WEBVTT
1
00:00:03.040 --> 00:00:05.089 A:middle L:90%
thank you very much steve. It's a pleasure to
2
00:00:05.099 --> 00:00:07.259 A:middle L:90%
be here. I really, really appreciate your,
3
00:00:08.039 --> 00:00:10.900 A:middle L:90%
your attention. Very nice to talk to a very
4
00:00:10.910 --> 00:00:16.879 A:middle L:90%
different audience. Yeah, so I am as as
5
00:00:16.879 --> 00:00:23.649 A:middle L:90%
steve mentioned a musician or composer and I have used
6
00:00:23.660 --> 00:00:29.609 A:middle L:90%
computers as part of my practice for more years than
7
00:00:29.620 --> 00:00:33.439 A:middle L:90%
I care to mention. Now. Today, I
8
00:00:33.439 --> 00:00:36.000 A:middle L:90%
wanted to talk to you a little bit about the
9
00:00:36.009 --> 00:00:41.799 A:middle L:90%
use of noise in computer music. And in order
10
00:00:41.799 --> 00:00:44.329 A:middle L:90%
to do that, I'm going to first give you
11
00:00:44.340 --> 00:00:50.390 A:middle L:90%
maybe a little bit of context in terms of artistic
12
00:00:50.390 --> 00:00:55.090 A:middle L:90%
practice and also talk a little bit about what noise
13
00:00:55.100 --> 00:01:00.429 A:middle L:90%
is. And then we'll dive into some musical examples
14
00:01:00.439 --> 00:01:03.719 A:middle L:90%
and eventually we will be getting to looking at coach
15
00:01:03.719 --> 00:01:06.859 A:middle L:90%
. I think that's, that's probably what gets everybody
16
00:01:06.859 --> 00:01:08.459 A:middle L:90%
excited around here. Is that right? Little bit
17
00:01:10.239 --> 00:01:12.530 A:middle L:90%
or or or you were overcome Bianca, it's sort
18
00:01:12.530 --> 00:01:19.299 A:middle L:90%
of like post code coding. Excellent. Okay,
19
00:01:19.310 --> 00:01:22.090 A:middle L:90%
well, first of all, there's the question of
20
00:01:22.090 --> 00:01:25.060 A:middle L:90%
noise, what why would you want to even mention
21
00:01:25.540 --> 00:01:29.989 A:middle L:90%
noise in the context of music? Which in almost
22
00:01:30.000 --> 00:01:34.390 A:middle L:90%
any definition of music is some form of organized.
23
00:01:34.340 --> 00:01:41.340 A:middle L:90%
So there are many different definitions of noise floating around
24
00:01:41.340 --> 00:01:44.359 A:middle L:90%
. Some of them are technical definitions. Some of
25
00:01:44.359 --> 00:01:51.450 A:middle L:90%
them are more aesthetic definitions and some of them are
26
00:01:51.459 --> 00:01:55.599 A:middle L:90%
really things that are physical measurements. Some are how
27
00:01:55.609 --> 00:02:01.140 A:middle L:90%
people perceive. So in terms in terms of music
28
00:02:01.150 --> 00:02:07.840 A:middle L:90%
, noise can were even larger context of noise is
29
00:02:07.849 --> 00:02:14.449 A:middle L:90%
sometimes referred to as unwanted sound. Yeah. In
30
00:02:14.449 --> 00:02:17.530 A:middle L:90%
the area of communication theory, it might be simply
31
00:02:17.539 --> 00:02:21.680 A:middle L:90%
sound that's not part of the message. It's And
32
00:02:21.680 --> 00:02:24.080 A:middle L:90%
in that in that sense, technically unwanted, even
33
00:02:24.080 --> 00:02:29.710 A:middle L:90%
though it might be interesting, mm there are further
34
00:02:29.710 --> 00:02:35.330 A:middle L:90%
more value judgment ways of using noise. Like if
35
00:02:35.330 --> 00:02:37.180 A:middle L:90%
somebody doesn't like a piece of music, they might
36
00:02:37.180 --> 00:02:38.259 A:middle L:90%
say, oh, that's that just sounds like noise
37
00:02:38.639 --> 00:02:43.370 A:middle L:90%
to me. So it does have kind of kind
38
00:02:43.370 --> 00:02:46.949 A:middle L:90%
of negative connotation, but at the same time,
39
00:02:46.639 --> 00:02:49.969 A:middle L:90%
in my own creative practice, I found ways to
40
00:02:49.979 --> 00:02:53.099 A:middle L:90%
use noise that seemed to it seems to be kind
41
00:02:53.099 --> 00:02:58.030 A:middle L:90%
of congruent with what I I think maybe some aspects
42
00:02:58.030 --> 00:03:00.530 A:middle L:90%
of our present culture. So I'll get to that
43
00:03:00.539 --> 00:03:02.949 A:middle L:90%
in a bit now, as far as some technical
44
00:03:04.110 --> 00:03:08.219 A:middle L:90%
definitions of noise. Um, I think the main
45
00:03:08.219 --> 00:03:13.159 A:middle L:90%
sort of technical definition of, let's say white noise
46
00:03:13.159 --> 00:03:15.919 A:middle L:90%
is the purest form of noise is just, it's
47
00:03:15.919 --> 00:03:21.360 A:middle L:90%
a random signal that has a constant power status.
48
00:03:22.039 --> 00:03:23.050 A:middle L:90%
So how many of you know what a spectrum at
49
00:03:24.939 --> 00:03:29.560 A:middle L:90%
? Okay, well, Spectra is basically the frequency
50
00:03:29.939 --> 00:03:34.219 A:middle L:90%
contact in some signal. And and since I was
51
00:03:34.229 --> 00:03:38.229 A:middle L:90%
a musician and listen to frequencies, spectrums are very
52
00:03:38.229 --> 00:03:43.569 A:middle L:90%
interesting too. two musicians. So a noise a
53
00:03:43.569 --> 00:03:46.819 A:middle L:90%
noise spectrum would be simply a spectrum that has every
54
00:03:46.819 --> 00:03:53.389 A:middle L:90%
imaginable every natural frequency in it. Now, moving
55
00:03:53.389 --> 00:03:57.659 A:middle L:90%
that from the area of just noise in general to
56
00:03:58.240 --> 00:04:00.490 A:middle L:90%
noise that we can actually play play with on our
57
00:04:00.500 --> 00:04:04.259 A:middle L:90%
computers. There is another definition of noise as a
58
00:04:04.270 --> 00:04:12.210 A:middle L:90%
discrete signal. We're the individual samples of the signal
59
00:04:12.219 --> 00:04:16.069 A:middle L:90%
are uncorrelated. Does does that make sense to people
60
00:04:16.360 --> 00:04:19.459 A:middle L:90%
? Like the difference between something being correlated and uncorrelated
61
00:04:19.939 --> 00:04:24.639 A:middle L:90%
? So the basic idea is that, and you
62
00:04:24.639 --> 00:04:28.180 A:middle L:90%
consider a sign that's a highly correlated signals. Everybody
63
00:04:28.180 --> 00:04:29.730 A:middle L:90%
knows what a sine wave is, right? Something
64
00:04:29.740 --> 00:04:30.459 A:middle L:90%
goes up and down and up and down and up
65
00:04:30.470 --> 00:04:35.379 A:middle L:90%
and down forever. So highly highly correlated signal generated
66
00:04:35.389 --> 00:04:40.240 A:middle L:90%
by a fairly simple equation. And no matter where
67
00:04:40.240 --> 00:04:42.949 A:middle L:90%
you are on that sine wave, you have a
68
00:04:42.959 --> 00:04:45.680 A:middle L:90%
really good idea. Let's say you sample that sign
69
00:04:45.689 --> 00:04:47.050 A:middle L:90%
. They say, what is the number of particular
70
00:04:47.050 --> 00:04:51.319 A:middle L:90%
point occurred? You have a pretty good idea what
71
00:04:51.319 --> 00:04:55.129 A:middle L:90%
the next point is going to be. This is
72
00:04:55.129 --> 00:04:57.360 A:middle L:90%
highly correlated. It's not going to jump around everywhere
73
00:04:57.839 --> 00:05:00.709 A:middle L:90%
. So noise is really the opposite of that.
74
00:05:00.720 --> 00:05:02.420 A:middle L:90%
No matter where you are in the noise signal signal
75
00:05:02.430 --> 00:05:06.370 A:middle L:90%
, you can basically, you have no idea what
76
00:05:06.379 --> 00:05:14.120 A:middle L:90%
the next sample in a discreet noise signal is going
77
00:05:14.120 --> 00:05:18.100 A:middle L:90%
to be. And an interesting thing about that is
78
00:05:18.110 --> 00:05:26.269 A:middle L:90%
that the single could actually look highly correlated that that's
79
00:05:26.269 --> 00:05:28.699 A:middle L:90%
voice some small segments. For example, if you're
80
00:05:28.699 --> 00:05:31.290 A:middle L:90%
just flipping a coin And you flip the coin 10
81
00:05:31.290 --> 00:05:34.060 A:middle L:90%
times in a row and you get heads 10 times
82
00:05:34.639 --> 00:05:36.970 A:middle L:90%
in. Rome, If it's an honest coin,
83
00:05:36.980 --> 00:05:41.839 A:middle L:90%
if every time you have a 5050 chance of being
84
00:05:41.850 --> 00:05:44.220 A:middle L:90%
either heads or tails, the fact that you happen
85
00:05:44.220 --> 00:05:47.459 A:middle L:90%
to get a stream stream of one's is completely consistent
86
00:05:47.939 --> 00:05:53.339 A:middle L:90%
with the idea of uncorrelated signal. Because you still
87
00:05:53.439 --> 00:05:56.050 A:middle L:90%
the next coins, you still point trust you still
88
00:05:56.050 --> 00:06:00.259 A:middle L:90%
have a 50 50 chance of getting either heads or
89
00:06:00.259 --> 00:06:05.379 A:middle L:90%
tails. Okay, now there's one last definition beyond
90
00:06:05.389 --> 00:06:10.750 A:middle L:90%
this idea of a a digital noise signal as a
91
00:06:10.750 --> 00:06:13.750 A:middle L:90%
series of a NCAA related numbers, which is that
92
00:06:14.139 --> 00:06:18.120 A:middle L:90%
in practice the kinds of noise that that one works
93
00:06:18.120 --> 00:06:24.300 A:middle L:90%
with on a computer are pseudo pseudo random noise.
