> 00:21:01,100
People understand that though
the larger ones are louder?

390
00:21:01,100 --> 00:21:02,300
People are getting this?

391
00:21:02,300 --> 00:21:03,650
AUDIENCE: Higher
amplitudes [INAUDIBLE]

392
00:21:03,650 --> 00:21:04,358
PROFESSOR: Right.

393
00:21:04,358 --> 00:21:07,190
Higher amplitude makes it
louder, in this top part--

394
00:21:07,190 --> 00:21:09,710
the wave form.

395
00:21:09,710 --> 00:21:10,870
This is different.

396
00:21:13,580 --> 00:21:14,480
So here's a question.

397
00:21:18,020 --> 00:21:19,980
If we can't tell the
difference between--

398
00:21:19,980 --> 00:21:25,310
or if the consonants are
so quiet, then how can

399
00:21:25,310 --> 00:21:29,210
we hear the difference
between two consonants?

400
00:21:29,210 --> 00:21:30,318
Yeah?

401
00:21:30,318 --> 00:21:33,082
AUDIENCE: Well one
thing is the shape

402
00:21:33,082 --> 00:21:38,418
of the volume of the vowel.

403
00:21:38,418 --> 00:21:40,987
If you take the
e, it usually goes

404
00:21:40,987 --> 00:21:42,362
from-- it starts
from really loud

405
00:21:42,362 --> 00:21:47,292
and goes to get
softer [INAUDIBLE]

406
00:21:47,292 --> 00:21:53,980
but if you say a dee
sound, then [INAUDIBLE]

407
00:21:53,980 --> 00:21:54,880
PROFESSOR: Yeah.

408
00:21:54,880 --> 00:21:58,500
The release of the stop is
quieter in the buh sound

409
00:21:58,500 --> 00:22:01,140
than in the duh
sound, which means

410
00:22:01,140 --> 00:22:06,280
that the abruptness of the
vowel is also different--

411
00:22:06,280 --> 00:22:09,000
just plain how loud
it is, is going

412
00:22:09,000 --> 00:22:12,090
to be more abrupt in the buh
sound than in the duh sound.

413
00:22:12,090 --> 00:22:14,170
You can see it here.

414
00:22:14,170 --> 00:22:17,790
People are understanding
what I'm saying?

415
00:22:17,790 --> 00:22:19,020
So that's one thing.

416
00:22:19,020 --> 00:22:22,800
So basically, the
volume of the release.

417
00:22:22,800 --> 00:22:23,360
What else?

418
00:22:27,960 --> 00:22:30,112
This is, if anyone
forgot, this is--

419
00:22:30,112 --> 00:22:32,080
COMPUTER: Buh, duh.

420
00:22:43,122 --> 00:22:44,872
AUDIENCE: I think the
starting frequencies

421
00:22:44,872 --> 00:22:47,850
have different formants.

422
00:22:47,850 --> 00:22:50,340
PROFESSOR: Yeah, they seem
to be different, right?

423
00:22:50,340 --> 00:22:54,240
So in this one, it
looks like the formants

424
00:22:54,240 --> 00:22:56,520
are pretty steady throughout--

425
00:22:56,520 --> 00:22:59,450
formants one and two, at least,
pretty steady throughout.

426
00:22:59,450 --> 00:23:01,890
Whereas in this one, it
looks like formant one

427
00:23:01,890 --> 00:23:05,405
is starting from a lower
frequency and getting higher.

428
00:23:05,405 --> 00:23:07,530
And formant two is starting
from a higher frequency

429
00:23:07,530 --> 00:23:09,690
and getting lower.

430
00:23:09,690 --> 00:23:11,010
People see this?

431
00:23:11,010 --> 00:23:12,390
How this is happening?

432
00:23:15,810 --> 00:23:18,600
Why might this be?

433
00:23:18,600 --> 00:23:20,580
Why would we see a
difference in how

434
00:23:20,580 --> 00:23:23,640
the beginnings of the formants
are depending on the preceding

435
00:23:23,640 --> 00:23:26,490
consonant?

436
00:23:26,490 --> 00:23:27,670
Yeah?

437
00:23:27,670 --> 00:23:31,300
AUDIENCE: Well, in daw,
the shape of your tongue

438
00:23:31,300 --> 00:23:33,734
is different than when you're
emphasizing the awe part.

439
00:23:33,734 --> 00:23:40,777
[INAUDIBLE] with buh, the
tongue stays [INAUDIBLE]

440
00:23:40,777 --> 00:23:42,610
PROFESSOR: Basically,
it must have something

441
00:23:42,610 --> 00:23:45,730
to do with that time--

442
00:23:45,730 --> 00:23:49,990
your lips are shaped one way
for the duh part of the sound,

443
00:23:49,990 --> 00:23:52,930
and they're shaped another way
for the awe part of the sound.

444
00:23:52,930 --> 00:23:57,880
And the mouth wasn't
designed to talk.

445
00:23:57,880 --> 00:23:59,990
It was designed to eat.

446
00:23:59,990 --> 00:24:03,580
I don't know if you
guys realize this--

447
00:24:03,580 --> 00:24:06,880
probably, yes.

448
00:24:06,880 --> 00:24:08,650
It's slow.

449
00:24:08,650 --> 00:24:13,030
Phoneticians hypothesize
that we're basically

450
00:24:13,030 --> 00:24:16,990
speaking as fast as we can.

451
00:24:16,990 --> 00:24:19,360
That rates of speech
in different languages

452
00:24:19,360 --> 00:24:21,340
are more or less similar.

453
00:24:21,340 --> 00:24:25,600
And if they got any louder,
the mechanics of the mouth

454
00:24:25,600 --> 00:24:28,907
wouldn't actually be
able to keep up with it.

455
00:24:28,907 --> 00:24:30,490
Whether perception
can keep up with it

456
00:24:30,490 --> 00:24:31,990
is maybe another question.

457
00:24:34,690 --> 00:24:37,150
There's going to be
some sort of lag.

458
00:24:37,150 --> 00:24:39,820
There's going to be some
lag time between when

459
00:24:39,820 --> 00:24:41,710
your mouth was in the
duh shape and when

460
00:24:41,710 --> 00:24:44,590
your mouth is in the awe shape.

461
00:24:44,590 --> 00:24:48,250
And similarly, some lag between
the buh shape and the awe

462
00:24:48,250 --> 00:24:52,000
shape, and the buh
shape and the e shape,

463
00:24:52,000 --> 00:24:54,434
and all these sorts of things.

464
00:24:54,434 --> 00:24:55,850
What we're seeing
here is the lag.

465
00:24:58,540 --> 00:25:01,120
This is the part between
the way your mouth was

466
00:25:01,120 --> 00:25:04,060
shaped from the duh,
and the way your mouth

467
00:25:04,060 --> 00:25:06,400
was shaped for the awe.

468
00:25:06,400 --> 00:25:09,160
It's kind of our ramp time.

469
00:25:09,160 --> 00:25:12,550
And it's very noticeable.

470
00:25:12,550 --> 00:25:18,110
So if you look at this,
it's pretty clear.

471
00:25:18,110 --> 00:25:20,070
This is exactly what
the cochlea is doing.

472
00:25:20,070 --> 00:25:22,090
The cochlea is
basically analyzing

473
00:25:22,090 --> 00:25:25,700
how loud the different
frequencies are in real time.

474
00:25:25,700 --> 00:25:27,850
And so basically, what
the cochlea is doing

475
00:25:27,850 --> 00:25:29,950
is showing you this
picture, showing your brain

476
00:25:29,950 --> 00:25:30,940
this picture.

477
00:25:30,940 --> 00:25:33,570
So the fact that
this is happening

478
00:25:33,570 --> 00:25:36,070
is sort of really noticeable.

479
00:25:36,070 --> 00:25:41,830
Whereas-- these are getting
kind of distracting.

480
00:25:41,830 --> 00:25:47,530
Whereas, the loudness of the
buh sound versus the duh sound,

481
00:25:47,530 --> 00:25:51,070
not really as noticeable.

482
00:25:51,070 --> 00:25:53,410
Does that makes sense?

483
00:25:53,410 --> 00:26:00,918
Let's look at bee and dee
[INAUDIBLE] How about this?

484
00:26:10,670 --> 00:26:11,580
What do you guys see?

485
00:26:11,580 --> 00:26:13,906
I'll play it for you.

486
00:26:13,906 --> 00:26:18,711
COMPUTER: Bee, dee, bee, dee.

487
00:26:27,370 --> 00:26:28,078
PROFESSOR: Ideas?

488
00:26:42,466 --> 00:26:45,358
AUDIENCE: It will take
the lag time [INAUDIBLE]

489
00:26:45,358 --> 00:26:51,862
because the dee shape is
closer to-- the duh shape is

490
00:26:51,862 --> 00:26:54,250
closer to the [INAUDIBLE]

491
00:26:54,250 --> 00:26:58,110
PROFESSOR: So it seems like
we don't see that drastic lag

492
00:26:58,110 --> 00:27:01,740
time that we saw with the
awe vowel, so exactly right.

493
00:27:01,740 --> 00:27:05,580
So it must be something like the
shape of your mouth for a duh

494
00:27:05,580 --> 00:27:08,730
is closer to the shape
of your mouth for e, dee.

495
00:27:08,730 --> 00:27:10,470
I mean, you can kind of tell--

496
00:27:10,470 --> 00:27:14,370
dee-- we're thinking mostly of
what your tongue is doing here,

497
00:27:14,370 --> 00:27:16,740
because that's the
part of your mouth

498
00:27:16,740 --> 00:27:20,190
that's moving the most
in this whole process.

499
00:27:20,190 --> 00:27:25,410
And b, we also
don't see some sort

500
00:27:25,410 --> 00:27:28,650
of exciting formant transition.

501
00:27:28,650 --> 00:27:30,750
This is what is called
formant transition.

502
00:27:30,750 --> 00:27:32,760
We're talking about
the beginnings and ends

503
00:27:32,760 --> 00:27:34,380
of the formants.

504
00:27:34,380 --> 00:27:39,520
So we do see something over
here in formants three and four.

505
00:27:39,520 --> 00:27:41,730
They seem to dip.

506
00:27:41,730 --> 00:27:44,370
I'm sorry-- they seem to
rise at the beginning,

507
00:27:44,370 --> 00:27:46,410
for bee more than with dee.

508
00:27:46,410 --> 00:27:49,660
But formants one-- formant
three is certainly important.

509
00:27:49,660 --> 00:27:56,030
But it seems that the lower
numbers of formants, the lower

510
00:27:56,030 --> 00:27:59,229
formants tends to be more
important in our perception

511
00:27:59,229 --> 00:28:00,270
than the higher formants.

512
00:28:03,780 --> 00:28:05,809
I guess because they
represent larger parts

513
00:28:05,809 --> 00:28:06,600
of the vocal tract.

514
00:28:06,600 --> 00:28:07,472
Yeah?

515
00:28:07,472 --> 00:28:17,119
AUDIENCE: Does each
formant [INAUDIBLE]

516
00:28:17,119 --> 00:28:17,910
PROFESSOR: Sort of.

517
00:28:17,910 --> 00:28:27,240
So basically for formants one
and two seem to be more or less

518
00:28:27,240 --> 00:28:29,100
that--

519
00:28:29,100 --> 00:28:31,680
there's typically
one obstruction

520
00:28:31,680 --> 00:28:35,350
in the vocal tract going on at
any one time, like one big one.

521
00:28:35,350 --> 00:28:37,890
And formants one and
two seem to be related

522
00:28:37,890 --> 00:28:41,610
to each other, in that formant
one is always the bigger half

523
00:28:41,610 --> 00:28:44,890
and formant two is
always the smaller half.

524
00:28:44,890 --> 00:28:48,990
And if you actually
look at the vowels

525
00:28:48,990 --> 00:28:53,340
and you think about
which formant corresponds

526
00:28:53,340 --> 00:28:58,980
to which half, you'll see that
they switch, as one half gets

527
00:28:58,980 --> 00:29:01,540
bigger and one half gets--

528
00:29:01,540 --> 00:29:03,420
it crosses the midpoint.

529
00:29:03,420 --> 00:29:04,770
But that's beside the point.

530
00:29:07,740 --> 00:29:12,300
And then after that, are
other resonant frequencies.

531
00:29:12,300 --> 00:29:14,730
The formants are going to take
into account basically all,

532
00:29:14,730 --> 00:29:15,400
like--

533
00:29:15,400 --> 00:29:16,640
they keep going.

534
00:29:16,640 --> 00:29:19,897
I've cut this off at 5,000
Hertz, but they keep going.

535
00:29:19,897 --> 00:29:21,480
There are many, many
formants up there

536
00:29:21,480 --> 00:29:23,730
and they're resonating with
all of the different parts

537
00:29:23,730 --> 00:29:26,010
of your head, all of
your various sinuses,

538
00:29:26,010 --> 00:29:29,650
and all this sort of thing,
and to various degrees.

539
00:29:29,650 --> 00:29:31,920
So they get quieter and
quieter as they go up.

540
00:29:35,010 --> 00:29:36,750
So does that answer
your question?

541
00:29:36,750 --> 00:29:38,250
Yeah.

542
00:29:38,250 --> 00:29:40,590
So what were we saying before?

543
00:29:40,590 --> 00:29:43,680
We don't have this
drastic difference

544
00:29:43,680 --> 00:29:49,290
in formants one and two for the
difference between bee and dee.

545
00:29:49,290 --> 00:29:52,215
So maybe we're not
going to really--

546
00:29:55,320 --> 00:30:00,480
what I'm trying to
get at is, what is--

547
00:30:00,480 --> 00:30:09,110
if you look at this,
and you ask yourself,

548
00:30:09,110 --> 00:30:14,320
how does a hearer perceive
the difference between a bee--

549
00:30:14,320 --> 00:30:15,850
the difference
between baw and daw?

550
00:30:19,100 --> 00:30:23,677
You might, without
thinking, or without looking

551
00:30:23,677 --> 00:30:25,260
at this picture, you
might think that,

552
00:30:25,260 --> 00:30:29,940
well, the difference between
baw and daw is the beginning buh

553
00:30:29,940 --> 00:30:31,900
and duh sounds.

554
00:30:35,230 --> 00:30:36,830
But as you can see
from this picture,

555
00:30:36,830 --> 00:30:39,170
the actual buh
and duh sounds are

556
00:30:39,170 --> 00:30:42,140
extremely quiet with
respect to the vowel.

557
00:30:42,140 --> 00:30:44,210
But if you look at the
formats of the vowels,

558
00:30:44,210 --> 00:30:47,960
you can clearly tell
that they are different.

559
00:30:47,960 --> 00:30:50,420
So one hypothesis
might be that when

560
00:30:50,420 --> 00:30:52,782
you're perceiving sound
it actually has more

561
00:30:52,782 --> 00:30:54,740
to do with the formant
transitions of the vowel

562
00:30:54,740 --> 00:30:56,120
than it does--

563
00:30:56,120 --> 00:30:59,810
with the difference in
release, the volumes of the

564
00:30:59,810 --> 00:31:03,140
stops, or whatever.

565
00:31:03,140 --> 00:31:12,350
And then, if you
look at bee and dee,

566
00:31:12,350 --> 00:31:14,570
there's not quite this
obvious difference

567
00:31:14,570 --> 00:31:19,390
in the formant transitions
from the buh and the duh sound

568
00:31:19,390 --> 00:31:20,810
to the e.

569
00:31:20,810 --> 00:31:24,650
So maybe when
perceiving bee and dee

570
00:31:24,650 --> 00:31:28,820
it might have more to do
with the loudness of the stop

571
00:31:28,820 --> 00:31:32,080
release or something like that.

572
00:31:32,080 --> 00:31:33,317
Does that makes sense?

573
00:31:33,317 --> 00:31:35,150
Go like this if you
understand what's going.

574
00:31:38,100 --> 00:31:39,830
Go like this if you don't.

575
00:31:45,790 --> 00:31:49,740
I didn't see any
perplexed faces, I guess.

576
00:31:52,332 --> 00:31:53,790
So let's look at
a different sound.

577
00:32:22,340 --> 00:32:23,350
I'll just play this.

578
00:32:23,350 --> 00:32:25,382
COMPUTER: Gun, gum.

579
00:32:25,382 --> 00:32:26,840
PROFESSOR: Can you
understand that?

580
00:32:26,840 --> 00:32:29,346
Is it a little loud?

581
00:32:29,346 --> 00:32:31,126
What did it say?

582
00:32:31,126 --> 00:32:32,382
AUDIENCE: Gun, gum.

583
00:32:32,382 --> 00:32:33,215
PROFESSOR: Gun, gum.

584
00:32:33,215 --> 00:32:33,913
Right.

585
00:32:33,913 --> 00:32:34,669
Gun, gum.

586
00:32:34,669 --> 00:32:35,460
COMPUTER: Gun, gum.

587
00:32:41,970 --> 00:32:46,150
PROFESSOR: Let's look at the
guh part for now, try and review

588
00:32:46,150 --> 00:32:47,636
what we learned.

589
00:32:47,636 --> 00:32:49,548
COMPUTER: Guh, guh.

590
00:32:49,548 --> 00:32:53,590
PROFESSOR: That's a
little bit zoomed in.

591
00:32:53,590 --> 00:32:56,340
So what's going on with
the-- this is gun, I guess.

592
00:32:56,340 --> 00:32:58,600
COMPUTER: Gun, gun.

593
00:32:58,600 --> 00:33:00,680
PROFESSOR: What's
going on with gun here?

594
00:33:00,680 --> 00:33:10,160
Can anyone summarize
this situation?

595
00:33:10,160 --> 00:33:11,530
Here, I'll give you a hint.

596
00:33:11,530 --> 00:33:13,980
I think the guh is
probably this much of it.

597
00:33:18,880 --> 00:33:20,306
COMPUTER: [SOUND SIGNAL]

598
00:33:20,306 --> 00:33:21,180
PROFESSOR: And then--

599
00:33:21,180 --> 00:33:23,372
COMPUTER: Gun, gun, gun.

600
00:33:26,730 --> 00:33:28,620
PROFESSOR: So what do
you think is going on?

601
00:33:43,350 --> 00:33:49,242
AUDIENCE: [INAUDIBLE] awe,
except the middle part of it,

602
00:33:49,242 --> 00:33:53,670
[INAUDIBLE] an un,
so that's part of it.

603
00:33:53,670 --> 00:33:55,130
PROFESSOR: Yeah, the n.

604
00:33:55,130 --> 00:33:57,463
Probably this--

605
00:33:57,463 --> 00:33:59,612
COMPUTER: un, un, un.

606
00:33:59,612 --> 00:34:00,820
PROFESSOR: Can you hear that?

607
00:34:00,820 --> 00:34:02,876
That's an un.

608
00:34:02,876 --> 00:34:05,330
COMPUTER: Un, un.

609
00:34:05,330 --> 00:34:10,670
PROFESSOR: So what is happening
with the formant transitions

610
00:34:10,670 --> 00:34:12,560
for just the guh part?

611
00:34:27,796 --> 00:34:29,352
Just the beginning.

612
00:34:40,668 --> 00:34:47,210
AUDIENCE: I think
one [INAUDIBLE]

613
00:34:47,210 --> 00:34:48,139
PROFESSOR: Right.

614
00:34:48,139 --> 00:34:49,550
Oh, yeah.

615
00:34:49,550 --> 00:34:52,340
These dots at the
beginning are trying

616
00:34:52,340 --> 00:34:56,520
to find formants in the guh,
and may not be successful.

617
00:34:56,520 --> 00:34:58,610
So sorry, I should have
mentioned probably,

618
00:34:58,610 --> 00:35:00,410
you can pretty much
ignore anything

619
00:35:00,410 --> 00:35:04,310
before about this line.

620
00:35:04,310 --> 00:35:07,660
Any of this stuff is probably--
here, we'll just turn this off.

621
00:35:11,440 --> 00:35:15,550
So the formant one has
this distinctive rise

622
00:35:15,550 --> 00:35:16,654
at the beginning.

623
00:35:16,654 --> 00:35:17,570
How about formant two?

624
00:35:29,920 --> 00:35:33,380
AUDIENCE: [INAUDIBLE]

625
00:35:33,380 --> 00:35:34,530
PROFESSOR: Yeah.

626
00:35:34,530 --> 00:35:35,924
What happens at the beginning?

627
00:35:39,382 --> 00:35:41,358
This is formant two, right?

628
00:35:45,284 --> 00:35:46,367
AUDIENCE: It's decreasing.

629
00:35:46,367 --> 00:35:47,182
It's going down.

630
00:35:47,182 --> 00:35:47,890
PROFESSOR: Right.

631
00:35:47,890 --> 00:35:51,470
So it seems to start high
and dip a little bit.

632
00:35:51,470 --> 00:35:54,680
And formant one, starts
low and rises a little bit,

633
00:35:54,680 --> 00:35:57,520
just in the beginning.

634
00:35:57,520 --> 00:36:01,370
Let's just, for kicks, see
if it's the same over here.

635
00:36:01,370 --> 00:36:06,050
This is also, if you remember,
also starts with a guh,

636
00:36:06,050 --> 00:36:08,910
so probably should be similar.

637
00:36:08,910 --> 00:36:11,210
What do we see?

638
00:36:11,210 --> 00:36:12,430
Similar?

639
00:36:12,430 --> 00:36:15,691
Go like this if you
think it's similar.

640
00:36:15,691 --> 00:36:16,190
Yeah.

641
00:36:16,190 --> 00:36:18,960
So a little bit of a rise here,
a little bit of a dip here.

642
00:36:18,960 --> 00:36:23,300
And formant three seems
to be having also a rise.

643
00:36:23,300 --> 00:36:27,290
So that's kind of interesting.

644
00:36:27,290 --> 00:36:29,120
So that seems to
be what's going on.

645
00:36:29,120 --> 00:36:33,500
Now let's look at
the other transition,

646
00:36:33,500 --> 00:36:37,260
the one that's going
into the n or the m.

647
00:36:42,064 --> 00:36:43,920
What's happening here?

648
00:36:43,920 --> 00:36:45,410
Can you you guys see this?

649
00:36:45,410 --> 00:36:49,630
So this dark part is the vowel.

650
00:36:49,630 --> 00:36:52,150
And then, you can see it
corresponds with loudness.

651
00:36:52,150 --> 00:36:53,470
It's very loud.

652
00:36:53,470 --> 00:36:56,770
And then this quieter
part is the n sound.

653
00:36:56,770 --> 00:36:59,570
So the line is probably
somewhere around here.

654
00:36:59,570 --> 00:37:01,750
COMPUTER: Guh, n, n, n.

655
00:37:05,171 --> 00:37:07,420
PROFESSOR: So what is going
on with the formants here.

656
00:37:07,420 --> 00:37:08,461
We can turn that back on.

657
00:37:16,200 --> 00:37:18,300
AUDIENCE: They
get a lot quieter.

658
00:37:18,300 --> 00:37:20,050
PROFESSOR: They get
quieter for sure.

659
00:37:20,050 --> 00:37:24,253
AUDIENCE: Looks like formant
one drops off almost to zero.

660
00:37:24,253 --> 00:37:26,440
PROFESSOR: Yeah,
formant one disappears.

661
00:37:31,520 --> 00:37:35,960
And then, formant two
is pretty much steady.

662
00:37:35,960 --> 00:37:36,680
Let's check out--

663
00:37:36,680 --> 00:37:37,900
AUDIENCE: [INAUDIBLE] formant

664
00:37:37,900 --> 00:37:42,020
PROFESSOR: Formant one is
going to be the bottom formant.

665
00:37:42,020 --> 00:37:43,860
You see this?

666
00:37:43,860 --> 00:37:45,380
So they count from
the bottom up.

667
00:37:45,380 --> 00:37:50,430
So this is the first formant,
second formant, third formant.

668
00:37:50,430 --> 00:37:55,820
The formants are these
big dark smudges we see.

669
00:37:55,820 --> 00:37:58,970
And also, the red dots
are trying to track them.

670
00:37:58,970 --> 00:38:02,150
So that's the program trying to
find formants for you, so you

671
00:38:02,150 --> 00:38:05,130
can measure them or whatever.