94
00:06:24.839 --> 00:06:29.379 A:middle L:90%
So what that means is that the actual spectrum of
95
00:06:29.379 --> 00:06:31.629 A:middle L:90%
the noise is in fact flat, but it has
96
00:06:31.629 --> 00:06:35.550 A:middle L:90%
to be generated by some algorithm. And so that
97
00:06:35.550 --> 00:06:40.259 A:middle L:90%
means that kind of, paradoxically, on one hand
98
00:06:40.740 --> 00:06:46.470 A:middle L:90%
, the noise seems to be completely unpredictable and it
99
00:06:46.470 --> 00:06:49.800 A:middle L:90%
has it has this sort of distribution of such that
100
00:06:49.810 --> 00:06:53.660 A:middle L:90%
if you're looking at it in terms of outputs,
101
00:06:53.670 --> 00:06:56.399 A:middle L:90%
you couldn't really predict what what the next thing was
102
00:06:56.399 --> 00:06:59.540 A:middle L:90%
going to be. But because it's an algorithm being
103
00:06:59.550 --> 00:07:00.459 A:middle L:90%
run on a computer, in fact, it's a
104
00:07:00.459 --> 00:07:05.949 A:middle L:90%
totally predictable sequence and it's totally repeatable sequence to so
105
00:07:05.959 --> 00:07:09.949 A:middle L:90%
there's that kind of little bit of a wire d
106
00:07:10.339 --> 00:07:13.310 A:middle L:90%
built into working with pseudo random noise, because you're
107
00:07:13.319 --> 00:07:17.980 A:middle L:90%
getting the effect of randomness while using a in principle
108
00:07:17.990 --> 00:07:24.759 A:middle L:90%
and actually practice totally controllable disreputable out. Okay.
109
00:07:25.540 --> 00:07:28.560 A:middle L:90%
Okay, so that's a little bit about definitions.
110
00:07:28.939 --> 00:07:31.329 A:middle L:90%
Now, let me try to put this in the
111
00:07:31.329 --> 00:07:38.529 A:middle L:90%
context of artistic practice. How many of you do
112
00:07:38.540 --> 00:07:42.439 A:middle L:90%
some kind of artwork, whether it's writing poetry or
113
00:07:42.449 --> 00:07:47.449 A:middle L:90%
making films or composing or writing novels. Okay,
114
00:07:47.839 --> 00:07:50.990 A:middle L:90%
and how many of you do not don't think of
115
00:07:50.990 --> 00:07:58.089 A:middle L:90%
yourself as being creative artists. Okay, all the
116
00:07:58.089 --> 00:08:05.209 A:middle L:90%
faculty raised their hand. Okay, so as as
117
00:08:05.209 --> 00:08:07.439 A:middle L:90%
we can imagine, I'm glad you put the word
118
00:08:07.439 --> 00:08:09.850 A:middle L:90%
. Or even even those of you who don't,
119
00:08:09.439 --> 00:08:13.790 A:middle L:90%
our artistic creators are no doubt artistic consumers of some
120
00:08:13.790 --> 00:08:16.410 A:middle L:90%
sort. You all listen to music, you watch
121
00:08:16.420 --> 00:08:22.050 A:middle L:90%
films, read books of books and additional that are
122
00:08:22.050 --> 00:08:26.350 A:middle L:90%
not just technical books. So, so it's it's
123
00:08:26.350 --> 00:08:31.329 A:middle L:90%
quite clear that artistic practice is such a broad thing
124
00:08:31.329 --> 00:08:33.159 A:middle L:90%
that there are many different ways of thinking about.
125
00:08:33.840 --> 00:08:37.120 A:middle L:90%
One way to think about artistic practice, let's say
126
00:08:37.120 --> 00:08:41.460 A:middle L:90%
that you're composing a piece of music and you,
127
00:08:43.039 --> 00:08:45.759 A:middle L:90%
one way to think about that is you're just making
128
00:08:45.759 --> 00:08:50.309 A:middle L:90%
a series of decisions. Okay, so let's let's
129
00:08:50.309 --> 00:08:50.889 A:middle L:90%
say that you're writing a melody and you're thinking,
130
00:08:50.899 --> 00:08:54.029 A:middle L:90%
well, should the next note b a long note
131
00:08:54.039 --> 00:08:56.649 A:middle L:90%
should be a short note. Do I want this
132
00:08:56.649 --> 00:09:03.159 A:middle L:90%
melody to be a, a really kind of dramatic
133
00:09:03.340 --> 00:09:05.679 A:middle L:90%
sort of sort of intense thing? Do I want
134
00:09:05.679 --> 00:09:09.970 A:middle L:90%
it to be relaxed thing? And so all all
135
00:09:09.970 --> 00:09:13.000 A:middle L:90%
those sorts of decisions are made at lots and lots
136
00:09:13.000 --> 00:09:16.659 A:middle L:90%
of different levels in creating a piece of music and
137
00:09:16.659 --> 00:09:18.830 A:middle L:90%
there's a lot of iteration, you might make a
138
00:09:18.830 --> 00:09:22.009 A:middle L:90%
decision and compose a little bit of peace and then
139
00:09:22.009 --> 00:09:24.879 A:middle L:90%
step back and think about it and say, well
140
00:09:24.950 --> 00:09:26.299 A:middle L:90%
actually that melody is pretty terrible. Maybe I should
141
00:09:26.299 --> 00:09:28.570 A:middle L:90%
write a new one. Or you can say,
142
00:09:28.580 --> 00:09:31.809 A:middle L:90%
well, This melody is interesting and I can do
143
00:09:31.820 --> 00:09:35.389 A:middle L:90%
the following 10 things with it that I think would
144
00:09:35.399 --> 00:09:37.649 A:middle L:90%
be enjoyable to listen to. And so that as
145
00:09:37.649 --> 00:09:43.009 A:middle L:90%
a result conditions further decisions. So if you,
146
00:09:43.019 --> 00:09:50.120 A:middle L:90%
if you think about on artistic practice as in this
147
00:09:50.120 --> 00:09:52.289 A:middle L:90%
kind of reductive way, a series of decisions are
148
00:09:52.289 --> 00:09:54.639 A:middle L:90%
made by an artist, then one interesting point comes
149
00:09:54.639 --> 00:10:01.049 A:middle L:90%
up, which is how constrained are you're making those
150
00:10:01.049 --> 00:10:03.850 A:middle L:90%
decisions or on the other hand, how free are
151
00:10:05.039 --> 00:10:09.720 A:middle L:90%
. And the the interesting thing is that sometimes you
152
00:10:09.720 --> 00:10:13.490 A:middle L:90%
might feel that you're free when you actually are being
153
00:10:13.490 --> 00:10:16.159 A:middle L:90%
constrained by forces that you're not even aware of.
154
00:10:16.639 --> 00:10:20.690 A:middle L:90%
So let's say, for example, that you're writing
155
00:10:20.769 --> 00:10:28.059 A:middle L:90%
a Yeah, technically, Well, the odds are
156
00:10:28.070 --> 00:10:31.389 A:middle L:90%
it's going to be in 44 right? Because it
157
00:10:31.389 --> 00:10:33.399 A:middle L:90%
has to work on the dance floor. So you
158
00:10:33.399 --> 00:10:39.019 A:middle L:90%
immediately have certain stylistic constraints. And if you're writing
159
00:10:39.029 --> 00:10:41.279 A:middle L:90%
right, techno track, it probably is not going
160
00:10:41.279 --> 00:10:43.769 A:middle L:90%
to be for strength work because it has to be
161
00:10:43.779 --> 00:10:46.429 A:middle L:90%
played by the DJ. So it's going to it's
162
00:10:46.429 --> 00:10:50.059 A:middle L:90%
going to have to be some form of electronic music
163
00:10:50.440 --> 00:10:52.559 A:middle L:90%
. So just by deciding what kind of peace you
164
00:10:52.559 --> 00:10:54.750 A:middle L:90%
want to write, you already have all sorts of
165
00:10:54.759 --> 00:10:58.259 A:middle L:90%
constraints that are coming into play. And if you
166
00:10:58.259 --> 00:11:03.129 A:middle L:90%
go further back in history, when people didn't have
167
00:11:03.139 --> 00:11:07.649 A:middle L:90%
access to nearly as much cultural information as we have
168
00:11:07.659 --> 00:11:11.830 A:middle L:90%
available to us instantaneously, you find that stylistic constraints
169
00:11:11.840 --> 00:11:16.529 A:middle L:90%
become considerably narrower in a way, so that so
170
00:11:16.529 --> 00:11:20.149 A:middle L:90%
that it becomes a lot easier to talk about common
171
00:11:20.149 --> 00:11:28.279 A:middle L:90%
practices or shared, shared stylistic constraints that affect a
172
00:11:28.289 --> 00:11:31.789 A:middle L:90%
practice of a large number of composers. And that's
173
00:11:31.029 --> 00:11:37.590 A:middle L:90%
why let's say if we were listening to the music
174
00:11:37.600 --> 00:11:41.549 A:middle L:90%
music of Bath, whose basically and early to mid
175
00:11:41.559 --> 00:11:46.159 A:middle L:90%
18th century composer, you here very different kind of
176
00:11:48.039 --> 00:11:50.259 A:middle L:90%
style, A different style, let's say then if
177
00:11:50.259 --> 00:11:56.149 A:middle L:90%
you're listening to a mentally 19th Century Proposal Like Block
178
00:11:58.139 --> 00:12:01.889 A:middle L:90%
. Okay, so another way to think about constraints
179
00:12:01.899 --> 00:12:05.549 A:middle L:90%
is how you actually make and use it and there
180
00:12:07.740 --> 00:12:11.129 A:middle L:90%
you could you could make a a sort of a
181
00:12:11.139 --> 00:12:16.299 A:middle L:90%
continuity between on the one hand, just freely making
182
00:12:16.309 --> 00:12:22.450 A:middle L:90%
decisions only constrained, let's say bye style or doing
183
00:12:22.450 --> 00:12:24.220 A:middle L:90%
the kind of composition. That's actually much more rule
184
00:12:24.220 --> 00:12:30.350 A:middle L:90%
based that actually has a kind of algorithmic aspect to
185
00:12:31.039 --> 00:12:35.159 A:middle L:90%
And how many of you were musicians. Okay,
186
00:12:35.169 --> 00:12:37.870 A:middle L:90%
so for those of you who are musicians, can
187
00:12:37.870 --> 00:12:43.450 A:middle L:90%
you think of any any form of music, any
188
00:12:43.450 --> 00:12:46.960 A:middle L:90%
kind of music music that is very strictly rule based
189
00:12:48.840 --> 00:12:52.019 A:middle L:90%
You can take 12 argues, you can do what
190
00:12:52.029 --> 00:12:56.549 A:middle L:90%
gratifying loose exactly. That is a beautiful event.