672
00:38:05,130 --> 00:38:06,380
Let's check out the other one.

673
00:38:06,380 --> 00:38:07,310
So that was gun.

674
00:38:11,330 --> 00:38:13,060
How about this?

675
00:38:13,060 --> 00:38:14,476
What's happening?

676
00:38:14,476 --> 00:38:16,428
AUDIENCE: Same thing.

677
00:38:16,428 --> 00:38:25,742
It's dropping off [INAUDIBLE]

678
00:38:25,742 --> 00:38:26,450
PROFESSOR: Right.

679
00:38:26,450 --> 00:38:28,521
So it's getting quieter.

680
00:38:28,521 --> 00:38:30,686
AUDIENCE: Also,
formant two jumps up

681
00:38:30,686 --> 00:38:32,572
a bit before it comes steady.

682
00:38:32,572 --> 00:38:33,280
PROFESSOR: Right.

683
00:38:33,280 --> 00:38:35,840
So formant two seems
to, rather than

684
00:38:35,840 --> 00:38:40,130
have this straight across
going from the uh sound

685
00:38:40,130 --> 00:38:44,120
to the n sound, or
going from the uh sound

686
00:38:44,120 --> 00:38:47,061
to the n sound in
gun, we saw that it

687
00:38:47,061 --> 00:38:49,310
was more or less-- formant
two was more or less steady

688
00:38:49,310 --> 00:38:50,900
all the way across.

689
00:38:50,900 --> 00:38:52,730
But this one, it's more--

690
00:38:52,730 --> 00:38:55,580
it dips and then rises.

691
00:38:55,580 --> 00:38:57,290
Well, the dip is
probably just a vowel.

692
00:38:57,290 --> 00:39:00,950
But it seems to rise
up abruptly for the n.

693
00:39:00,950 --> 00:39:02,788
So if we listen--

694
00:39:02,788 --> 00:39:06,180
COMPUTER: M, m, m.

695
00:39:06,180 --> 00:39:08,540
PROFESSOR: You can hear
that it's an m sound.

696
00:39:08,540 --> 00:39:11,430
You can hear that
it's mm and not nn.

697
00:39:11,430 --> 00:39:12,140
Yeah?

698
00:39:12,140 --> 00:39:14,270
You agree?

699
00:39:14,270 --> 00:39:19,180
And it seems like the main
difference between mm and nn

700
00:39:19,180 --> 00:39:23,420
is where the second formant is.

701
00:39:23,420 --> 00:39:25,940
Assuming that the rest of the
world is the same, which I I

702
00:39:25,940 --> 00:39:29,360
think we can pretty
much safely assume that.

703
00:39:29,360 --> 00:39:30,846
Where did my mouse go?

704
00:39:30,846 --> 00:39:31,640
Down here?

705
00:39:31,640 --> 00:39:32,550
Oh, there it is.

706
00:39:35,781 --> 00:39:36,780
We can even see it here.

707
00:39:36,780 --> 00:39:44,510
So pretty steady transition
here, whereas sharp rise here.

708
00:39:44,510 --> 00:39:45,050
Make sense?

709
00:39:49,580 --> 00:39:52,640
We could hypothesize
that the difference

710
00:39:52,640 --> 00:40:00,800
between when you're perceiving
gum and gun is whether the--

711
00:40:00,800 --> 00:40:04,340
is going to be the
formant transitions.

712
00:40:04,340 --> 00:40:08,330
Or we could hypothesize that
it's the actual sound of the mm

713
00:40:08,330 --> 00:40:11,000
and nn.

714
00:40:11,000 --> 00:40:16,040
You could hear the
difference, right?

715
00:40:16,040 --> 00:40:18,440
When I didn't play
the vowel, you

716
00:40:18,440 --> 00:40:21,500
could hear the difference
between mm and nn, right?

717
00:40:21,500 --> 00:40:24,696
Let's just try again.

718
00:40:24,696 --> 00:40:31,062
COMPUTER: Nn, nn, mm, mm, mm.

719
00:40:31,062 --> 00:40:32,770
AUDIENCE: [INAUDIBLE]
play the beginning.

720
00:40:32,770 --> 00:40:34,478
I think if you play
it from the beginning

721
00:40:34,478 --> 00:40:35,740
of the vowel [INAUDIBLE].

722
00:40:39,830 --> 00:40:42,680
PROFESSOR: But I think
even if you don't play--

723
00:40:42,680 --> 00:40:44,894
if you don't get the
formant transition,

724
00:40:44,894 --> 00:40:46,060
you can hear the difference.

725
00:40:46,060 --> 00:40:47,252
COMPUTER: Mm, nn.

726
00:40:50,479 --> 00:40:52,139
AUDIENCE: Unless this a placebo.

727
00:40:52,139 --> 00:40:54,534
Why don't we try it by
[INAUDIBLE] Everybody

728
00:40:54,534 --> 00:40:58,366
close our eyes [INAUDIBLE]
and then raise your right hand

729
00:40:58,366 --> 00:40:59,906
if it's nn, and
left hand it's mm.

730
00:40:59,906 --> 00:41:00,780
PROFESSOR: Good idea.

731
00:41:00,780 --> 00:41:02,090
Everyone close your eyes.

732
00:41:02,090 --> 00:41:04,742
We'll see if we can hear this.

733
00:41:04,742 --> 00:41:06,350
AUDIENCE: [INAUDIBLE]

734
00:41:06,350 --> 00:41:08,183
PROFESSOR: I'll just
say who thinks it's mm?

735
00:41:08,183 --> 00:41:11,564
Who thinks it's nn?

736
00:41:11,564 --> 00:41:12,925
I'm going to play it.

737
00:41:12,925 --> 00:41:16,422
COMPUTER: Nn, nn, nn, nn.

738
00:41:16,422 --> 00:41:19,955
PROFESSOR: Who thinks it's m?

739
00:41:19,955 --> 00:41:21,220
Who thinks it's n?

740
00:41:24,996 --> 00:41:26,620
PROFESSOR: That was
like four and five,

741
00:41:26,620 --> 00:41:28,960
so I guess we can't hear it.

742
00:41:28,960 --> 00:41:30,330
OK-- can't hear it.

743
00:41:30,330 --> 00:41:33,400
But it is maybe different.

744
00:41:33,400 --> 00:41:34,210
It was n.

745
00:41:34,210 --> 00:41:35,560
It was an n sound.

746
00:41:35,560 --> 00:41:39,040
Congrats to the four people
who thought it was n--

747
00:41:39,040 --> 00:41:40,760
half of the class.

748
00:41:40,760 --> 00:41:44,650
So maybe not-- maybe we
can't really hear it.

749
00:41:44,650 --> 00:41:48,400
So if we can't really hear the
difference between mm and nn,

750
00:41:48,400 --> 00:41:50,800
how can we tell the difference
between gum and gun?

751
00:41:59,710 --> 00:42:02,185
AUDIENCE: [INAUDIBLE] Context.

752
00:42:08,055 --> 00:42:09,693
AUDIENCE: But she
just said both and it

753
00:42:09,693 --> 00:42:10,942
wasn't in any [INAUDIBLE]

754
00:42:10,942 --> 00:42:11,650
PROFESSOR: Right.

755
00:42:11,650 --> 00:42:13,108
I think I can say
it out of nowhere

756
00:42:13,108 --> 00:42:18,220
and you can probably tell,
even if you can't see my face.

757
00:42:18,220 --> 00:42:22,310
Gum, gun, gun, gum, gun, gun.

758
00:42:22,310 --> 00:42:22,810
Yeah?

759
00:42:22,810 --> 00:42:23,434
Maybe you can--

760
00:42:23,434 --> 00:42:25,830
AUDIENCE: [INAUDIBLE]

761
00:42:25,830 --> 00:42:27,710
PROFESSOR: Yeah.

762
00:42:27,710 --> 00:42:29,410
Or you could-- sometimes, like--

763
00:42:40,140 --> 00:42:46,100
So if it's not the sound
of the m or n sound,

764
00:42:46,100 --> 00:42:49,740
then how can we
tell the difference?

765
00:42:49,740 --> 00:42:54,610
I just played it for you
just without the transitions.

766
00:42:54,610 --> 00:42:56,430
COMPUTER: Mm, nn.

767
00:42:56,430 --> 00:42:59,000
PROFESSOR: And you couldn't
really tell the difference.

768
00:42:59,000 --> 00:43:05,620
But what if I played
you the whole thing?

769
00:43:08,760 --> 00:43:09,804
COMPUTER: Gun.

770
00:43:09,804 --> 00:43:10,470
PROFESSOR: Yeah?

771
00:43:10,470 --> 00:43:11,922
AUDIENCE: Is it
the actual change

772
00:43:11,922 --> 00:43:13,684
from the vowel to the consonant?

773
00:43:13,684 --> 00:43:14,350
PROFESSOR: Yeah.

774
00:43:14,350 --> 00:43:18,270
So it's probably the change
from the vowel to the consonant,

775
00:43:18,270 --> 00:43:22,540
because this is pretty much
steady, and this is different.

776
00:43:22,540 --> 00:43:24,820
So it's probably like
our baw, daw example

777
00:43:24,820 --> 00:43:27,640
where the formant transitions
were different, really

778
00:43:27,640 --> 00:43:28,550
different.

779
00:43:28,550 --> 00:43:36,670
And in comparison, the
actual stop release

780
00:43:36,670 --> 00:43:39,760
was not such a
drastic difference.

781
00:43:39,760 --> 00:43:43,270
So we might hypothesize that
it's the formant transitions

782
00:43:43,270 --> 00:43:44,880
that we can tell.

783
00:43:44,880 --> 00:43:48,221
So if this is our hypothesis,
how would we test it?

784
00:43:48,221 --> 00:43:48,970
What do you think?

785
00:43:58,630 --> 00:44:02,510
AUDIENCE: You mean by
different [INAUDIBLE]

786
00:44:02,510 --> 00:44:03,820
PROFESSOR: Yeah.

787
00:44:03,820 --> 00:44:05,521
AUDIENCE: We can
start off how we just

788
00:44:05,521 --> 00:44:08,407
did with our eyes closed
and half of us [INAUDIBLE]

789
00:44:08,407 --> 00:44:12,255
if you started it from an
earlier point where the

790
00:44:12,255 --> 00:44:14,756
transitions [INAUDIBLE]
and see how many people

791
00:44:14,756 --> 00:44:16,120
could guess what it was.

792
00:44:16,120 --> 00:44:17,270
PROFESSOR: Yeah, exactly.

793
00:44:17,270 --> 00:44:21,130
So maybe we could play smaller
parts of the speech signal

794
00:44:21,130 --> 00:44:25,810
and see what parts
of speech signal

795
00:44:25,810 --> 00:44:30,400
you actually have to hear in
order to hear the difference.

796
00:44:33,910 --> 00:44:36,970
I will point out, by the way,
that when I first brought this

797
00:44:36,970 --> 00:44:38,770
up, all I did was play
it and I asked you

798
00:44:38,770 --> 00:44:41,610
what it said-- everybody
knew, even though it

799
00:44:41,610 --> 00:44:44,940
was completely out of context.

800
00:44:44,940 --> 00:44:47,380
So maybe if we
chop up the speech

801
00:44:47,380 --> 00:44:53,380
signal in strategic
ways, and then play it,

802
00:44:53,380 --> 00:44:56,560
we can maybe hear--

803
00:44:56,560 --> 00:45:00,280
statistically figure out what
people are actually hearing.

804
00:45:00,280 --> 00:45:02,080
So why don't we do
something like that?

805
00:45:02,080 --> 00:45:15,320
So we could, for example, if
we just play the guh part.

806
00:45:19,630 --> 00:45:22,130
COMPUTER: Guh, guh, guh.

807
00:45:25,130 --> 00:45:26,130
AUDIENCE: [INAUDIBLE]

808
00:45:26,130 --> 00:45:29,470
COMPUTER: Gum, gum, gum.

809
00:45:29,470 --> 00:45:32,876
PROFESSOR: Yeah just
through the transition.

810
00:45:32,876 --> 00:45:37,866
COMPUTER: Gum, gum, gum.

811
00:45:37,866 --> 00:45:40,361
Gun, gun.

812
00:45:40,361 --> 00:45:44,540
PROFESSOR: Can you
hear the difference?

813
00:45:44,540 --> 00:45:47,380
So it seems to me more
this very beginning

814
00:45:47,380 --> 00:45:51,460
part of the nasal or the
end part of the vowel

815
00:45:51,460 --> 00:45:55,550
is doing more for us.

816
00:45:55,550 --> 00:45:57,350
Maybe we could try
just the end of it.

817
00:45:57,350 --> 00:45:58,058
COMPUTER: Un, un.

818
00:46:01,930 --> 00:46:04,724
Um, um, um.

819
00:46:04,724 --> 00:46:06,640
PROFESSOR: Could you
hear the difference then?

820
00:46:06,640 --> 00:46:07,360
AUDIENCE: Yes.

821
00:46:07,360 --> 00:46:08,980
PROFESSOR: Yeah,
much clearer, right?

822
00:46:08,980 --> 00:46:10,280
Much clearer.

823
00:46:10,280 --> 00:46:10,930
You agree?

824
00:46:10,930 --> 00:46:12,626
Go like this if you agree.

825
00:46:12,626 --> 00:46:14,137
Go like this if you don't agree.

826
00:46:16,880 --> 00:46:18,122
Everyone agrees.

827
00:46:18,122 --> 00:46:18,830
That's very nice.

828
00:46:21,830 --> 00:46:28,260
Or let's say that we
were to zoom in so we

829
00:46:28,260 --> 00:46:30,780
can do some precision work.

830
00:46:36,429 --> 00:46:37,470
Oh, this is gum, I think.

831
00:46:48,220 --> 00:46:52,400
If we were to just
stick the m in here--

832
00:47:00,460 --> 00:47:02,530
I included the
transition here sort of.

833
00:47:02,530 --> 00:47:05,380
You see, it might sound weird.

834
00:47:05,380 --> 00:47:07,000
Let's see what it sounds like.

835
00:47:09,700 --> 00:47:11,890
COMPUTER: Gun, gun, gun.

836
00:47:15,250 --> 00:47:18,398
PROFESSOR: It sounded
like gun, I think.

837
00:47:21,320 --> 00:47:31,812
Let's try-- that's
not what I want to do.

838
00:47:31,812 --> 00:47:32,640
Can I undo again?

839
00:47:35,540 --> 00:47:37,600
No.

840
00:47:37,600 --> 00:47:39,980
Sorry.

841
00:47:39,980 --> 00:47:42,148
Where am I?

842
00:47:42,148 --> 00:47:43,636
[INAUDIBLE] at the end.

843
00:47:51,076 --> 00:47:52,580
Is it back?

844
00:47:52,580 --> 00:47:54,430
Oh, no, it looks scary.

845
00:47:54,430 --> 00:47:55,070
Hold on.

846
00:47:58,046 --> 00:48:00,030
let me fix this.

847
00:48:26,800 --> 00:48:30,790
Let's try and grab this end bit.

848
00:48:37,412 --> 00:48:39,270
We should copy.

849
00:48:39,270 --> 00:48:40,916
That sounds good.

850
00:48:40,916 --> 00:48:44,620
We'll stick it here,
make sure that we get--

851
00:48:44,620 --> 00:48:46,830
we're moving over the
transition as well.

852
00:48:53,200 --> 00:48:59,290
COMPUTER: Gum, gun, gun, gum.

853
00:48:59,290 --> 00:49:01,000
PROFESSOR: What did I do?

854
00:49:01,000 --> 00:49:03,760
AUDIENCE: You changed
it from gun to gum.

855
00:49:03,760 --> 00:49:04,520
PROFESSOR: Yeah.

856
00:49:04,520 --> 00:49:05,020
So--

857
00:49:05,020 --> 00:49:07,690
COMPUTER: Gun, gum, gum.

858
00:49:07,690 --> 00:49:09,370
PROFESSOR: So it
seems that if I make

859
00:49:09,370 --> 00:49:13,660
very careful to include
the end bit of a vowel,

860
00:49:13,660 --> 00:49:18,850
then I can successfully
change the word.

861
00:49:18,850 --> 00:49:20,230
So to review, what
I did here was

862
00:49:20,230 --> 00:49:23,682
I took the m, including,
making sure to include

863
00:49:23,682 --> 00:49:24,640
the formant transition.

864
00:49:24,640 --> 00:49:26,860
So I was including
the end of the vowel,

865
00:49:26,860 --> 00:49:30,910
and I successfully
changed the word to gum.

866
00:49:30,910 --> 00:49:34,660
So let's try again,
except this time,

867
00:49:34,660 --> 00:49:37,645
I'll not include the
end bit of the vowel.

868
00:49:37,645 --> 00:49:39,399
That's actually what
I did the first time.

869
00:49:51,320 --> 00:49:53,530
Now it's here.

870
00:49:53,530 --> 00:49:56,307
COMPUTER: Gun, gun, gun.

871
00:49:56,307 --> 00:49:57,390
PROFESSOR: Now what is it?

872
00:49:57,390 --> 00:49:58,590
AUDIENCE: [INAUDIBLE]

873
00:49:58,590 --> 00:50:02,380
PROFESSOR: Now it's gun, right?

874
00:50:02,380 --> 00:50:04,340
So this is very interesting.

875
00:50:04,340 --> 00:50:08,840
So what this means is
that it must be something

876
00:50:08,840 --> 00:50:13,220
about the formant transitions
that we're hearing, because--

877
00:50:13,220 --> 00:50:19,520
that is the most
salient auditory cue.

878
00:50:19,520 --> 00:50:22,790
Because if I included
the formant transitions,

879
00:50:22,790 --> 00:50:27,080
I could successfully take
the n and paste it onto

880
00:50:27,080 --> 00:50:28,690
and change the word.

881
00:50:28,690 --> 00:50:31,850
But if I didn't include
the formant transitions,

882
00:50:31,850 --> 00:50:33,140
I couldn't.

883
00:50:33,140 --> 00:50:34,795
I didn't change the word.

884
00:50:34,795 --> 00:50:36,420
People following
what's happening here?

885
00:50:39,770 --> 00:50:43,850
So this experiment
suggests that it

886
00:50:43,850 --> 00:50:47,540
must be something about the
formant transitions that's

887
00:50:47,540 --> 00:50:50,040
making us hear the difference.

888
00:50:50,040 --> 00:50:51,800
And there may be other cues.

889
00:50:51,800 --> 00:50:54,350
Certainly there are other
cues, like m and then

890
00:50:54,350 --> 00:50:56,460
n itself look different.

891
00:50:56,460 --> 00:50:57,850
Let me just undo this.

892
00:51:01,420 --> 00:51:04,240
A little bit.

893
00:51:04,240 --> 00:51:05,980
But for one thing,
the m and the n parts

894
00:51:05,980 --> 00:51:09,000
are significantly
quieter than the vowels.

895
00:51:09,000 --> 00:51:10,060
See, here is the vowel--

896
00:51:10,060 --> 00:51:11,470
much louder.

897
00:51:11,470 --> 00:51:13,420
Here's the nasal-- not loud.

898
00:51:16,060 --> 00:51:19,000
And for another thing,
they're not that different.

899
00:51:19,000 --> 00:51:23,320
It must be-- so it
makes sense that it's

900
00:51:23,320 --> 00:51:25,194
the formant transitions
that we listen to.

901
00:51:25,194 --> 00:51:27,967
AUDIENCE: Do you
call it [INAUDIBLE]

902
00:51:27,967 --> 00:51:29,800
PROFESSOR: Yeah, sorry,
the nasal consonant.

903
00:51:29,800 --> 00:51:30,300
Right.

904
00:51:33,640 --> 00:51:36,382
The consonants are much
quieter than the vowels.

905
00:51:36,382 --> 00:51:38,590
So it makes sense that what
we're really listening to

906
00:51:38,590 --> 00:51:39,250
is the vowel.

907
00:51:43,110 --> 00:51:44,609
Let me play something
else for you.

908
00:52:10,557 --> 00:52:12,793
COMPUTER: Lose, lose.

909
00:52:12,793 --> 00:52:13,918
PROFESSOR: What's the word?

910
00:52:13,918 --> 00:52:15,630
AUDIENCE: Lose.

911
00:52:15,630 --> 00:52:19,330
PROFESSOR: Lose, like
lose a game, right?

912
00:52:19,330 --> 00:52:21,120
Something like that?

913
00:52:21,120 --> 00:52:23,591
AUDIENCE: [INAUDIBLE]

914
00:52:23,591 --> 00:52:24,840
PROFESSOR: This guy's British.

915
00:52:24,840 --> 00:52:27,560
AUDIENCE: Oh, yeah.

916
00:52:27,560 --> 00:52:28,534
That makes sense.

917
00:52:28,534 --> 00:52:30,092
COMPUTER: Lose.

918
00:52:30,092 --> 00:52:31,300
PROFESSOR: He's very British.

919
00:52:31,300 --> 00:52:32,962
AUDIENCE: Lose.

920
00:52:32,962 --> 00:52:33,920
PROFESSOR: Lose-- yeah.

921
00:52:38,200 --> 00:52:40,570
It's not quite the
vowel we use, right?

922
00:52:40,570 --> 00:52:42,054
It's not quite the
American vowel.

923
00:52:42,054 --> 00:52:43,720
There's something
about the vowel that's

924
00:52:43,720 --> 00:52:45,284
a little bit
different, and that's

925
00:52:45,284 --> 00:52:46,450
what makes it sound British.

926
00:52:46,450 --> 00:52:48,730
So let me break it down for you.

927
00:52:48,730 --> 00:52:52,150
Here is the l
part, the luh part.

928
00:52:52,150 --> 00:52:54,370
COMPUTER: Luh, luh.

929
00:52:54,370 --> 00:52:56,946
PROFESSOR: Well, maybe I
got a little bit of oo.

930
00:52:56,946 --> 00:52:59,170
COMPUTER: Loo, loo.

931
00:52:59,170 --> 00:53:01,810
PROFESSOR: And then
we have the ooh.

932
00:53:01,810 --> 00:53:02,746
See this.

933
00:53:02,746 --> 00:53:04,450
COMPUTER: Ooh, ooh.

934
00:53:04,450 --> 00:53:07,342
PROFESSOR: And then
here is the zuh.

935
00:53:07,342 --> 00:53:11,530
COMPUTER: [SOUND SIGNAL]

936
00:53:11,530 --> 00:53:13,720
PROFESSOR: So let
me ask you, what

937
00:53:13,720 --> 00:53:18,790
is the difference between
between the word lose

938
00:53:18,790 --> 00:53:21,770
and the word loose?

939
00:53:21,770 --> 00:53:22,270
Yeah?

940
00:53:25,150 --> 00:53:28,520
AUDIENCE: The [INAUDIBLE]
consonant in his voice.

941
00:53:28,520 --> 00:53:29,805
PROFESSOR: In which one?

942
00:53:29,805 --> 00:53:30,695
AUDIENCE: In lose.

943
00:53:30,695 --> 00:53:31,778
PROFESSOR: In lose, right.

944
00:53:31,778 --> 00:53:36,380
So in lose, it's a Z sound.

945
00:53:36,380 --> 00:53:39,079
Lose-- I'm going
to write them down.

946
00:53:39,079 --> 00:53:39,620
Make sure I--

947
00:53:48,106 --> 00:53:49,900
So this a Z sound.

948
00:53:49,900 --> 00:53:55,370
And this is an S.
People understand?

949
00:53:55,370 --> 00:53:56,440
You get this?

950
00:53:56,440 --> 00:53:57,760
This is a Z-- lose.

951
00:53:57,760 --> 00:53:59,350
This is an S-- loose.

952
00:53:59,350 --> 00:53:59,850
Yeah?

953
00:54:04,390 --> 00:54:07,620
So this is lose, so it should
have a Z sound at the end here.

954
00:54:11,157 --> 00:54:14,930
COMPUTER: [SOUND SIGNAL]

955
00:54:14,930 --> 00:54:17,580
PROFESSOR: What
does it sound like?

956
00:54:17,580 --> 00:54:19,890
It sounds like an s, right?