191
00:12:56.039 --> 00:13:01.529 A:middle L:90%
So you have a series of chords and and that
192
00:13:01.529 --> 00:13:05.230 A:middle L:90%
series is basically a fixed form that you have to
193
00:13:05.240 --> 00:13:07.220 A:middle L:90%
work with. And then of course, within that
194
00:13:07.230 --> 00:13:09.820 A:middle L:90%
form, there's any number of ways that it can
195
00:13:09.830 --> 00:13:11.870 A:middle L:90%
be expressed, but it's an algorithm. I mean
196
00:13:11.870 --> 00:13:16.330 A:middle L:90%
if you do a 13 are Sequence all of a
197
00:13:16.330 --> 00:13:20.629 A:middle L:90%
sudden is not 12 lose it has rules. Another
198
00:13:20.629 --> 00:13:24.440 A:middle L:90%
example would be a cab That's sort of like the
199
00:13:24.440 --> 00:13:28.330 A:middle L:90%
song three blind mice where you just you basically using
200
00:13:28.330 --> 00:13:31.269 A:middle L:90%
a melody and then somebody else sings exactly the same
201
00:13:31.269 --> 00:13:35.909 A:middle L:90%
melody at a delay and so forth. And it's
202
00:13:35.909 --> 00:13:41.149 A:middle L:90%
composed in such a way that that the piece builds
203
00:13:41.149 --> 00:13:43.379 A:middle L:90%
up harmoniously, but nonetheless, there was a very
204
00:13:43.379 --> 00:13:46.750 A:middle L:90%
strict constraint there there's a very strict, I'll.
205
00:13:48.639 --> 00:13:50.730 A:middle L:90%
So, in that way of thinking about music,
206
00:13:50.740 --> 00:13:58.059 A:middle L:90%
you really kind of have a a story of greed
207
00:13:58.539 --> 00:14:05.629 A:middle L:90%
from highly constrained ways of composing. Too much freer
208
00:14:05.639 --> 00:14:07.190 A:middle L:90%
ways of composing. And the interesting thing is,
209
00:14:07.200 --> 00:14:11.210 A:middle L:90%
even in the freer ways of composing, you still
210
00:14:11.220 --> 00:14:18.220 A:middle L:90%
ultimately have some constraints coming from your own personal habits
211
00:14:18.460 --> 00:14:24.220 A:middle L:90%
, especially after you, after you compose for four
212
00:14:24.220 --> 00:14:26.480 A:middle L:90%
years for decades, you start doing the same things
213
00:14:26.490 --> 00:14:31.460 A:middle L:90%
over and over again. So, in that,
214
00:14:31.940 --> 00:14:37.659 A:middle L:90%
in that context, it's useful to have compositional techniques
215
00:14:37.799 --> 00:14:43.350 A:middle L:90%
that can pull you outside of stylistic constraints. Mhm
216
00:14:45.240 --> 00:14:48.570 A:middle L:90%
. Your own personal habits to get you into other
217
00:14:48.580 --> 00:14:52.210 A:middle L:90%
areas to get you two kinds of musical expression that
218
00:14:52.220 --> 00:14:56.860 A:middle L:90%
you might not have gone to without without using these
219
00:14:56.860 --> 00:15:01.529 A:middle L:90%
techniques. So, in a sense, noise could
220
00:15:01.539 --> 00:15:07.889 A:middle L:90%
be used as a way of disrupting some existing patterns
221
00:15:09.000 --> 00:15:11.250 A:middle L:90%
that might get you to a different kind of music
222
00:15:11.340 --> 00:15:15.600 A:middle L:90%
. Okay, so what I'd like to do now
223
00:15:15.610 --> 00:15:18.929 A:middle L:90%
very quickly is before getting to the actual deployment of
224
00:15:18.929 --> 00:15:26.230 A:middle L:90%
noise is give you a little bit of historical background
225
00:15:26.240 --> 00:15:28.909 A:middle L:90%
of a few different techniques that, Well, they're
226
00:15:28.909 --> 00:15:33.279 A:middle L:90%
not exactly noise techniques. They lead to certain unpredictable
227
00:15:33.289 --> 00:15:37.840 A:middle L:90%
results. And one of the first one is the
228
00:15:37.850 --> 00:15:43.669 A:middle L:90%
14th century practice of ice arrhythmic motet and that's that's
229
00:15:43.860 --> 00:15:50.059 A:middle L:90%
quite interesting approach to music, essentially, when we
230
00:15:50.059 --> 00:15:54.899 A:middle L:90%
think about music in terms of melody as as a
231
00:15:54.899 --> 00:16:00.169 A:middle L:90%
central feature, the melody is a is a sort
232
00:16:00.169 --> 00:16:07.320 A:middle L:90%
of an indestructible unit of a combination of pitch patterns
233
00:16:07.320 --> 00:16:12.309 A:middle L:90%
and rhythms. They go together. Now in later
234
00:16:12.309 --> 00:16:15.360 A:middle L:90%
in development, you might try different, different patterns
235
00:16:15.940 --> 00:16:19.580 A:middle L:90%
associated with those, but the melody itself as a
236
00:16:19.590 --> 00:16:25.230 A:middle L:90%
unit really puts these things together. So the interesting
237
00:16:25.230 --> 00:16:30.509 A:middle L:90%
thing about the ice arrhythmic attack is that it separates
238
00:16:30.519 --> 00:16:34.960 A:middle L:90%
out rhythm and pitch contours. And very often what
239
00:16:34.960 --> 00:16:38.419 A:middle L:90%
will happen is there will be a melody with no
240
00:16:38.429 --> 00:16:45.590 A:middle L:90%
rhythm and rhythms that repeat at various at various time
241
00:16:45.590 --> 00:16:48.440 A:middle L:90%
scales. So the result is that you get some
242
00:16:48.450 --> 00:16:53.210 A:middle L:90%
rather complex music and also once you start this process
243
00:16:53.210 --> 00:16:56.580 A:middle L:90%
off, let's say that you come up with some
244
00:16:56.580 --> 00:16:59.419 A:middle L:90%
sort of pitch melody and some sort of rhythm,
245
00:16:59.419 --> 00:17:03.069 A:middle L:90%
melt Melody, you couldn't necessarily predict in advance without
246
00:17:03.079 --> 00:17:07.059 A:middle L:90%
working through the algorithm what the results will be.
247
00:17:07.069 --> 00:17:10.259 A:middle L:90%
And you get these very interesting permutations. So,
248
00:17:10.259 --> 00:17:12.539 A:middle L:90%
I'd like to very quickly play you an example of
249
00:17:12.549 --> 00:17:17.779 A:middle L:90%
a piece by the 14th century composer, Philippe de
250
00:17:17.779 --> 00:17:22.140 A:middle L:90%
Vitry. He's part of the ars nova uh Movement
251
00:17:22.150 --> 00:17:26.720 A:middle L:90%
in the early 14th century France. And so we
252
00:17:26.720 --> 00:17:29.759 A:middle L:90%
would like you to be aware of is the,
253
00:17:29.769 --> 00:17:33.630 A:middle L:90%
just the complexity as a result of what is essentially
254
00:17:33.640 --> 00:17:40.180 A:middle L:90%
a musical algorithm where you are playing out pitch pitch
255
00:17:40.180 --> 00:17:44.670 A:middle L:90%
structures and rhythm structures that are not in line.
256
00:18:34.470 --> 00:18:40.619 A:middle L:90%
Oh, okay. I want him to play enough
257
00:18:40.619 --> 00:18:42.990 A:middle L:90%
of that for you here in the sense of how
258
00:18:42.990 --> 00:18:45.619 A:middle L:90%
it plays out. And this to me is a
259
00:18:45.630 --> 00:18:51.339 A:middle L:90%
really good example of how a compositional technique creating to
260
00:18:51.339 --> 00:18:53.670 A:middle L:90%
a kind of music that you might not close into
261
00:18:53.670 --> 00:18:57.759 A:middle L:90%
it. If you were just sort of strong son
262
00:18:59.339 --> 00:19:02.950 A:middle L:90%
towards on allude and singing along with it, you
263
00:19:02.960 --> 00:19:04.460 A:middle L:90%
would wind up with a much simpler kind of music
264
00:19:04.839 --> 00:19:07.730 A:middle L:90%
. But what's interesting here is, did you notice
265
00:19:07.730 --> 00:19:12.299 A:middle L:90%
how some parts of the songs are moving very fast
266
00:19:12.369 --> 00:19:15.369 A:middle L:90%
and some parts of moving very slow? It's a
267
00:19:15.380 --> 00:19:21.240 A:middle L:90%
very interesting, intricate thing. I mean, The
268
00:19:21.250 --> 00:19:25.079 A:middle L:90%
14th century was really very interesting, kind of an
269
00:19:25.089 --> 00:19:30.220 A:middle L:90%
avant garde period in in french and and other european
270
00:19:30.220 --> 00:19:33.950 A:middle L:90%
areas of composition. So I want to jump now
271
00:19:33.960 --> 00:19:38.890 A:middle L:90%
ahead. That that's an example of a an algorithm
272
00:19:38.890 --> 00:19:45.109 A:middle L:90%
which is relatively deterministic. It's not completely automatic because
273
00:19:45.109 --> 00:19:48.000 A:middle L:90%
the composer then decides which, you know, how
274
00:19:48.000 --> 00:19:52.119 A:middle L:90%
long to stretch these things out, how to combine
275
00:19:52.130 --> 00:19:56.940 A:middle L:90%
the different elements to get harmonic relationships that you now
276
00:19:56.950 --> 00:20:02.240 A:middle L:90%
, I'm going to jump ahead to another kind of
277
00:20:02.250 --> 00:20:04.990 A:middle L:90%
composition which is again not noise based, but which
278
00:20:04.990 --> 00:20:07.559 A:middle L:90%
seems to take you to the edge of noise.