957
00:54:19,890 --> 00:54:22,860
COMPUTER: [SOUND SIGNAL]

958
00:54:22,860 --> 00:54:25,110
PROFESSOR: So in
theory, this is voice,

959
00:54:25,110 --> 00:54:29,550
but in actuality kind of not.

960
00:54:29,550 --> 00:54:30,570
If we look at--

961
00:54:30,570 --> 00:54:32,730
so the characteristic
of voicing is

962
00:54:32,730 --> 00:54:37,050
this striations, these up and
down stripes, in the bottom

963
00:54:37,050 --> 00:54:38,030
here.

964
00:54:38,030 --> 00:54:41,820
If we look, they
peter out really fast.

965
00:54:41,820 --> 00:54:45,330
So for the zuh
part of this, it's

966
00:54:45,330 --> 00:54:49,110
only really a Z for this much
time, and then it is all s.

967
00:54:52,530 --> 00:54:54,690
Furthermore, the
furcation noise--

968
00:54:54,690 --> 00:54:57,240
that's the fuzziness
that happens

969
00:54:57,240 --> 00:54:59,530
when you say a Z or an S--

970
00:54:59,530 --> 00:55:03,460
is really loud, compared
to what's going on here.

971
00:55:03,460 --> 00:55:08,786
This is not very loud, and
this is kind of a big deal.

972
00:55:08,786 --> 00:55:10,680
Does this makes sense?

973
00:55:10,680 --> 00:55:19,620
So I ask you, if it's not
really a zuh that's happening,

974
00:55:19,620 --> 00:55:21,420
if it's really more like an S--

975
00:55:21,420 --> 00:55:23,670
This is s sound--

976
00:55:23,670 --> 00:55:30,694
then how can we tell that
this is lose and not loose?

977
00:55:39,143 --> 00:55:42,840
AUDIENCE: [INAUDIBLE]

978
00:55:42,840 --> 00:55:46,820
PROFESSOR: Sure, maybe it's
just the beginning voicing.

979
00:55:46,820 --> 00:55:47,435
Yeah?

980
00:55:47,435 --> 00:55:49,710
AUDIENCE: Does it have
to do with the volume--

981
00:55:49,710 --> 00:55:51,080
like what happens after?

982
00:55:51,080 --> 00:55:52,005
PROFESSOR: After what?

983
00:55:52,005 --> 00:55:57,055
AUDIENCE: After the transition
from the ooh to the suh.

984
00:55:57,055 --> 00:55:59,430
PROFESSOR: Yeah, maybe it's
something about the furcation

985
00:55:59,430 --> 00:56:01,520
noise.

986
00:56:01,520 --> 00:56:04,800
What it probably is,
and I don't really

987
00:56:04,800 --> 00:56:06,690
expect you guys to
have known this,

988
00:56:06,690 --> 00:56:12,540
but the actual time that
it takes to say the vowel,

989
00:56:12,540 --> 00:56:16,920
the actual length the vowel
is going to be much longer in

990
00:56:16,920 --> 00:56:18,930
lose than in loose.

991
00:56:18,930 --> 00:56:20,250
If and you actually--

992
00:56:20,250 --> 00:56:21,420
I don't have loose.

993
00:56:21,420 --> 00:56:22,430
I really should.

994
00:56:22,430 --> 00:56:25,590
If you actually look
at lose versus loose,

995
00:56:25,590 --> 00:56:27,586
lose is going to be--

996
00:56:27,586 --> 00:56:28,350
are you OK?

997
00:56:31,624 --> 00:56:32,832
AUDIENCE: What are you doing?

998
00:56:32,832 --> 00:56:35,170
AUDIENCE: [INAUDIBLE]

999
00:56:35,170 --> 00:56:35,814
PROFESSOR: Oh.

1000
00:56:35,814 --> 00:56:37,266
AUDIENCE: Never mind.

1001
00:56:37,266 --> 00:56:46,181
[INAUDIBLE]

1002
00:56:46,181 --> 00:56:47,181
That was really strange.

1003
00:56:50,010 --> 00:56:55,520
PROFESSOR: So the vowel
in lose is actually

1004
00:56:55,520 --> 00:56:57,830
like-- it takes
twice as long to say.

1005
00:56:57,830 --> 00:57:02,060
You actually spend
more time saying it.

1006
00:57:02,060 --> 00:57:03,250
It's a long vowel.

1007
00:57:03,250 --> 00:57:04,670
Whereas in loose, it's not.

1008
00:57:04,670 --> 00:57:06,320
And I'm not going to do
this, because it's going

1009
00:57:06,320 --> 00:57:07,444
to take a really long time.

1010
00:57:07,444 --> 00:57:13,400
But if you actually take half of
the length of this vowel away,

1011
00:57:13,400 --> 00:57:18,480
and you just cut, use
edit, cut, control

1012
00:57:18,480 --> 00:57:23,150
whatever, X or whatever,
then it will actually

1013
00:57:23,150 --> 00:57:25,617
start sounding like loose.

1014
00:57:25,617 --> 00:57:27,200
The reason I'm not
going to do for you

1015
00:57:27,200 --> 00:57:31,490
is that to preserve
this beautiful arc here,

1016
00:57:31,490 --> 00:57:35,480
you actually have to take
away every other period

1017
00:57:35,480 --> 00:57:40,940
sort of thing, which
is painstaking work.

1018
00:57:40,940 --> 00:57:43,160
This is kind of
another example of--

1019
00:57:43,160 --> 00:57:46,029
what we are trying to signal is
a difference in the consonant.

1020
00:57:46,029 --> 00:57:47,570
We're trying to
signal the difference

1021
00:57:47,570 --> 00:57:50,030
between a ss and a zz.

1022
00:57:50,030 --> 00:57:52,580
And that's supposed to
be a voicing difference.

1023
00:57:52,580 --> 00:57:54,230
But in English, we don't--

1024
00:57:54,230 --> 00:57:56,510
it's not that the voicing
is what's different.

1025
00:57:56,510 --> 00:57:59,030
The voicing is hardly different.

1026
00:57:59,030 --> 00:58:01,350
It's just this
amount of different.

1027
00:58:01,350 --> 00:58:02,750
Hardly different at all.

1028
00:58:02,750 --> 00:58:07,580
But the vowel length is
really very different.

1029
00:58:07,580 --> 00:58:09,920
So this is just another
example of where

1030
00:58:09,920 --> 00:58:13,420
we're listening to the much
louder vowel to get our signal,

1031
00:58:13,420 --> 00:58:18,140
to get relevant cues, than
to the actual consonant.

1032
00:58:18,140 --> 00:58:18,783
Yeah?

1033
00:58:18,783 --> 00:58:21,550
AUDIENCE: Doesn't it have to be
something like a consonant too,

1034
00:58:21,550 --> 00:58:24,576
because if you
say two words that

1035
00:58:24,576 --> 00:58:26,780
have the exact
same length, vowel,

1036
00:58:26,780 --> 00:58:33,550
you can still the difference
between an S and a Z.

1037
00:58:33,550 --> 00:58:34,780
PROFESSOR: I mean, maybe.

1038
00:58:34,780 --> 00:58:36,445
Although-- what?

1039
00:58:36,445 --> 00:58:41,200
AUDIENCE: [INAUDIBLE] hearing
sounds [INAUDIBLE] same vowel

1040
00:58:41,200 --> 00:58:41,700
length.

1041
00:58:41,700 --> 00:58:44,910
PROFESSOR: I mean you can
say a Z with full voicing.

1042
00:58:44,910 --> 00:58:47,700
You can go zzz, if you want to.

1043
00:58:47,700 --> 00:58:51,420
It's possible to make
that distinction.

1044
00:58:51,420 --> 00:58:55,470
But, like I said, if I
were to take this signal

1045
00:58:55,470 --> 00:58:59,640
and make the duration
of the vowel shorter,

1046
00:58:59,640 --> 00:59:02,160
you would hear a loose.

1047
00:59:02,160 --> 00:59:04,470
If I were to make
this half the length,

1048
00:59:04,470 --> 00:59:06,210
then you would
actually hear loose.

1049
00:59:06,210 --> 00:59:09,420
So that suggests
that the vowel length

1050
00:59:09,420 --> 00:59:12,464
is some really relevant signal.

1051
00:59:12,464 --> 00:59:13,380
Does that makes sense?

1052
00:59:17,250 --> 00:59:18,584
That's very interesting.

1053
00:59:22,940 --> 00:59:23,908
Here's one.

1054
00:59:29,920 --> 00:59:30,420
Hello?

1055
00:59:35,914 --> 00:59:37,705
You guys don't need to
see that, I realize.

1056
00:59:42,190 --> 00:59:44,620
So before I play this whole
thing for you, let me just

1057
00:59:44,620 --> 00:59:49,456
play the n bit, and you guys
can tell me what you hear.

1058
00:59:49,456 --> 00:59:53,134
COMPUTER: A, a, a, a,

1059
00:59:53,134 --> 00:59:54,592
PROFESSOR: What
does it sound like?

1060
00:59:54,592 --> 00:59:55,980
AUDIENCE: A.

1061
00:59:55,980 --> 00:59:57,230
PROFESSOR: A, sure.

1062
00:59:57,230 --> 00:59:57,893
Yeah.

1063
00:59:57,893 --> 01:00:00,160
AUDIENCE: The consonant a.

1064
01:00:00,160 --> 01:00:04,180
PROFESSOR: Yeah, an a that comes
out of nowhere, or something.

1065
01:00:04,180 --> 01:00:05,970
It's not a or I don't know--

1066
01:00:05,970 --> 01:00:06,970
hard to do.

1067
01:00:06,970 --> 01:00:09,160
Yeah, some sort of
consonant maybe.

1068
01:00:09,160 --> 01:00:13,710
COMPUTER: A, a, a, a, a.

1069
01:00:13,710 --> 01:00:15,047
PROFESSOR: So it sounded like a.

1070
01:00:15,047 --> 01:00:15,630
AUDIENCE: Yes.

1071
01:00:19,060 --> 01:00:23,419
COMPUTER: A, a, a, a.

1072
01:00:23,419 --> 01:00:25,710
PROFESSOR: Is it sounding
different at all to you guys?

1073
01:00:25,710 --> 01:00:27,520
AUDIENCE: [INAUDIBLE]

1074
01:00:27,520 --> 01:00:28,690
PROFESSOR: More like bay.

1075
01:00:28,690 --> 01:00:30,280
AUDIENCE: [INAUDIBLE]

1076
01:00:30,280 --> 01:00:31,780
PROFESSOR: Yeah, right.

1077
01:00:31,780 --> 01:00:34,107
Sounds like-- kind of like bay.

1078
01:00:34,107 --> 01:00:35,240
AUDIENCE: Sounds like grey.

1079
01:00:35,240 --> 01:00:36,191
PROFESSOR: Grey?

1080
01:00:36,191 --> 01:00:38,532
AUDIENCE: Sounds like gay to me.

1081
01:00:38,532 --> 01:00:40,240
PROFESSOR: Gay, bay,
something like that.

1082
01:00:40,240 --> 01:00:40,740
Yeah.

1083
01:00:40,740 --> 01:00:43,520
It sounds like there's
a consonant there--

1084
01:00:43,520 --> 01:00:45,290
some consonant at the beginning.

1085
01:00:45,290 --> 01:00:46,810
How about here?

1086
01:00:46,810 --> 01:00:49,000
COMPUTER: Bay, bay, bay.

1087
01:00:49,000 --> 01:00:50,670
AUDIENCE: Sounds
like [INAUDIBLE]

1088
01:00:50,670 --> 01:00:52,902
PROFESSOR: What?

1089
01:00:52,902 --> 01:00:53,860
AUDIENCE: Bay.

1090
01:00:53,860 --> 01:00:56,618
PROFESSOR: Bay, still-- is
that what you're hearing?

1091
01:00:56,618 --> 01:00:59,546
AUDIENCE: [INAUDIBLE]
I heard a B at the end.

1092
01:00:59,546 --> 01:01:00,530
PROFESSOR: At the end?

1093
01:01:00,530 --> 01:01:01,956
COMPUTER: Bay, bay, bay.

1094
01:01:01,956 --> 01:01:03,164
AUDIENCE: It's not one vowel.

1095
01:01:03,164 --> 01:01:03,872
It's a diphthong.

1096
01:01:03,872 --> 01:01:06,330
PROFESSOR: Yes,
it's a diphthong.

1097
01:01:06,330 --> 01:01:09,620
We're focusing on
the beginning here.

1098
01:01:09,620 --> 01:01:12,342
COMPUTER: Bay, bay, bay, bay.

1099
01:01:12,342 --> 01:01:14,140
PROFESSOR: What
do you guys hear?

1100
01:01:14,140 --> 01:01:15,020
Still the same?

1101
01:01:15,020 --> 01:01:16,250
AUDIENCE: Day.

1102
01:01:16,250 --> 01:01:18,080
PROFESSOR: Day?

1103
01:01:18,080 --> 01:01:19,310
hearing day.

1104
01:01:19,310 --> 01:01:21,910
Some people are hearing bay.

1105
01:01:21,910 --> 01:01:23,910
AUDIENCE: [INAUDIBLE]

1106
01:01:23,910 --> 01:01:30,347
COMPUTER: Bay,
bay, day, day, day.

1107
01:01:30,347 --> 01:01:31,826
PROFESSOR: Still bay, day--

1108
01:01:33,305 --> 01:01:36,263
AUDIENCE: It switches
off between bay and day.

1109
01:01:36,263 --> 01:01:38,740
COMPUTER: Day, day, day, day.

1110
01:01:38,740 --> 01:01:40,434
PROFESSOR: How about now?

1111
01:01:40,434 --> 01:01:42,250
AUDIENCE: Date.

1112
01:01:42,250 --> 01:01:43,860
PROFESSOR: Day, they, date.

1113
01:01:43,860 --> 01:01:45,702
COMPUTER: Date, date, date.

1114
01:01:45,702 --> 01:01:47,910
PROFESSOR: This was actually
taken out of a sentence.

1115
01:01:47,910 --> 01:01:51,530
So it makes sense to you here
also a consonant at the end,

1116
01:01:51,530 --> 01:01:53,840
because the vowel
stops abruptly,

1117
01:01:53,840 --> 01:01:57,050
because it was just
sort of cut and pasted

1118
01:01:57,050 --> 01:01:59,495
from the middle of a sentence.

1119
01:02:02,310 --> 01:02:04,370
COMPUTER: Day, day.

1120
01:02:04,370 --> 01:02:06,640
PROFESSOR: People are
hearing mostly day here?

1121
01:02:06,640 --> 01:02:08,110
AUDIENCE: Yeah.

1122
01:02:08,110 --> 01:02:08,754
PROFESSOR: Day?

1123
01:02:08,754 --> 01:02:09,670
Let's keep going back.

1124
01:02:09,670 --> 01:02:13,630
COMPUTER: Day, day, day, day.

1125
01:02:13,630 --> 01:02:14,774
PROFESSOR: Still day?

1126
01:02:14,774 --> 01:02:17,680
AUDIENCE: [INAUDIBLE]

1127
01:02:17,680 --> 01:02:20,465
COMPUTER: Day, day, day, day.

1128
01:02:20,465 --> 01:02:22,370
PROFESSOR: Still day?

1129
01:02:22,370 --> 01:02:23,675
I'll play you the whole thing.

1130
01:02:23,675 --> 01:02:24,625
COMPUTER: Say.

1131
01:02:24,625 --> 01:02:26,198
AUDIENCE: Say.

1132
01:02:26,198 --> 01:02:28,447
AUDIENCE: [INAUDIBLE] the
last time right before that,

1133
01:02:28,447 --> 01:02:29,625
and I was like, oh.

1134
01:02:29,625 --> 01:02:31,530
COMPUTER: Say, say.

1135
01:02:31,530 --> 01:02:32,490
AUDIENCE: [INAUDIBLE]

1136
01:02:32,490 --> 01:02:35,090
PROFESSOR: Say--
anyone surprised?

1137
01:02:35,090 --> 01:02:37,473
AUDIENCE: [INAUDIBLE]

1138
01:02:37,473 --> 01:02:40,146
[INTERPOSING VOICES]

1139
01:02:40,146 --> 01:02:40,770
PROFESSOR: Huh?

1140
01:02:40,770 --> 01:02:41,220
Oh, yeah.

1141
01:02:41,220 --> 01:02:41,720
Right.

1142
01:02:41,720 --> 01:02:44,340
The name of the
file is set to yes.

1143
01:02:44,340 --> 01:02:46,728
What's going on here?

1144
01:02:46,728 --> 01:02:48,960
AUDIENCE: I heard
it right before.

1145
01:02:48,960 --> 01:02:50,620
PROFESSOR: Yeah.

1146
01:02:50,620 --> 01:02:53,140
COMPUTER: Say, say, say.

1147
01:02:53,140 --> 01:02:58,320
PROFESSOR: You can see
from this demonstration

1148
01:02:58,320 --> 01:03:00,660
how we're interpreting
the signal.

1149
01:03:00,660 --> 01:03:03,000
So back when we were
over here, you guys

1150
01:03:03,000 --> 01:03:07,290
were saying it sounds like
a or bay or something.

1151
01:03:07,290 --> 01:03:10,015
It sounds like a vowel and
it comes out of nowhere,

1152
01:03:10,015 --> 01:03:11,890
so it must have a
consonant at the beginning.

1153
01:03:11,890 --> 01:03:14,006
So your brain sort of
fills that in, like oh,

1154
01:03:14,006 --> 01:03:15,630
there must have been
a consonant there,

1155
01:03:15,630 --> 01:03:17,855
because it's a vowel
coming out of nowhere.

1156
01:03:17,855 --> 01:03:20,640
If you were just to say
a, then you would probably

1157
01:03:20,640 --> 01:03:24,030
have a sort of characteristic
rise that we've seen before,

1158
01:03:24,030 --> 01:03:26,130
sort of swell into it.

1159
01:03:26,130 --> 01:03:28,747
And then as I got farther--

1160
01:03:28,747 --> 01:03:29,330
what happened?

1161
01:03:36,720 --> 01:03:38,090
My computer is freaking out.

1162
01:03:38,090 --> 01:03:42,104
So as I got farther back,
you guys started hearing day.

1163
01:03:42,104 --> 01:03:43,853
Why do you think you
would have heard day?

1164
01:03:47,240 --> 01:03:49,330
As I got closer
and closer to here,

1165
01:03:49,330 --> 01:03:53,190
you guys started hearing
day rather than bay, right?

1166
01:03:53,190 --> 01:03:54,496
So why was that?

1167
01:03:59,809 --> 01:04:04,594
AUDIENCE: Maybe because
bay was more abrupt--

1168
01:04:04,594 --> 01:04:09,590
stop [INAUDIBLE] it's
so abrupt, and so we

1169
01:04:09,590 --> 01:04:12,360
assumed it was that until we
started hearing the ss sound.

1170
01:04:12,360 --> 01:04:13,850
PROFESSOR: Right.

1171
01:04:13,850 --> 01:04:15,362
So you guys started
hearing day--

1172
01:04:15,362 --> 01:04:17,320
I don't know when my
computer is freaking out--

1173
01:04:17,320 --> 01:04:20,220
you guys started
hearing day around here,

1174
01:04:20,220 --> 01:04:25,062
Right so maybe it was
a little less abrupt.

1175
01:04:25,062 --> 01:04:26,550
Anything else?

1176
01:04:37,958 --> 01:04:41,330
AUDIENCE: [INAUDIBLE]

1177
01:04:41,330 --> 01:04:43,980
PROFESSOR: Yeah, you didn't
hear the furcation noise,

1178
01:04:43,980 --> 01:04:47,280
is what this is called, of
the-- because S is a fricative,

1179
01:04:47,280 --> 01:04:49,260
and that's a furcation noise.

1180
01:04:49,260 --> 01:04:51,390
So you didn't hear
that happening.

1181
01:04:51,390 --> 01:04:56,370
So you weren't perceiving
S. But the question is,

1182
01:04:56,370 --> 01:04:59,790
since you weren't
perceiving S, why was it day

1183
01:04:59,790 --> 01:05:04,350
that you were perceiving, and
not something else, like bay.

1184
01:05:04,350 --> 01:05:06,405
Because you were perceiving
bay a lot in here.

1185
01:05:09,030 --> 01:05:11,370
So you suggested that
it's a little less

1186
01:05:11,370 --> 01:05:16,260
abrupt to go into there, maybe
because of the release for day

1187
01:05:16,260 --> 01:05:17,130
being louder.

1188
01:05:17,130 --> 01:05:18,666
What do you think?

1189
01:05:18,666 --> 01:05:25,911
AUDIENCE: Maybe the shape of
[INAUDIBLE] the formant changes

1190
01:05:25,911 --> 01:05:30,014
the closer you're [INAUDIBLE]

1191
01:05:30,014 --> 01:05:30,680
PROFESSOR: Yeah.

1192
01:05:30,680 --> 01:05:33,250
So if you think
about it, the way

1193
01:05:33,250 --> 01:05:36,430
your mouth is for an
S and a d, there's

1194
01:05:36,430 --> 01:05:39,424
a closure in a
very similar place.

1195
01:05:39,424 --> 01:05:44,344
AUDIENCE: Didn't we notice
before that with the d

1196
01:05:44,344 --> 01:05:49,270
that there was less
of a formant change.

1197
01:05:49,270 --> 01:05:53,140
PROFESSOR: Yeah for awe, right?

1198
01:05:53,140 --> 01:05:58,730
So we have yet to see that
for the a vowel, diphthong.

1199
01:06:02,400 --> 01:06:04,580
But if you think
about it, the way

1200
01:06:04,580 --> 01:06:10,620
your mouth is for an S and duh,
both have a closure right here.

1201
01:06:10,620 --> 01:06:13,660
Try it-- ss, duh, duh, duh--

1202
01:06:13,660 --> 01:06:18,370
very close, like behind your
top teeth, behind your teeth,

1203
01:06:18,370 --> 01:06:20,170
that's where the
closure is happening.

1204
01:06:23,260 --> 01:06:26,530
Basically, what that means
is that the different sort

1205
01:06:26,530 --> 01:06:29,455
of cavities in your vocal tract
are going to be very similar.

1206
01:06:32,930 --> 01:06:33,740
That's too bad.

1207
01:06:41,180 --> 01:06:42,940
The cavities in
your mouth are going

1208
01:06:42,940 --> 01:06:48,721
to be really similar
for day and say.

1209
01:06:48,721 --> 01:06:50,470
So the formant transitions
for day and say

1210
01:06:50,470 --> 01:06:52,460
are also going to
be really similar.

1211
01:06:52,460 --> 01:06:54,950
So the way to figure out the
difference between day and say

1212
01:06:54,950 --> 01:07:00,220
is going to be something like is
it a fricative or is it a stop?

1213
01:07:00,220 --> 01:07:03,340
So when we were starting just
at the end of the furcation

1214
01:07:03,340 --> 01:07:05,702
or the beginning of a
vowel, somewhere in there,

1215
01:07:05,702 --> 01:07:07,660
we were thinking it's a
stop, because we didn't

1216
01:07:07,660 --> 01:07:09,270
hear all that furcation noise.

1217
01:07:09,270 --> 01:07:13,210
We heard the vowel coming
very loudly out of nowhere.

1218
01:07:13,210 --> 01:07:16,570
But when we moved it back to
hear the whole furcation noise,

1219
01:07:16,570 --> 01:07:18,550
then we could hear
that it was say.

1220
01:07:18,550 --> 01:07:20,600
People are getting this?

1221
01:07:20,600 --> 01:07:21,902
Understanding this?

1222
01:07:21,902 --> 01:07:22,402
Yeah?

1223
01:07:40,944 --> 01:07:41,860
Do you guys like this?

1224
01:07:41,860 --> 01:07:44,170
I'm trying out a new background.

1225
01:07:44,170 --> 01:07:45,550
I just found this yesterday.

1226
01:07:45,550 --> 01:07:48,070
I'm not sure whether I like it.

1227
01:07:48,070 --> 01:07:50,410
AUDIENCE: [INAUDIBLE]

1228
01:07:50,410 --> 01:07:51,160
PROFESSOR: I know.