279
00:20:08.640 --> 00:20:12.400 A:middle L:90%
In the early 20th century, there was a technique
280
00:20:12.410 --> 00:20:18.150 A:middle L:90%
that was developed in the post tonal period called 12
281
00:20:18.150 --> 00:20:21.539 A:middle L:90%
tone music. And the idea talked on music was
282
00:20:21.549 --> 00:20:23.299 A:middle L:90%
that you would take all of the possible notes,
283
00:20:23.359 --> 00:20:29.579 A:middle L:90%
all of the chromatic notes of the of the european
284
00:20:29.589 --> 00:20:33.400 A:middle L:90%
musical scale and put them in a particular order and
285
00:20:33.410 --> 00:20:37.470 A:middle L:90%
then use that ordering of those notes as the basis
286
00:20:37.470 --> 00:20:40.599 A:middle L:90%
for entire piece. So in a way that was
287
00:20:40.599 --> 00:20:44.910 A:middle L:90%
a much tire constraint than the way music was often
288
00:20:44.910 --> 00:20:48.730 A:middle L:90%
composed prior to that. So it was an algorithm
289
00:20:48.740 --> 00:20:53.720 A:middle L:90%
to tighten up the compositional procedure. But the result
290
00:20:53.730 --> 00:20:56.900 A:middle L:90%
was to get the kind of music which sounds nearly
291
00:20:56.910 --> 00:21:00.029 A:middle L:90%
random. Now, what I'm going to play you
292
00:21:00.029 --> 00:21:03.220 A:middle L:90%
a little bit of is a piece by a french
293
00:21:03.220 --> 00:21:04.210 A:middle L:90%
composer, he's still alive. His name is Pierre
294
00:21:04.210 --> 00:21:10.339 A:middle L:90%
boulez And this is a piece from 1952 called structures
295
00:21:10.349 --> 00:21:14.170 A:middle L:90%
or structures. And it takes the idea of a
296
00:21:14.180 --> 00:21:18.009 A:middle L:90%
fixed structure that is employed over and over again to
297
00:21:18.019 --> 00:21:21.890 A:middle L:90%
pitch. And like the ice arrhythmic attack of de
298
00:21:21.890 --> 00:21:26.390 A:middle L:90%
Vitry employ applies it to other aspects of the music
299
00:21:26.400 --> 00:21:29.869 A:middle L:90%
. So in addition to applying it to pitch,
300
00:21:29.880 --> 00:21:32.549 A:middle L:90%
he applies it to register. Is it a high
301
00:21:32.549 --> 00:21:34.029 A:middle L:90%
note? Is it a low note? He applies
302
00:21:34.029 --> 00:21:37.940 A:middle L:90%
it to dynamics? Is it loud? Is it
303
00:21:37.940 --> 00:21:42.579 A:middle L:90%
quiet? He applies it to articulations. So very
304
00:21:42.579 --> 00:21:45.670 A:middle L:90%
much like an ice arrhythmic motet. The idea of
305
00:21:45.670 --> 00:21:51.170 A:middle L:90%
the coherence of a melody is completely split apart and
306
00:21:51.170 --> 00:21:55.589 A:middle L:90%
instead you have all of these structures that separately are
307
00:21:55.589 --> 00:21:59.900 A:middle L:90%
imposed to create an accurate result. So let's listen
308
00:21:59.900 --> 00:22:10.690 A:middle L:90%
to a little bit of struck tours snake. Mm
309
00:22:17.539 --> 00:22:23.329 A:middle L:90%
Yeah. Mm. Now, the interesting thing about
310
00:22:23.329 --> 00:22:26.269 A:middle L:90%
that, did that, is there anybody who didn't
311
00:22:26.269 --> 00:22:30.089 A:middle L:90%
hear that as having a large amount of randomness in
312
00:22:30.089 --> 00:22:34.230 A:middle L:90%
it? So the interesting thing is very much like
313
00:22:34.230 --> 00:22:40.619 A:middle L:90%
pseudo random noise, that score was totally determined by
314
00:22:40.619 --> 00:22:47.539 A:middle L:90%
algorithms. There was actually almost no composer intervention.
315
00:22:47.549 --> 00:22:51.289 A:middle L:90%
The composer intervention was basically about making the rules and
316
00:22:51.289 --> 00:22:53.539 A:middle L:90%
once the rules were made that music is what came
317
00:22:53.539 --> 00:23:00.059 A:middle L:90%
out now. Uh huh. About the same time
318
00:23:00.039 --> 00:23:04.049 A:middle L:90%
, there was another area of research which was in
319
00:23:04.049 --> 00:23:11.170 A:middle L:90%
the area of really trying to create a stochastic music
320
00:23:11.539 --> 00:23:14.990 A:middle L:90%
, a kind of music that had algorithms but those
321
00:23:15.000 --> 00:23:18.099 A:middle L:90%
algorithms were driven by noise. And so I'd like
322
00:23:18.099 --> 00:23:23.140 A:middle L:90%
to play you an example of what is considered by
323
00:23:23.140 --> 00:23:26.230 A:middle L:90%
some, to some people to be the first computer
324
00:23:26.230 --> 00:23:30.519 A:middle L:90%
using this is the Iliac Suite, which was which
325
00:23:30.519 --> 00:23:33.799 A:middle L:90%
was created by a composer named jerry Heller. Uh
326
00:23:33.809 --> 00:23:37.410 A:middle L:90%
he was a chemistry professor working at the University of
327
00:23:37.420 --> 00:23:41.980 A:middle L:90%
Illinois in Urbana Champagne. And you'll hear that there's
328
00:23:41.980 --> 00:23:45.839 A:middle L:90%
a certain amount of composer intervention in the style,
329
00:23:45.849 --> 00:23:49.779 A:middle L:90%
but the overall structure of this is generated by the
330
00:23:49.789 --> 00:23:55.009 A:middle L:90%
application of noise to a set of rules to generate
331
00:23:55.009 --> 00:23:57.089 A:middle L:90%
music. So this is this is algorithmic music with
332
00:23:57.089 --> 00:24:06.269 A:middle L:90%
noise coming in and at the top end. Okay
333
00:24:11.440 --> 00:24:30.990 A:middle L:90%
. Okay. Yeah. Yeah. Did that sound
334
00:24:30.990 --> 00:24:36.059 A:middle L:90%
gravitating? It sounds highly structured. And it's a
335
00:24:36.059 --> 00:24:40.599 A:middle L:90%
really interesting thing that, you know, like so
336
00:24:40.609 --> 00:24:41.730 A:middle L:90%
much about music. On the one hand, we
337
00:24:41.740 --> 00:24:47.480 A:middle L:90%
have technical definitions of what we're working with, The
338
00:24:47.480 --> 00:24:49.099 A:middle L:90%
technical definition of white noise. But on the other
339
00:24:49.099 --> 00:24:52.640 A:middle L:90%
side we have how we actually perceive these things and
340
00:24:52.640 --> 00:24:57.670 A:middle L:90%
very often there's there's tension or real dissidence. So
341
00:24:57.680 --> 00:25:00.980 A:middle L:90%
on the one hand you have, we heard the
342
00:25:00.990 --> 00:25:03.450 A:middle L:90%
peace bible s which was totally determined and highly pattern
343
00:25:04.039 --> 00:25:07.430 A:middle L:90%
sounding totally random. And on the other hand we
344
00:25:07.430 --> 00:25:12.549 A:middle L:90%
have the killer which was very largely ran in how
345
00:25:12.549 --> 00:25:17.259 A:middle L:90%
it was assembled. But with some rules to force
346
00:25:17.269 --> 00:25:21.609 A:middle L:90%
that those random decisions to sell musical and we just
347
00:25:21.609 --> 00:25:22.900 A:middle L:90%
don't hear it is rand. So I think that's
348
00:25:23.349 --> 00:25:29.500 A:middle L:90%
very interesting. And I will be getting shortly to
349
00:25:29.500 --> 00:25:32.049 A:middle L:90%
my own compositional practice, ensure you showing you what
350
00:25:32.049 --> 00:25:36.269 A:middle L:90%
I actually do that. Now I'd like to Mhm
351
00:25:37.240 --> 00:25:45.440 A:middle L:90%
. Now move to some actual code and so yeah
352
00:25:45.339 --> 00:25:51.339 A:middle L:90%
, actual code here. So what I'm showing you
353
00:25:51.339 --> 00:25:55.740 A:middle L:90%
now is what is, in my opinion, one
354
00:25:55.740 --> 00:25:57.869 A:middle L:90%
of the most interesting computer music languages. It's a
355
00:25:57.869 --> 00:26:04.569 A:middle L:90%
language called Supercollider. And uh it's interesting parts important
356
00:26:04.569 --> 00:26:08.259 A:middle L:90%
because it actually really is the language it really is
357
00:26:08.259 --> 00:26:11.549 A:middle L:90%
a computer music language. Where as some of the
358
00:26:11.559 --> 00:26:14.910 A:middle L:90%
other computer music languages, something like C sound is
359
00:26:14.910 --> 00:26:18.059 A:middle L:90%
really a little bit more like assembler code. Uh
360
00:26:18.539 --> 00:26:21.950 A:middle L:90%
And another program that's very popular, which we'll see
361
00:26:21.950 --> 00:26:25.069 A:middle L:90%
in a bit called maX MSP, is a visual
362
00:26:25.069 --> 00:26:27.839 A:middle L:90%
data float program. And as you probably know,
363
00:26:27.849 --> 00:26:33.000 A:middle L:90%
actually coded with data flow programs can be incredibly frustrating
364
00:26:33.000 --> 00:26:34.829 A:middle L:90%
when you just want to write a line of code
365
00:26:34.839 --> 00:26:38.519 A:middle L:90%
to say exactly what you want to have. So
366
00:26:38.529 --> 00:26:45.210 A:middle L:90%
what supercollider is essentially is a sort of a mix
367
00:26:45.220 --> 00:26:48.970 A:middle L:90%
of a sound synthesis language that generates audio samples.
368
00:26:48.980 --> 00:26:56.119 A:middle L:90%
That is controlled by a high level language which will
369
00:26:56.119 --> 00:26:59.130 A:middle L:90%
have a lot of features that will be familiar to
370
00:26:59.130 --> 00:27:03.960 A:middle L:90%
you. It has things like like generations and functions
371
00:27:03.960 --> 00:27:11.960 A:middle L:90%
and variables. Okay, so let's start with noise
372
00:27:11.960 --> 00:27:17.579 A:middle L:90%
and just how we perceive it. So one way
373
00:27:17.579 --> 00:27:21.460 A:middle L:90%
the easiest way to create sound in Super Collyer is
374
00:27:21.470 --> 00:27:26.200 A:middle L:90%
to create a function and evaluated by sending a message
375
00:27:26.210 --> 00:27:29.480 A:middle L:90%
place. So inside here I'm going to write a
376
00:27:29.490 --> 00:27:41.160 A:middle L:90%
very simple program and executed. So one thing about
377
00:27:41.160 --> 00:27:45.369 A:middle L:90%
supercollider is it is totally object oriented. Every everything
378
00:27:45.380 --> 00:27:48.849 A:middle L:90%
in supercollider, including numbers, they're all objects.
379
00:27:48.849 --> 00:27:56.420 A:middle L:90%
So for example, if I type the number five
380
00:27:56.430 --> 00:27:59.279 A:middle L:90%
and evaluate it, I'm not going to show you
381
00:27:59.279 --> 00:28:00.789 A:middle L:90%
exactly how I evaluate it. It just evaluates to
382
00:28:00.799 --> 00:28:03.950 A:middle L:90%
five, but I can send it the message,
383
00:28:03.960 --> 00:28:08.859 A:middle L:90%
reciprocal And it will evaluate to 1/5. So it's
384
00:28:08.859 --> 00:28:12.279 A:middle L:90%
a really convenient sort of interpreted language that's that's a
385
00:28:12.279 --> 00:28:15.170 A:middle L:90%
lot of fun to play with. Now, what's
386
00:28:15.170 --> 00:28:19.329 A:middle L:90%
happening here is white noise. Is an audio object
387
00:28:19.329 --> 00:28:23.240 A:middle L:90%
is an object that will generate audio samples if you
388
00:28:23.240 --> 00:28:26.529 A:middle L:90%
send the right message and dot A. R.
389
00:28:26.539 --> 00:28:32.460 A:middle L:90%
Is a message to tell this object to start generating
390
00:28:32.940 --> 00:28:37.559 A:middle L:90%
pseudo random noise at the audio raid. And this
391
00:28:37.559 --> 00:28:41.250 A:middle L:90%
one parameter in here is a an amplitude value just
392
00:28:41.250 --> 00:28:47.299 A:middle L:90%
says multiply the output by a certain amount. So
393
00:28:47.309 --> 00:28:51.049 A:middle L:90%
if I evaluate this, then we get white.
394
00:28:52.039 --> 00:28:55.859 A:middle L:90%
Yeah, I could also just take this part out
395
00:28:56.440 --> 00:29:02.750 A:middle L:90%
and just multiply 5.1 and we should be only getting
396
00:29:02.779 --> 00:29:07.819 A:middle L:90%
in one speaker. So if I, another nice
397
00:29:07.819 --> 00:29:11.210 A:middle L:90%
thing about supercollider is involved multi channel expansion. So
398
00:29:11.210 --> 00:29:15.170 A:middle L:90%
if I want Rather than one channel, 128 channels
399
00:29:15.539 --> 00:29:18.619 A:middle L:90%
, which might be useful for the cube. As
400
00:29:18.619 --> 00:29:21.390 A:middle L:90%
a side note how many people are aware of the
401
00:29:21.390 --> 00:29:22.259 A:middle L:90%
existence of the cube at the Center for the Arts
402
00:29:23.140 --> 00:29:26.059 A:middle L:90%
. Okay. It's a very cool place that has
403
00:29:26.440 --> 00:29:29.759 A:middle L:90%
Upwards of 128. So super flyer is really the
404
00:29:29.759 --> 00:29:32.109 A:middle L:90%
language to use. But if I expand to two
405
00:29:32.109 --> 00:29:37.609 A:middle L:90%
channels then stones out now that is going to just
406
00:29:37.609 --> 00:29:41.130 A:middle L:90%
generate a sequence of white noise and it has a
407
00:29:41.130 --> 00:29:42.109 A:middle L:90%
kind of a sound to it. If I if
408
00:29:42.109 --> 00:29:44.619 A:middle L:90%
I just let it run for a little bit,
409
00:29:44.619 --> 00:29:45.460 A:middle L:90%
maybe I should just play a little bit louder there
410
00:29:48.640 --> 00:30:00.509 A:middle L:90%
. Mhm. Mhm. Does that sound more or
411
00:30:00.509 --> 00:30:07.400 A:middle L:90%
less uniform to you? It has a sound more
412
00:30:07.400 --> 00:30:11.250 A:middle L:90%
or less. I mean it's not exactly 100 uniform
413
00:30:11.250 --> 00:30:14.170 A:middle L:90%
, but it's sort of sort of comic a little
414
00:30:14.170 --> 00:30:18.930 A:middle L:90%
bit like maybe maybe fan noise or or the sound
415
00:30:18.930 --> 00:30:22.869 A:middle L:90%
of a breeze. But it's definitely this kind of
416
00:30:22.880 --> 00:30:27.630 A:middle L:90%
ambiance signal, Interesting thing is remember what I mentioned
417
00:30:27.630 --> 00:30:30.400 A:middle L:90%
about, you could be flipping a coin which is
418
00:30:30.400 --> 00:30:34.069 A:middle L:90%
a random signal and you might wind up with 10
419
00:30:34.079 --> 00:30:38.109 A:middle L:90%
heads in a row. The interesting thing is if
420
00:30:38.109 --> 00:30:45.509 A:middle L:90%
we actually mhm capture a small amount of sound of
421
00:30:45.519 --> 00:30:48.900 A:middle L:90%
noise, actually hear patterns and the noise, which
422
00:30:48.900 --> 00:30:49.900 A:middle L:90%
is kind of interesting. So what I'm, what
423
00:30:49.900 --> 00:30:55.190 A:middle L:90%
I have here right now is um I'm going to
424
00:30:55.190 --> 00:30:56.920 A:middle L:90%
create a buffer and buffer is just an array that's
425
00:30:56.920 --> 00:31:03.289 A:middle L:90%
going to store a series of numbers. So that's
426
00:31:03.289 --> 00:31:06.670 A:middle L:90%
what this thing is and I'm going to store basically
427
00:31:07.839 --> 00:31:15.500 A:middle L:90%
, let's say one second worth of this. And
428
00:31:15.500 --> 00:31:17.779 A:middle L:90%
now what I'm going to do is I'm going to
429
00:31:17.779 --> 00:31:26.059 A:middle L:90%
record into that buffer and we're not actually hearing anything
430
00:31:26.069 --> 00:31:29.220 A:middle L:90%
, but that was more than a second. So
431
00:31:29.220 --> 00:31:30.369 A:middle L:90%
if I wanted to I could go ahead and plot
432
00:31:30.369 --> 00:31:34.559 A:middle L:90%
this buffer and you can see there there is that
433
00:31:34.569 --> 00:31:40.309 A:middle L:90%
noise and we're not I don't know how well we
434
00:31:40.309 --> 00:31:41.430 A:middle L:90%
consume in here, but you can see this is
435
00:31:41.430 --> 00:31:47.670 A:middle L:90%
very spiky signal, there's really no no correlation from
436
00:31:47.670 --> 00:31:51.109 A:middle L:90%
one sample to the next. That is evident even
437
00:31:51.109 --> 00:31:56.549 A:middle L:90%
though this is was generated by a totally repeatable algorithm
438
00:31:56.940 --> 00:32:00.750 A:middle L:90%
. So now what happens if I play that noise
439
00:32:00.759 --> 00:32:15.000 A:middle L:90%
as a loop? Does anybody? Yeah, Okay
440
00:32:15.000 --> 00:32:20.849 A:middle L:90%
, now let's let's even even shorter here. So
441
00:32:20.859 --> 00:32:22.569 A:middle L:90%
instead, what I'm gonna do is make this buffer
442
00:32:22.569 --> 00:32:31.440 A:middle L:90%
size, let's say 200 milliseconds and then we reported
443
00:32:31.450 --> 00:32:40.470 A:middle L:90%
to it and then we'll play you're going to hear
444
00:32:40.470 --> 00:32:45.970 A:middle L:90%
that. And we can actually if we took it
445
00:32:45.980 --> 00:32:52.029 A:middle L:90%
inside of the make it short enough That the repetition
446
00:32:52.039 --> 00:32:55.490 A:middle L:90%
would go above 20 Harris would move into human audio
447
00:32:57.140 --> 00:32:59.640 A:middle L:90%
, we would actually hear this as a pitch and
448
00:32:59.640 --> 00:33:07.359 A:middle L:90%
I'll demonstrate that very waiting for you. Okay,
449
00:33:08.140 --> 00:33:10.279 A:middle L:90%
so we're still listening to noise. I mean if
450
00:33:10.279 --> 00:33:15.640 A:middle L:90%
we wanted to plot this thing thus, that's what
451
00:33:15.640 --> 00:33:20.279 A:middle L:90%
we're hearing. What we're hearing is that repeated over
452
00:33:20.279 --> 00:33:22.880 A:middle L:90%
and over again. And it gives us the perception
453
00:33:22.880 --> 00:33:27.460 A:middle L:90%
of pitch because we hear the larger pattern of repetition
454
00:33:27.839 --> 00:33:31.190 A:middle L:90%
over there. Okay, now let me give you
455
00:33:31.200 --> 00:33:37.670 A:middle L:90%
another example of noise here in the sense of pattern
456
00:33:39.740 --> 00:33:45.960 A:middle L:90%
. So what I have here is a bunch of
457
00:33:45.970 --> 00:33:50.960 A:middle L:90%
code that I'm going to gloss over really quickly but
458
00:33:51.740 --> 00:33:55.230 A:middle L:90%
supercollider as a full fledged computer music language allows you
459
00:33:55.230 --> 00:33:59.660 A:middle L:90%
to read in sounds you might have lying around your
460
00:33:59.670 --> 00:34:01.640 A:middle L:90%
hard drive and I do happen to have a bass
461
00:34:01.640 --> 00:34:06.210 A:middle L:90%
drum and snare drum want line wrong. And then
462
00:34:06.220 --> 00:34:12.869 A:middle L:90%
these symptoms definitions are basically ways of creating audio synthesis
463
00:34:12.869 --> 00:34:15.360 A:middle L:90%
graphs that can be repeated over and over again.
464
00:34:15.539 --> 00:34:17.530 A:middle L:90%
So, really nice thing about super flyer. All
465
00:34:17.530 --> 00:34:22.429 A:middle L:90%
right. Is that in a total pattern of synthesis
466
00:34:22.429 --> 00:34:25.769 A:middle L:90%
of sex can be made as a bunch of individual
467
00:34:25.769 --> 00:34:29.460 A:middle L:90%
notes that you can have a lot of control.
468
00:34:29.840 --> 00:34:34.389 A:middle L:90%
So this is this is creating some structure and 11
469
00:34:34.389 --> 00:34:36.619 A:middle L:90%
structure is just this place sound thing, which is
470
00:34:36.619 --> 00:34:38.090 A:middle L:90%
going to just play a sound in a towel supercollider
471
00:34:38.090 --> 00:34:40.460 A:middle L:90%
to play. And the other thing is what I'm
472
00:34:40.469 --> 00:34:45.659 A:middle L:90%
calling acid because is anybody familiar with acid house music
473
00:34:45.739 --> 00:34:49.510 A:middle L:90%
a little bit. Okay, so acid has a
474
00:34:49.519 --> 00:34:52.679 A:middle L:90%
lot to do with the idea of a tone that's
475
00:34:52.679 --> 00:34:57.719 A:middle L:90%
rich in harmonics that you sweep a filter over so
476
00:34:57.719 --> 00:35:00.889 A:middle L:90%
we can acquire that kind of self. So that
477
00:35:00.900 --> 00:35:05.360 A:middle L:90%
being said, I could haven't loaded this sound.
478
00:35:06.239 --> 00:35:09.099 A:middle L:90%
I can play it as much as I wanted to
479
00:35:09.099 --> 00:35:12.360 A:middle L:90%
and if I wanted to test the acid sound.
480
00:35:12.739 --> 00:35:16.079 A:middle L:90%
Sure. Okay. Yeah, yeah. So that's
481
00:35:16.079 --> 00:35:20.590 A:middle L:90%
that kind of sense of having a filter sweet down
482
00:35:20.599 --> 00:35:24.260 A:middle L:90%
on looks like a sergeant were. Okay. Now
483
00:35:25.139 --> 00:35:28.550 A:middle L:90%
, the next thing is going to be very similar
484
00:35:28.559 --> 00:35:32.719 A:middle L:90%
to what I did with the, with the noise
485
00:35:34.099 --> 00:35:37.519 A:middle L:90%
. I created a pattern, a random pattern and
486
00:35:37.519 --> 00:35:39.239 A:middle L:90%
then just repeated it over and over again. We
487
00:35:39.250 --> 00:35:45.980 A:middle L:90%
heard that pattern as as meaningful in some sense.