1229
01:07:51,160 --> 01:07:56,050
I kind of wish that they
were leaves all the way down,

1230
01:07:56,050 --> 01:07:57,348
and not stars.

1231
01:07:57,348 --> 01:07:59,540
AUDIENCE: Did you make it?

1232
01:07:59,540 --> 01:08:01,352
PROFESSOR: I didn't make it.

1233
01:08:01,352 --> 01:08:02,525
[INTERPOSING VOICES]

1234
01:08:02,525 --> 01:08:04,870
AUDIENCE: It's
like maple leaves.

1235
01:08:04,870 --> 01:08:08,005
PROFESSOR: I mean, this one
looks like a star, right?

1236
01:08:08,005 --> 01:08:10,430
AUDIENCE: Oh, I see it.

1237
01:08:10,430 --> 01:08:13,340
[INTERPOSING VOICES]

1238
01:08:17,763 --> 01:08:21,609
COMPUTER: Say, say.

1239
01:08:21,609 --> 01:08:23,100
PROFESSOR: What
was I going to say?

1240
01:08:23,100 --> 01:08:23,600
Oh, yeah.

1241
01:08:23,600 --> 01:08:28,229
Incidentally, a lot of these
studies are done with women--

1242
01:08:28,229 --> 01:08:31,180
sorry-- a lot of these
studies are done with men.

1243
01:08:31,180 --> 01:08:34,120
[INTERPOSING VOICES]

1244
01:08:34,120 --> 01:08:38,979
This is a sort of exceptional,
in that this is a woman.

1245
01:08:38,979 --> 01:08:40,840
It's sort of exceptional,
because the thing

1246
01:08:40,840 --> 01:08:44,100
is that, on average,
men and women speak

1247
01:08:44,100 --> 01:08:46,660
at different pitches.

1248
01:08:46,660 --> 01:08:49,450
And yes, really.

1249
01:08:53,680 --> 01:08:56,340
So those interact
with the formants.

1250
01:08:56,340 --> 01:09:00,189
So if you think of
this, the formants

1251
01:09:00,189 --> 01:09:02,890
have something to do with the
resonances in your vocal tract.

1252
01:09:02,890 --> 01:09:04,560
That's what they
have to do with.

1253
01:09:04,560 --> 01:09:06,310
And the pitch of
your voice has to do

1254
01:09:06,310 --> 01:09:09,790
with how quickly your
vocal chords are vibrating.

1255
01:09:09,790 --> 01:09:12,550
So if you think about it, these
are completely independent

1256
01:09:12,550 --> 01:09:13,810
things.

1257
01:09:13,810 --> 01:09:15,130
Do you understand this?

1258
01:09:20,512 --> 01:09:22,720
The formants have to do with
the shape of your mouth,

1259
01:09:22,720 --> 01:09:25,956
and the pitch has to do
with the speed of air

1260
01:09:25,956 --> 01:09:27,080
and all that sort of thing.

1261
01:09:27,080 --> 01:09:30,200
And these are independent.

1262
01:09:30,200 --> 01:09:34,399
But what you can see is that,
depending on the frequency--

1263
01:09:34,399 --> 01:09:38,830
So if you're producing
sound at a certain pitch,

1264
01:09:38,830 --> 01:09:44,180
if you're producing a pitch of
a vowel, then it has harmonics.

1265
01:09:44,180 --> 01:09:48,160
So if you produce something
at like 400 Hertz,

1266
01:09:48,160 --> 01:09:52,029
then it's going to be loudest
at 400, and 800, and 1200,

1267
01:09:52,029 --> 01:09:55,620
and 1600, and so on.

1268
01:09:55,620 --> 01:09:57,598
And then, in between--

1269
01:09:57,598 --> 01:10:00,530
AUDIENCE: Don't you
multiply be two each time?

1270
01:10:00,530 --> 01:10:05,330
PROFESSOR: 1200 comes
in there, I believe.

1271
01:10:10,530 --> 01:10:14,490
And in between them,
you're going to have like--

1272
01:10:14,490 --> 01:10:17,280
at 400-- sorry, 400, 800--

1273
01:10:17,280 --> 01:10:20,970
so at 600, it's going
to be less quiet.

1274
01:10:20,970 --> 01:10:25,560
So what ends up happening is
that like regardless you're

1275
01:10:25,560 --> 01:10:30,930
not going to see
much going on at 600.

1276
01:10:30,930 --> 01:10:39,420
Basically, these little
dots are indicative of where

1277
01:10:39,420 --> 01:10:40,470
those gaps are.

1278
01:10:40,470 --> 01:10:42,060
If you see the
white dots, that's

1279
01:10:42,060 --> 01:10:44,900
where you can't really hear
it because of the pitch

1280
01:10:44,900 --> 01:10:46,140
that her voice is at.

1281
01:10:46,140 --> 01:10:49,530
And for men, because
it's a lower pitch,

1282
01:10:49,530 --> 01:10:52,910
they're going to be
differently spaced.

1283
01:10:52,910 --> 01:10:58,270
So if we were to
look at Edward again,

1284
01:10:58,270 --> 01:11:00,360
which we'll do in
a second, I guess,

1285
01:11:00,360 --> 01:11:06,370
then we could see that he didn't
interfere with the formant

1286
01:11:06,370 --> 01:11:07,180
here.

1287
01:11:07,180 --> 01:11:10,510
Sometimes, you have some
issues actually finding

1288
01:11:10,510 --> 01:11:12,280
where is formant two,
or where it's not,

1289
01:11:12,280 --> 01:11:14,955
because you have this
pitch thing going on.

1290
01:11:14,955 --> 01:11:16,330
So that's why a
lot of this stuff

1291
01:11:16,330 --> 01:11:18,540
is done with men's
voices, because it

1292
01:11:18,540 --> 01:11:20,740
tends not to interfere
with the lower formants

1293
01:11:20,740 --> 01:11:23,490
when you're just
looking at the pictures.

1294
01:11:23,490 --> 01:11:24,820
That isn't really relevant.

1295
01:11:30,560 --> 01:11:33,140
I guess that is more or
less what I had to say.

1296
01:11:33,140 --> 01:11:35,540
Do you guys I have questions?

1297
01:11:40,240 --> 01:11:42,080
No questions?

1298
01:11:42,080 --> 01:11:43,308
What?

1299
01:11:43,308 --> 01:11:46,242
AUDIENCE: I see the stars.

1300
01:11:46,242 --> 01:11:51,140
[INTERPOSING VOICES]

1301
01:11:51,140 --> 01:11:54,481
PROFESSOR: Do you guys want to
see some cool Burmese stuff?

1302
01:11:54,481 --> 01:11:54,980
Burmese?

1303
01:11:54,980 --> 01:11:56,107
AUDIENCE: Yeah.

1304
01:11:56,107 --> 01:11:56,690
PROFESSOR: OK.

1305
01:11:56,690 --> 01:11:59,570
[INTERPOSING VOICES]

1306
01:11:59,570 --> 01:12:00,892
AUDIENCE: Oh, Burmese.

1307
01:12:00,892 --> 01:12:02,776
[INTERPOSING VOICES]

1308
01:12:05,470 --> 01:12:08,108
PROFESSOR: Yes,
with the Buddhists.

1309
01:12:08,108 --> 01:12:11,084
[INTERPOSING VOICES]

1310
01:14:00,020 --> 01:14:01,350
PROFESSOR: This is Burmese.

1311
01:14:01,350 --> 01:14:03,170
[INTERPOSING VOICES]

1312
01:14:09,950 --> 01:14:11,569
PROFESSOR: That was a guy, yeah.

1313
01:14:11,569 --> 01:14:12,610
AUDIENCE: That was a guy?

1314
01:14:12,610 --> 01:14:13,235
PROFESSOR: Yes.

1315
01:14:16,600 --> 01:14:19,277
COMPUTER: Ma, ma.

1316
01:14:19,277 --> 01:14:20,610
PROFESSOR: This is the same guy.

1317
01:14:20,610 --> 01:14:22,362
AUDIENCE: Oh, it's a guy.

1318
01:14:22,362 --> 01:14:24,146
COMPUTER: Ma, ma.

1319
01:14:26,790 --> 01:14:30,390
PROFESSOR: This guy--
so Burmese has tones.

1320
01:14:34,194 --> 01:14:35,110
It's a tonal language.

1321
01:14:35,110 --> 01:14:36,330
Do you guys know
what that means?

1322
01:14:36,330 --> 01:14:37,662
AUDIENCE: Is it like Chinese.

1323
01:14:37,662 --> 01:14:38,828
AUDIENCE: Japanese, Chinese.

1324
01:14:38,828 --> 01:14:40,320
PROFESSOR: It's like Chinese.

1325
01:14:40,320 --> 01:14:40,820
Yeah.

1326
01:14:40,820 --> 01:14:43,106
AUDIENCE: Japanese [INAUDIBLE]

1327
01:14:43,106 --> 01:14:44,730
PROFESSOR: Japanese
isn't really tonal.

1328
01:14:44,730 --> 01:14:46,970
AUDIENCE: [INAUDIBLE]

1329
01:14:46,970 --> 01:14:48,780
PROFESSOR: Yeah.

1330
01:14:48,780 --> 01:14:52,710
So the difference-- So what
makes it a tonal language is

1331
01:14:52,710 --> 01:14:58,130
that something about the pitch
of the vowel or the shape--

1332
01:14:58,130 --> 01:15:01,350
whether it's going up or
down, makes a distinction.

1333
01:15:01,350 --> 01:15:04,140
So like in Chinese,
there's going

1334
01:15:04,140 --> 01:15:08,460
to be difference between Ma
and ma, something like that.

1335
01:15:08,460 --> 01:15:09,786
And those are different words.

1336
01:15:09,786 --> 01:15:11,535
AUDIENCE: It's like
how you express anger.

1337
01:15:11,535 --> 01:15:13,593
because like in English,
you use your tone

1338
01:15:13,593 --> 01:15:14,860
to express differently.

1339
01:15:14,860 --> 01:15:16,222
You just yell the word loudly.

1340
01:15:16,222 --> 01:15:17,590
[INTERPOSING VOICES]

1341
01:15:24,387 --> 01:15:25,970
PROFESSOR: So this
is a good question.

1342
01:15:25,970 --> 01:15:28,780
And the answer to your--

1343
01:15:28,780 --> 01:15:31,250
is the word you should
Google is prosody.

1344
01:15:42,400 --> 01:15:46,600
Prosody is the study of
the pitches of sentences

1345
01:15:46,600 --> 01:15:50,680
and entire utterances, and
how those pitches change,

1346
01:15:50,680 --> 01:15:53,510
and how that relates
to what you're saying.

1347
01:15:53,510 --> 01:15:55,690
So for example, in
English, we can use prosody

1348
01:15:55,690 --> 01:15:58,600
to make the difference
between a declarative sentence

1349
01:15:58,600 --> 01:16:04,240
and a question, like
you ate earlier today.

1350
01:16:04,240 --> 01:16:06,930
Versus, you ate earlier today?

1351
01:16:06,930 --> 01:16:09,610
Or something like that.

1352
01:16:09,610 --> 01:16:13,050
So it's not the words
themselves, but the porosity

1353
01:16:13,050 --> 01:16:14,050
that makes a difference.

1354
01:16:14,050 --> 01:16:16,330
And porosity in Chinese
is different from porosity

1355
01:16:16,330 --> 01:16:20,580
in English is the short answer.

1356
01:16:20,580 --> 01:16:22,090
Burmese is tonal.

1357
01:16:22,090 --> 01:16:28,930
Burmese also has this weird
thing called voiceless nasals.

1358
01:16:28,930 --> 01:16:31,030
So let's just look at one.

1359
01:16:35,620 --> 01:16:39,280
If you think of a nasal
like mm in English,

1360
01:16:39,280 --> 01:16:41,830
pretty much whenever you say
it, it has voicing going on.

1361
01:16:41,830 --> 01:16:45,450
Right there-- their vocal chords
are vibrating during the mm.

1362
01:16:45,450 --> 01:16:46,810
Mm-- it's happening.

1363
01:16:46,810 --> 01:16:47,800
You can feel it--

1364
01:16:47,800 --> 01:16:50,530
mm.

1365
01:16:50,530 --> 01:16:55,375
And in Burmese, this
is true as well.