488
00:35:45.090 --> 00:35:47.610 A:middle L:90%
So the next thing I'm going to do is use
489
00:35:47.610 --> 00:35:52.199 A:middle L:90%
a few more patterns. And so this thing over
490
00:35:52.199 --> 00:35:59.269 A:middle L:90%
here, I think this time smaller, is this
491
00:35:59.579 --> 00:36:02.340 A:middle L:90%
still legible? More or less? That might be
492
00:36:02.929 --> 00:36:07.250 A:middle L:90%
mhm. Well, I can always go of and
493
00:36:07.250 --> 00:36:09.019 A:middle L:90%
illness. So this is a little bit more like
494
00:36:09.030 --> 00:36:13.750 A:middle L:90%
actual supercollider codes just generate things in real time.
495
00:36:14.230 --> 00:36:17.119 A:middle L:90%
And so what I'm doing is I'm setting a bpm
496
00:36:17.130 --> 00:36:22.630 A:middle L:90%
beach permitted as 130 and then I'm calculating an actual
497
00:36:22.639 --> 00:36:29.590 A:middle L:90%
beat as the actual duration beef is 60 Because if
498
00:36:29.590 --> 00:36:32.289 A:middle L:90%
we're looking at seconds bpm for 1/2 per b would
499
00:36:32.300 --> 00:36:38.820 A:middle L:90%
be 60. Okay, so essentially we get a
500
00:36:38.829 --> 00:36:42.309 A:middle L:90%
beat that's going to be somewhat faster than a second
501
00:36:42.320 --> 00:36:47.340 A:middle L:90%
, and then I divide to get the A 16
502
00:36:49.030 --> 00:36:51.550 A:middle L:90%
Or rather a quarter to be divided by four,
503
00:36:52.130 --> 00:36:53.469 A:middle L:90%
and this duration is just how many seconds to go
504
00:36:53.469 --> 00:36:59.849 A:middle L:90%
in. Now. A routine is basically a uh
505
00:37:00.530 --> 00:37:05.849 A:middle L:90%
a sequence that is able to be paused or and
506
00:37:05.860 --> 00:37:07.980 A:middle L:90%
or continue. So, so if I wanted to
507
00:37:07.989 --> 00:37:12.630 A:middle L:90%
, let's say regularly play beef, which is what
508
00:37:12.639 --> 00:37:14.599 A:middle L:90%
one often does in techno, all you have to
509
00:37:14.610 --> 00:37:17.449 A:middle L:90%
do is play a note, wait while play another
510
00:37:17.449 --> 00:37:20.809 A:middle L:90%
note, wait while play another note, Wait awhile
511
00:37:20.949 --> 00:37:24.460 A:middle L:90%
. So this routine basically does that. Mhm.
512
00:37:24.829 --> 00:37:28.670 A:middle L:90%
So then for the, the bass drum, I'm
513
00:37:28.670 --> 00:37:30.420 A:middle L:90%
going to now, the cool thing for those of
514
00:37:30.420 --> 00:37:34.869 A:middle L:90%
you who are fans, I can just comment our
515
00:37:34.869 --> 00:37:37.760 A:middle L:90%
code that I don't want to actually execute. So
516
00:37:37.760 --> 00:37:43.909 A:middle L:90%
I'm gonna comment out mm few of these routines and
517
00:37:43.909 --> 00:37:46.659 A:middle L:90%
then I can execute this part here. Okay.
518
00:37:46.670 --> 00:37:53.309 A:middle L:90%
Okay. So that's that is pretty simple code.
519
00:37:53.309 --> 00:37:57.849 A:middle L:90%
But what's happening here is remember that since definition that
520
00:37:57.849 --> 00:38:00.500 A:middle L:90%
I made a while back to play a noon this
521
00:38:00.500 --> 00:38:04.130 A:middle L:90%
place sound is just playing what's in the buffer of
522
00:38:04.139 --> 00:38:06.260 A:middle L:90%
that base drop. And I could replace that with
523
00:38:06.260 --> 00:38:09.389 A:middle L:90%
anything else. So you could think of supercollider potentially
524
00:38:09.420 --> 00:38:15.090 A:middle L:90%
as a drum machine that could have 1000 different drugs
525
00:38:15.090 --> 00:38:17.449 A:middle L:90%
out. That's that's the view of your program.
526
00:38:20.820 --> 00:38:22.130 A:middle L:90%
And all we do is we play the sound,
527
00:38:22.139 --> 00:38:27.780 A:middle L:90%
we send the message, wait two the being and
528
00:38:27.780 --> 00:38:32.559 A:middle L:90%
then I use this kind of updating of how far
529
00:38:32.559 --> 00:38:36.039 A:middle L:90%
we've gone in time just so that I can stop
530
00:38:36.050 --> 00:38:38.349 A:middle L:90%
them. We've got to a certain point because this
531
00:38:38.349 --> 00:38:40.849 A:middle L:90%
is a classic, wildly. It just says,
532
00:38:40.860 --> 00:38:45.750 A:middle L:90%
wow, now is less than the duration we run
533
00:38:45.750 --> 00:38:50.840 A:middle L:90%
this thing. Okay, so let me add in
534
00:38:50.840 --> 00:38:52.510 A:middle L:90%
another element. This is the snare drum has its
535
00:38:52.510 --> 00:39:00.230 A:middle L:90%
own pattern here. So now the patterns themselves are
536
00:39:00.239 --> 00:39:02.960 A:middle L:90%
global. So if I wanted to I could change
537
00:39:02.969 --> 00:39:13.539 A:middle L:90%
the pattern is being read, go for it.
538
00:39:15.320 --> 00:39:19.659 A:middle L:90%
Okay, now all of that is totally deterministic.
539
00:39:19.670 --> 00:39:22.449 A:middle L:90%
Let's add in the random part of this now,
540
00:39:22.119 --> 00:39:30.940 A:middle L:90%
which is essentially to take uh to fill an array
541
00:39:30.940 --> 00:39:36.239 A:middle L:90%
of 16 beach because we'll have Basically 16 16th notes
542
00:39:36.250 --> 00:39:40.639 A:middle L:90%
through the bar with a random pattern each and these
543
00:39:40.639 --> 00:39:45.260 A:middle L:90%
are frequencies, Each frequency will be between 100 and
544
00:39:45.269 --> 00:39:51.530 A:middle L:90%
500. Yeah, so I'm gonna comment out now
545
00:39:51.539 --> 00:39:54.760 A:middle L:90%
the drunk part of this and just do the bass
546
00:39:54.760 --> 00:40:05.539 A:middle L:90%
part every time I do this I'll get a different
547
00:40:05.539 --> 00:40:16.260 A:middle L:90%
pattern. Okay, now, if I really wanted
548
00:40:16.269 --> 00:40:20.909 A:middle L:90%
to keep things static, this is this is the
549
00:40:20.920 --> 00:40:24.130 A:middle L:90%
beauty of of the sort of pseudo random noise.
550
00:40:24.610 --> 00:40:29.139 A:middle L:90%
Uh huh. Each each of these routines has its
551
00:40:29.139 --> 00:40:32.519 A:middle L:90%
own thread so I can say this thread Grand C
552
00:40:34.210 --> 00:40:49.090 A:middle L:90%
equals some value. So did I? Very interesting
553
00:40:49.159 --> 00:40:54.260 A:middle L:90%
. Um I couldn't tell you why that was I
554
00:40:54.260 --> 00:40:57.349 A:middle L:90%
think I might have to put it outside routine but
555
00:40:57.349 --> 00:41:00.440 A:middle L:90%
we'll give we'll give that a try here. Um
556
00:41:00.710 --> 00:41:08.400 A:middle L:90%
see that you can hear that that's that's not painted
557
00:41:08.400 --> 00:41:10.900 A:middle L:90%
and no matter how many times I try this,
558
00:41:10.909 --> 00:41:19.170 A:middle L:90%
that particular random sequence is preserved. Okay, so
559
00:41:19.179 --> 00:41:21.619 A:middle L:90%
now let's bring the whole thing back in and then
560
00:41:21.619 --> 00:41:24.570 A:middle L:90%
you will get a kind of an acid uh results
561
00:41:30.900 --> 00:41:34.559 A:middle L:90%
and then I could change other aspects of the synthesis
562
00:41:39.980 --> 00:41:44.429 A:middle L:90%
. Okay, so you can see that this is
563
00:41:44.500 --> 00:41:50.369 A:middle L:90%
this is a combination of straight, straight up procedural
564
00:41:50.369 --> 00:41:53.889 A:middle L:90%
programming and real time sound. Some possess that you
565
00:41:53.889 --> 00:41:58.179 A:middle L:90%
can interact in a bunch of ways for the main
566
00:41:58.179 --> 00:42:02.059 A:middle L:90%
point is that we're using noise as a way of
567
00:42:02.070 --> 00:42:06.269 A:middle L:90%
getting two different kinds of results. So let's say
568
00:42:06.269 --> 00:42:07.550 A:middle L:90%
that I take out this RAMsi thing, I just
569
00:42:07.550 --> 00:42:15.119 A:middle L:90%
let it be whatever is going to be okay,
570
00:42:21.760 --> 00:42:23.920 A:middle L:90%
you can hear that's different every time. Great.
571
00:42:24.300 --> 00:42:29.090 A:middle L:90%
Yes, we're last. So I mean one of
572
00:42:29.099 --> 00:42:31.800 A:middle L:90%
, one of the classic definition of insanity is doing
573
00:42:31.800 --> 00:42:36.309 A:middle L:90%
the same thing time and expecting different results. But
574
00:42:36.320 --> 00:42:37.900 A:middle L:90%
actually that's exactly what we do when we're doing algorithm
575
00:42:37.900 --> 00:42:40.369 A:middle L:90%
. Mark, you have you have an algorithm.
576
00:42:40.380 --> 00:42:44.829 A:middle L:90%
But the introduction of noise means that the results actually
577
00:42:44.840 --> 00:42:46.329 A:middle L:90%
are going to be different every time you do the
578
00:42:46.329 --> 00:42:50.130 A:middle L:90%
same things. And yeah, just one question can
579
00:42:50.690 --> 00:42:53.079 A:middle L:90%
be applied to rhythms as well. Then you apply
580
00:42:53.449 --> 00:42:57.469 A:middle L:90%
rhythms in order to sort of. Absolutely yeah.
581
00:42:57.480 --> 00:43:01.199 A:middle L:90%
The thing is because you guys know programming can control
582
00:43:01.210 --> 00:43:05.750 A:middle L:90%
anything you want. So, so where I'm just
583
00:43:05.750 --> 00:43:07.840 A:middle L:90%
having this weight thing via costing, it could be
584
00:43:07.849 --> 00:43:12.630 A:middle L:90%
a random value. I could generate random patterns.
585
00:43:12.730 --> 00:43:15.019 A:middle L:90%
I could generate groups of patterns that repeat over and
586
00:43:15.019 --> 00:43:17.780 A:middle L:90%
over again. So, so yeah, the sky
587
00:43:17.789 --> 00:43:20.780 A:middle L:90%
, the sky is kind of limited. Now,
588
00:43:20.789 --> 00:43:23.719 A:middle L:90%
I'd like to look at a couple of other examples
589
00:43:23.730 --> 00:43:27.880 A:middle L:90%
really quickly of the application of noise because what we've
590
00:43:27.880 --> 00:43:30.730 A:middle L:90%
seen here is kind of very literal minded one.
591
00:43:31.199 --> 00:43:36.820 A:middle L:90%
But I also, in my own work tried to
592
00:43:36.829 --> 00:43:40.159 A:middle L:90%
use noise as a way of guiding other kinds of
593
00:43:40.170 --> 00:43:45.420 A:middle L:90%
procedures. So you'll recall that the, um,
594
00:43:47.300 --> 00:43:52.059 A:middle L:90%
that the killer that we heard was largely regimen and
595
00:43:52.070 --> 00:43:54.380 A:middle L:90%
, but did not sound right. So I'm going
596
00:43:54.380 --> 00:43:57.900 A:middle L:90%
to show you another example of that. This is
597
00:43:57.909 --> 00:44:00.000 A:middle L:90%
a piece I wrote in 2011 called Noise trip deck
598
00:44:00.010 --> 00:44:04.940 A:middle L:90%
, which had a few different ways of approaching the
599
00:44:04.940 --> 00:44:07.750 A:middle L:90%
composition with noise. So it's for two islands.
600
00:44:07.750 --> 00:44:10.199 A:middle L:90%
I'll play you a little bit of it here.
601
00:44:10.199 --> 00:44:42.860 A:middle L:90%
It yeah. Mhm Yeah. So does that sound
602
00:44:42.860 --> 00:44:49.349 A:middle L:90%
totally around you? Okay, so here go ahead
603
00:44:49.360 --> 00:44:53.710 A:middle L:90%
, defeating explaining evolving down which keeps flowing order,
604
00:44:53.719 --> 00:44:59.980 A:middle L:90%
which uh, they keep getting slightly slightly like there
605
00:44:59.980 --> 00:45:01.670 A:middle L:90%
are slight changes. But actually what I'm doing here
606
00:45:01.670 --> 00:45:06.579 A:middle L:90%
is something I call articulated noise, which is to
607
00:45:06.590 --> 00:45:09.420 A:middle L:90%
have to work in combination with the computer. So
608
00:45:09.420 --> 00:45:14.949 A:middle L:90%
there is some super flyer code here, which,
609
00:45:14.960 --> 00:45:16.539 A:middle L:90%
which I'll show you very, very quickly, a
610
00:45:17.289 --> 00:45:22.610 A:middle L:90%
bunch of different functions that do different things. And
611
00:45:22.989 --> 00:45:28.889 A:middle L:90%
as I'm proposing essentially this first, this, this
612
00:45:28.889 --> 00:45:31.480 A:middle L:90%
first thing has had to do with different kinds of
613
00:45:31.489 --> 00:45:36.909 A:middle L:90%
melodies That I would use. one was from a
614
00:45:36.909 --> 00:45:39.710 A:middle L:90%
book of north american folk songs. The second source
615
00:45:39.710 --> 00:45:44.090 A:middle L:90%
of melodies was a book of plain chant melodies,
616
00:45:44.099 --> 00:45:47.019 A:middle L:90%
Ecuadorian Chant melodies that go back to the 8th and
617
00:45:47.019 --> 00:45:51.239 A:middle L:90%
9th century. And the third was melodies that I
618
00:45:51.239 --> 00:45:52.400 A:middle L:90%
would have to make up myself. So as I
619
00:45:52.400 --> 00:45:55.610 A:middle L:90%
was composing section by section, I would just execute
620
00:45:57.289 --> 00:46:00.179 A:middle L:90%
on a very simple piece of code. And the
621
00:46:00.190 --> 00:46:02.869 A:middle L:90%
first thing that happens is doing anomaly. Because I
622
00:46:02.869 --> 00:46:06.280 A:middle L:90%
had built into this, the idea that whatever the
623
00:46:06.280 --> 00:46:08.199 A:middle L:90%
structure was, I might well composing, just ask
624
00:46:08.199 --> 00:46:12.250 A:middle L:90%
myself to do something that had absolutely nothing to do
625
00:46:12.260 --> 00:46:14.489 A:middle L:90%
with what else was going on in the piece.
626
00:46:14.500 --> 00:46:17.460 A:middle L:90%
But it gives me some other guidelines or rules.
627
00:46:17.460 --> 00:46:20.780 A:middle L:90%
So it says, okay, this this bar is
628
00:46:20.780 --> 00:46:22.039 A:middle L:90%
gonna be a 54 bar and then I have a
629
00:46:22.039 --> 00:46:27.139 A:middle L:90%
transition. Yeah, in terms of how I moved
630
00:46:27.139 --> 00:46:29.000 A:middle L:90%
to other things. So let's try this again.
631
00:46:30.679 --> 00:46:34.090 A:middle L:90%
A source melody is make it up. So that
632
00:46:34.090 --> 00:46:36.010 A:middle L:90%
means that I have after, in that case,
633
00:46:36.010 --> 00:46:38.510 A:middle L:90%
make up an original um kind of melody. It
634
00:46:38.510 --> 00:46:42.449 A:middle L:90%
tells me about the articulation, tells me how long
635
00:46:42.460 --> 00:46:45.530 A:middle L:90%
it should be and how I transition. So as
636
00:46:45.530 --> 00:46:50.030 A:middle L:90%
a result of this composing through this piece, some
637
00:46:50.030 --> 00:46:53.210 A:middle L:90%
melodies moved gradually to the others, some instantaneously.
638
00:46:53.320 --> 00:46:58.550 A:middle L:90%
And there's a huge amount of randomness at the high
639
00:46:58.550 --> 00:47:00.440 A:middle L:90%
level structure, but at the lower level structure,
640
00:47:00.530 --> 00:47:05.400 A:middle L:90%
I'm making decisions in a very, very determined fashion
641
00:47:06.980 --> 00:47:13.010 A:middle L:90%
. Uh I'll give you another example of this which
642
00:47:13.019 --> 00:47:16.690 A:middle L:90%
is the noise in prompted. Now the noise and
643
00:47:16.690 --> 00:47:21.690 A:middle L:90%
prompted is a guided in private improvisation. And now
644
00:47:21.690 --> 00:47:23.730 A:middle L:90%
we are going to see max MSP, which is
645
00:47:23.730 --> 00:47:29.690 A:middle L:90%
another, another kind of language. And the way
646
00:47:29.690 --> 00:47:37.119 A:middle L:90%
this works here is that basically the two violinists have
647
00:47:37.130 --> 00:47:42.599 A:middle L:90%
to two laptops, they're connected together over wifi,
648
00:47:43.179 --> 00:47:49.480 A:middle L:90%
and one of them essentially executes code to create a
649
00:47:49.489 --> 00:47:52.659 A:middle L:90%
piece that will guide their improvisations. So every time
650
00:47:52.659 --> 00:47:57.389 A:middle L:90%
they do this piece, the guiding of the improvisation
651
00:47:57.389 --> 00:47:59.269 A:middle L:90%
is going to be different and then they actually play
652
00:47:59.269 --> 00:48:04.940 A:middle L:90%
it in real time. So first of all,
653
00:48:04.949 --> 00:48:07.260 A:middle L:90%
I'll play you a little bit about what that sounds
654
00:48:07.260 --> 00:48:23.500 A:middle L:90%
like. Mhm. Mhm. Yeah. Mhm.
655
00:48:29.469 --> 00:48:39.280 A:middle L:90%
Mhm. Yeah. Yeah. Mhm. Mhm.
656
00:48:39.670 --> 00:49:15.280 A:middle L:90%
Mhm. Mhm. Yeah. 13. Mhm.
657
00:49:22.070 --> 00:49:23.679 A:middle L:90%
Right. Oh yeah, I just wanted you to
658
00:49:23.679 --> 00:49:28.360 A:middle L:90%
hear enough of that here. They're improvising. But
659
00:49:28.369 --> 00:49:30.119 A:middle L:90%
could you all hear that? There are different sections
660
00:49:30.119 --> 00:49:32.179 A:middle L:90%
where they're different improvising in different manners. So I'll
661
00:49:32.179 --> 00:49:36.260 A:middle L:90%
show you a little bit of what they're actually seeing
662
00:49:36.449 --> 00:49:37.599 A:middle L:90%
. And again, this is this is basically a
663
00:49:37.599 --> 00:49:42.989 A:middle L:90%
noise guided improvisation. So what's the starts? One
664
00:49:43.000 --> 00:49:45.329 A:middle L:90%
computer will communicate to the other. And so you're
665
00:49:45.329 --> 00:49:50.239 A:middle L:90%
seeing the two parts. The first violinist would only
666
00:49:50.250 --> 00:49:52.000 A:middle L:90%
be following this part in the second, this one
667
00:50:00.059 --> 00:50:06.199 A:middle L:90%
. So you could hear that. Then basically the
668
00:50:06.210 --> 00:50:12.079 A:middle L:90%
instructions are if one violence is going completely crazy,
669
00:50:12.760 --> 00:50:15.289 A:middle L:90%
the other has to follow. Some of the instructions
670
00:50:15.300 --> 00:50:16.780 A:middle L:90%
are the same so that this was at the beginning
671
00:50:16.780 --> 00:50:21.780 A:middle L:90%
occurred. These really scratching things that was that particular
672
00:50:23.059 --> 00:50:30.690 A:middle L:90%
instructions and the durations are totally grand and the sequencing
673
00:50:30.699 --> 00:50:34.429 A:middle L:90%
. So this person had to do what they're doing
674
00:50:34.440 --> 00:50:37.730 A:middle L:90%
. Mika is actually a reference to a piece by
675
00:50:37.730 --> 00:50:40.980 A:middle L:90%
Yannis Xenakis that involves lots of lasagna. So it's
676
00:50:40.980 --> 00:50:45.110 A:middle L:90%
a sort of shorthand, the person what I want
677
00:50:45.110 --> 00:50:45.840 A:middle L:90%
you now, this is funny, but one has
678
00:50:45.840 --> 00:50:50.099 A:middle L:90%
to go playing extremely quiet and the other person has
679
00:50:50.099 --> 00:50:52.360 A:middle L:90%
to go completely mental on that. So, so
680
00:50:52.360 --> 00:50:57.550 A:middle L:90%
this is a really interesting for in terms of guidance
681
00:50:57.550 --> 00:51:01.429 A:middle L:90%
improvisation, because just as much as composition improvisation,
682
00:51:01.429 --> 00:51:05.340 A:middle L:90%
even though it's very free, is very heavily conditioned
683
00:51:05.349 --> 00:51:07.969 A:middle L:90%
by stylistic constraints, habits what you can do on
684
00:51:07.969 --> 00:51:10.699 A:middle L:90%
the instrument. So if you have something like this
685
00:51:10.710 --> 00:51:16.420 A:middle L:90%
where you have the computer making random choices about how
686
00:51:16.610 --> 00:51:21.159 A:middle L:90%
each person should play and they're related to each other
687
00:51:21.349 --> 00:51:27.079 A:middle L:90%
. You might be forced into a really unusual situation
688
00:51:27.559 --> 00:51:29.750 A:middle L:90%
. And the structuring of the piece will be very
689
00:51:29.760 --> 00:51:34.500 A:middle L:90%
different than an improvisation that that's completely free. And
690
00:51:34.500 --> 00:51:38.280 A:middle L:90%
sometimes you get these really interesting things where 11 violinist
691
00:51:38.289 --> 00:51:42.989 A:middle L:90%
is instructed to imitate what the other violinist is doing
692
00:51:43.360 --> 00:51:46.019 A:middle L:90%
and the other violence at the same exact, you
693
00:51:46.019 --> 00:51:49.530 A:middle L:90%
know, instruction. So they're all listening each list
694
00:51:49.539 --> 00:51:52.739 A:middle L:90%
, trying to figure out what the other violinists playing
695
00:51:52.750 --> 00:51:55.340 A:middle L:90%
. It goes into this kind of uh you know
696
00:51:55.349 --> 00:52:00.900 A:middle L:90%
, an infinite loop that way. So this this
697
00:52:00.900 --> 00:52:04.980 A:middle L:90%
is an example of how you can you can use
698
00:52:04.980 --> 00:52:10.760 A:middle L:90%
noise in the context of applying basically computational thinking.