1366
01:16:55,375 --> 01:16:58,180
But they have something
called the voiceless nasal.

1367
01:16:58,180 --> 01:17:02,950
So this is supposed to be the
same word with the same tone,

1368
01:17:02,950 --> 01:17:04,960
and the difference is
that this one-- or sorry,

1369
01:17:04,960 --> 01:17:05,905
these are supposed to
be different words,

1370
01:17:05,905 --> 01:17:08,530
but they're supposed to have the
same tone and the same vowel--

1371
01:17:08,530 --> 01:17:13,190
and the difference is that
this has a voiceless m

1372
01:17:13,190 --> 01:17:17,587
versus a voiced m here.

1373
01:17:17,587 --> 01:17:19,420
So let's see if you can
hear the difference.

1374
01:17:19,420 --> 01:17:21,415
COMPUTER: Ma.

1375
01:17:21,415 --> 01:17:21,915
Muh.

1376
01:17:24,570 --> 01:17:25,320
PROFESSOR: Anyone?

1377
01:17:25,320 --> 01:17:27,680
AUDIENCE: Yeah, slightly.

1378
01:17:27,680 --> 01:17:29,976
PROFESSOR: The vowel is shorter.

1379
01:17:29,976 --> 01:17:33,830
COMPUTER: Muh, ma.

1380
01:17:33,830 --> 01:17:36,030
PROFESSOR: I mean, in
truth, they're both voiced.

1381
01:17:36,030 --> 01:17:38,420
We can see this
beautiful voicing

1382
01:17:38,420 --> 01:17:41,604
down here, and down here.

1383
01:17:41,604 --> 01:17:43,020
And maybe the
length is different.

1384
01:17:43,020 --> 01:17:43,561
I don't know.

1385
01:17:43,561 --> 01:17:44,210
Let's see.

1386
01:17:44,210 --> 01:17:50,870
The length of this one
is like 57 milliseconds,

1387
01:17:50,870 --> 01:17:54,770
and this one is like
42 milliseconds.

1388
01:17:54,770 --> 01:17:56,720
So it's maybe a
little bit shorter.

1389
01:17:56,720 --> 01:17:57,450
Yeah?

1390
01:17:57,450 --> 01:17:58,950
AUDIENCE: I think
it might be easier

1391
01:17:58,950 --> 01:18:01,400
to pronounce voiceless
nasal [INAUDIBLE] shorter.

1392
01:18:01,400 --> 01:18:04,894
It might be easier, because--

1393
01:18:04,894 --> 01:18:06,060
PROFESSOR: So phoneticians--

1394
01:18:06,060 --> 01:18:09,770
[INTERPOSING VOICES]

1395
01:18:09,770 --> 01:18:12,930
PROFESSOR: Yeah, you have to
voice it to hear it basically.

1396
01:18:12,930 --> 01:18:15,990
So phoneticians and
Burmese speakers

1397
01:18:15,990 --> 01:18:18,870
have noticed that in the
voiceless nasals, what happens

1398
01:18:18,870 --> 01:18:21,360
is that before
the nasal happens,

1399
01:18:21,360 --> 01:18:23,740
there's some air flowing
out of your mouth--

1400
01:18:23,740 --> 01:18:25,715
or sorry, flowing
out of your nose.

1401
01:18:25,715 --> 01:18:27,840
AUDIENCE: [INAUDIBLE]

1402
01:18:27,840 --> 01:18:29,640
PROFESSOR: Ma, right?

1403
01:18:29,640 --> 01:18:31,170
So ma.

1404
01:18:31,170 --> 01:18:35,040
So there's air flowing out
of the nose before the nasal.

1405
01:18:35,040 --> 01:18:40,560
But I found this
kind of unlikely.

1406
01:18:40,560 --> 01:18:43,170
It's really unlikely that the
difference that you're hearing

1407
01:18:43,170 --> 01:18:47,090
is actually the air
flowing out of your nose--

1408
01:18:47,090 --> 01:18:48,450
not very loud, right?

1409
01:18:51,300 --> 01:18:52,020
Really quiet.

1410
01:18:54,790 --> 01:19:00,140
So I actually did a
study where I took--

1411
01:19:00,140 --> 01:19:03,360
I basically did a
copy and paste study.

1412
01:19:03,360 --> 01:19:08,831
So it took everything
before the vowel

1413
01:19:08,831 --> 01:19:10,830
for two words that were
supposed to be the same,

1414
01:19:10,830 --> 01:19:13,770
except for the
voicing of the nasal,

1415
01:19:13,770 --> 01:19:16,710
and I pasted them
in the wrong way,

1416
01:19:16,710 --> 01:19:19,170
and then I also pasted
them in the right way

1417
01:19:19,170 --> 01:19:21,250
with other instances
of the same word,

1418
01:19:21,250 --> 01:19:25,410
and they always judged
the meaning of the word

1419
01:19:25,410 --> 01:19:29,370
to be the one that associated
with the correct vowel.

1420
01:19:29,370 --> 01:19:33,990
So if I took this off here,
and put on the voice one,

1421
01:19:33,990 --> 01:19:37,620
they would still
think it was unvoiced.

1422
01:19:37,620 --> 01:19:39,350
And if I took off
the voiceless one--

1423
01:19:39,350 --> 01:19:43,350
or if I put a voiceless
onto here instead of this,

1424
01:19:43,350 --> 01:19:46,020
they would still think
that it was voiced.

1425
01:19:46,020 --> 01:19:48,550
So I never figured out exactly
what the difference was,

1426
01:19:48,550 --> 01:19:50,460
but it must be something
to do with the vowel that's

1427
01:19:50,460 --> 01:19:51,376
making the difference.

1428
01:19:51,376 --> 01:19:52,640
Does that make sense?

1429
01:19:52,640 --> 01:19:56,340
This is another one of
those stereotypical things.

1430
01:19:56,340 --> 01:19:58,024
AUDIENCE: [INAUDIBLE]

1431
01:19:58,024 --> 01:19:59,440
PROFESSOR: Totally
out of context.

1432
01:19:59,440 --> 01:20:01,470
So I just played them the word.

1433
01:20:01,470 --> 01:20:04,530
I randomly mixed them up using--

1434
01:20:04,530 --> 01:20:07,440
I think, actually, I drew
numbers out of a hat--

1435
01:20:07,440 --> 01:20:08,940
named them things
that had nothing--

1436
01:20:08,940 --> 01:20:11,670
like just letters or
something-- had nothing

1437
01:20:11,670 --> 01:20:14,310
to do with what they meant,
played them for them,

1438
01:20:14,310 --> 01:20:17,440
had lots of controls, and just
said, what does this mean?

1439
01:20:17,440 --> 01:20:22,362
Does it mean to marinate or a
celestial body, or whatever,

1440
01:20:22,362 --> 01:20:24,570
because this is like the
differences that these have.

1441
01:20:24,570 --> 01:20:27,160
And they give me really
significant results.

1442
01:20:27,160 --> 01:20:29,790
So it was really cool.

1443
01:20:29,790 --> 01:20:31,730
And it has everything
to do with the vowel.

1444
01:20:31,730 --> 01:20:33,660
I hypothesized that it
might have something

1445
01:20:33,660 --> 01:20:36,090
to do with breathiness
of the vowel.

1446
01:20:36,090 --> 01:20:38,910
But that's also intertwined
with the Burmese tone system,

1447
01:20:38,910 --> 01:20:44,320
so it's definitely
not conclusive.

1448
01:20:44,320 --> 01:20:49,440
So that's an interesting
thing that Burmese has.

1449
01:20:49,440 --> 01:20:52,109
It also has tones, so I
wanted to play those for you.

1450
01:20:52,109 --> 01:20:53,400
AUDIENCE: Do you speak Burmese?

1451
01:20:53,400 --> 01:20:56,235
PROFESSOR: I do not.

1452
01:20:56,235 --> 01:21:01,525
AUDIENCE: So wasn't
it really [INAUDIBLE]

1453
01:21:01,525 --> 01:21:02,150
PROFESSOR: Hmm?

1454
01:21:02,150 --> 01:21:03,898
AUDIENCE: Wasn't it like
really difficult [INAUDIBLE]

1455
01:21:03,898 --> 01:21:05,210
experiment [INAUDIBLE]

1456
01:21:05,210 --> 01:21:07,380
PROFESSOR: I was asking
Burmese speakers.

1457
01:21:07,380 --> 01:21:08,980
AUDIENCE: Oh.

1458
01:21:08,980 --> 01:21:15,380
PROFESSOR: I wasn't doing
it with myself, obviously.

1459
01:21:15,380 --> 01:21:18,896
So here's a voiced--

1460
01:21:26,720 --> 01:21:29,050
So here is one tone.

1461
01:21:29,050 --> 01:21:30,480
COMPUTER: Ma.

1462
01:21:30,480 --> 01:21:32,210
PROFESSOR: Here's another tone.

1463
01:21:32,210 --> 01:21:34,385
COMPUTER: Mah, mah.

1464
01:21:34,385 --> 01:21:36,010
PROFESSOR: Can you
hear the difference?

1465
01:21:36,010 --> 01:21:37,326
AUDIENCE: Yeah.

1466
01:21:37,326 --> 01:21:38,700
PROFESSOR: So the
first time is--

1467
01:21:38,700 --> 01:21:42,225
COMPUTER: Ma, ma.

1468
01:21:42,225 --> 01:21:43,590
Mah.

1469
01:21:43,590 --> 01:21:48,000
PROFESSOR: So the first tone
is falling and shorter--

1470
01:21:48,000 --> 01:21:53,520
falling, shorter, creaky,
while this is kind of breathy.

1471
01:21:53,520 --> 01:21:56,580
It's not really a tone system,
because there's actually

1472
01:21:56,580 --> 01:22:00,990
not just pitches involved, but
also different stuff going on.

1473
01:22:00,990 --> 01:22:03,470
So it's kind of a crazy
tone system Burmese has.

1474
01:22:03,470 --> 01:22:06,090
And it has two more tones
in addition to this.

1475
01:22:06,090 --> 01:22:07,880
And it also has--

1476
01:22:07,880 --> 01:22:10,930
some syllables have
no tone, somehow.

1477
01:22:10,930 --> 01:22:13,810
But they are special syllables.

1478
01:22:13,810 --> 01:22:15,300
So it's pretty complicated.

1479
01:22:19,560 --> 01:22:22,440
So for example--
maybe I'll play you

1480
01:22:22,440 --> 01:22:25,440
one more thing since we have
two minutes or zero minutes.

1481
01:22:53,870 --> 01:22:58,199
COMPUTER: Ma, ma, mah.

1482
01:22:58,199 --> 01:22:59,607
PROFESSOR: Those are different.

1483
01:23:02,740 --> 01:23:07,800
And I think this
doesn't always go--

1484
01:23:07,800 --> 01:23:09,245
COMPUTER: Ma, ma.

1485
01:23:09,245 --> 01:23:10,452
PROFESSOR: --as reliably.

1486
01:23:10,452 --> 01:23:11,660
And this one sometimes falls.

1487
01:23:11,660 --> 01:23:13,200
It's kind of complicated.

1488
01:23:13,200 --> 01:23:15,534
So these are--

1489
01:23:15,534 --> 01:23:17,850
AUDIENCE: What does ma mean?

1490
01:23:17,850 --> 01:23:20,370
PROFESSOR: I don't know.

1491
01:23:20,370 --> 01:23:23,250
I don't know.

1492
01:23:23,250 --> 01:23:24,690
Once upon a time I knew.

1493
01:23:24,690 --> 01:23:26,790
COMPUTER: Mah, mah.

1494
01:23:26,790 --> 01:23:30,720
PROFESSOR: So this is
supposed to be a falling tone,

1495
01:23:30,720 --> 01:23:32,910
but it doesn't fall
very far, as you see.

1496
01:23:32,910 --> 01:23:34,160
Maybe it isn't a falling tone.

1497
01:23:34,160 --> 01:23:35,240
I don't know.

1498
01:23:35,240 --> 01:23:38,230
So that was kind of cool.

1499
01:23:38,230 --> 01:23:39,790
That's basically it.

1500
01:23:39,790 --> 01:23:41,940
Class is over.