699
00:52:12.349 --> 00:52:22.980 A:middle L:90%
Uh huh That that was that's really great. Mm
700
00:52:24.449 --> 00:52:30.010 A:middle L:90%
But anyhow that did bring me to pretty much what
701
00:52:30.010 --> 00:52:31.389 A:middle L:90%
I want to share with you, which is the
702
00:52:31.389 --> 00:52:37.070 A:middle L:90%
idea that, you know, artistic practice can very
703
00:52:37.070 --> 00:52:43.349 A:middle L:90%
easily today incorporate computation. But also the idea of
704
00:52:43.349 --> 00:52:49.940 A:middle L:90%
noise is a very valuable input to artistic thinking that
705
00:52:49.940 --> 00:52:52.690 A:middle L:90%
can get you to places that you wouldn't necessarily be
706
00:52:52.690 --> 00:52:57.250 A:middle L:90%
able to get to by yourself. So thanks very
707
00:52:57.250 --> 00:52:58.619 A:middle L:90%
much for having me. And I'll be happy to
708
00:52:58.619 --> 00:53:13.639 A:middle L:90%
answer questions. So I guess the question is this
709
00:53:14.469 --> 00:53:19.980 A:middle L:90%
a turing test? Well, the turing test already
710
00:53:19.989 --> 00:53:24.150 A:middle L:90%
had, which is that the that random folk noise
711
00:53:24.150 --> 00:53:28.409 A:middle L:90%
piece that I played? I played it for,
712
00:53:28.420 --> 00:53:31.250 A:middle L:90%
you know, large numbers of people and yeah,
713
00:53:32.650 --> 00:53:36.610 A:middle L:90%
They hear his music. And then I say and
714
00:53:36.610 --> 00:53:39.269 A:middle L:90%
by the way that piece was 80 random and they
715
00:53:39.269 --> 00:53:45.889 A:middle L:90%
go what? Well, But is that because you
716
00:53:50.090 --> 00:53:54.239 A:middle L:90%
but a musician extra experts? Well, as I
717
00:53:54.239 --> 00:53:57.969 A:middle L:90%
said, I played it for musicians, you know
718
00:53:57.980 --> 00:54:00.559 A:middle L:90%
, in other words. And this is a really
719
00:54:00.570 --> 00:54:02.519 A:middle L:90%
interesting question. There is actually some research having to
720
00:54:02.530 --> 00:54:07.039 A:middle L:90%
do with how well we here. So because a
721
00:54:07.039 --> 00:54:10.269 A:middle L:90%
large amount of our theory about total music in the
722
00:54:10.280 --> 00:54:15.340 A:middle L:90%
european common practice has to do with the argument for
723
00:54:15.349 --> 00:54:20.840 A:middle L:90%
coherence, that you can just take one piece of
724
00:54:20.840 --> 00:54:23.170 A:middle L:90%
music and chop it into little bits and rearrange it
725
00:54:23.179 --> 00:54:28.489 A:middle L:90%
because there's a narrative structure and further all a total
726
00:54:28.489 --> 00:54:30.909 A:middle L:90%
structure, which is which almost looks a little bit
727
00:54:30.920 --> 00:54:37.550 A:middle L:90%
like uh like chomsky's linguistic analyses of spoken language.
728
00:54:37.940 --> 00:54:39.849 A:middle L:90%
So that shouldn't work. But there was some there
729
00:54:39.849 --> 00:54:45.170 A:middle L:90%
was some research back in the late 90s where they
730
00:54:45.739 --> 00:54:49.179 A:middle L:90%
sure no listeners who have been exposed to classical music
731
00:54:49.190 --> 00:54:52.179 A:middle L:90%
took a Mozart symphony, chopped it into individual sections
732
00:54:52.289 --> 00:54:55.760 A:middle L:90%
and rearranged in ways that were plausible and most people
733
00:54:55.760 --> 00:55:00.159 A:middle L:90%
didn't notice the difference. So I actually think that
734
00:55:00.170 --> 00:55:04.829 A:middle L:90%
you know the kinds of structural hearing our ears today
735
00:55:04.840 --> 00:55:07.849 A:middle L:90%
and our experience of culture I think is just completely
736
00:55:07.860 --> 00:55:12.820 A:middle L:90%
reprogrammed From what it used to be in the 18th
737
00:55:12.829 --> 00:55:15.159 A:middle L:90%
or 19th century. So we just don't seem to
738
00:55:15.170 --> 00:55:16.889 A:middle L:90%
be responding to that as much as we respond to
739
00:55:16.889 --> 00:55:22.929 A:middle L:90%
surfaces. So he's like this which is very plausible
740
00:55:22.960 --> 00:55:25.230 A:middle L:90%
in terms of its individual elements. And as you
741
00:55:25.230 --> 00:55:28.809 A:middle L:90%
notice in the coding pays a lot of attention to
742
00:55:28.809 --> 00:55:31.099 A:middle L:90%
how things transition, which is something that people do
743
00:55:31.099 --> 00:55:35.809 A:middle L:90%
notice resulted in a peace that if I can tell
744
00:55:35.809 --> 00:55:37.760 A:middle L:90%
people that it was random, nobody would have guessed
745
00:55:39.139 --> 00:55:40.849 A:middle L:90%
. So. So I mean that seems to be
746
00:55:40.849 --> 00:55:43.800 A:middle L:90%
about as close to the turing test if you're talking
747
00:55:43.800 --> 00:55:46.840 A:middle L:90%
about like can you tell the difference between algorithmic music
748
00:55:46.849 --> 00:55:51.530 A:middle L:90%
or music created by computer computer? But the funny
749
00:55:51.530 --> 00:55:54.659 A:middle L:90%
thing is, and I actually just posted this to
750
00:55:54.670 --> 00:55:58.789 A:middle L:90%
my twitter feed because I was just in a mood
751
00:55:58.800 --> 00:56:00.030 A:middle L:90%
. So I said that, you know, in
752
00:56:00.030 --> 00:56:02.690 A:middle L:90%
the 20th century you say isn't libraries and memories,
753
00:56:02.769 --> 00:56:06.719 A:middle L:90%
right. I don't know very many people are old
754
00:56:06.719 --> 00:56:07.639 A:middle L:90%
enough to remember what that means. But it's like
755
00:56:07.650 --> 00:56:09.909 A:middle L:90%
if you're listening in a room, could you tell
756
00:56:09.909 --> 00:56:15.250 A:middle L:90%
the difference between a live person singing or is it
757
00:56:15.260 --> 00:56:17.340 A:middle L:90%
memorex the team good enough. Is reproduction that you
758
00:56:17.349 --> 00:56:21.119 A:middle L:90%
could not hear? The difference between something being played
759
00:56:21.119 --> 00:56:24.960 A:middle L:90%
back speakers or played or some live All right.
760
00:56:25.929 --> 00:56:29.349 A:middle L:90%
Today, I think the question is, you know
761
00:56:29.360 --> 00:56:30.110 A:middle L:90%
, if you read a news article, was it
762
00:56:30.110 --> 00:56:32.170 A:middle L:90%
written by a human or written by a computer?
763
00:56:34.030 --> 00:56:37.280 A:middle L:90%
Because there's a big thing about how computer programs are
764
00:56:37.289 --> 00:56:40.659 A:middle L:90%
getting pretty good at writing news articles. So that
765
00:56:40.670 --> 00:56:45.389 A:middle L:90%
idea of distinguishing the human from from the the computer
766
00:56:45.389 --> 00:56:49.599 A:middle L:90%
program, I think it's getting hard, let's say
767
00:56:49.599 --> 00:56:51.920 A:middle L:90%
the computer programs are starting to get better and better
768
00:56:51.920 --> 00:56:54.510 A:middle L:90%
at fooling us. Yes. Since I I just
769
00:56:54.510 --> 00:56:57.670 A:middle L:90%
wanted to know, I've been a lot of times
770
00:56:57.690 --> 00:57:00.639 A:middle L:90%
feel like I can't describe what I feel is when
771
00:57:00.639 --> 00:57:02.150 A:middle L:90%
I go back. Yes, it's hard to describe
772
00:57:02.159 --> 00:57:05.510 A:middle L:90%
. But then going back here, does it again
773
00:57:05.519 --> 00:57:07.289 A:middle L:90%
bring me to the point that what I see as
774
00:57:07.289 --> 00:57:09.539 A:middle L:90%
speed is really is a condition of the programming of
775
00:57:09.550 --> 00:57:12.949 A:middle L:90%
my brain that has they just learned over time.
776
00:57:13.230 --> 00:57:17.070 A:middle L:90%
And then uh so it is just algorithm, is
777
00:57:17.070 --> 00:57:19.940 A:middle L:90%
there nothing we feel at all? And it's just
778
00:57:20.070 --> 00:57:22.150 A:middle L:90%
that question for you, there was definitely a sound
779
00:57:22.150 --> 00:57:25.530 A:middle L:90%
of algorithmic music. I mean if you sort of
780
00:57:25.539 --> 00:57:30.539 A:middle L:90%
make a lot of random decisions and you know,
781
00:57:30.550 --> 00:57:32.230 A:middle L:90%
there's there's a there's a feel of that, I
782
00:57:32.230 --> 00:57:35.909 A:middle L:90%
mean just as the feel of that sort of machine
783
00:57:35.920 --> 00:57:39.170 A:middle L:90%
regularity of that's not a human rhythm, it's an
784
00:57:39.179 --> 00:57:42.239 A:middle L:90%
inhuman rhythm. But we seem to like it.
785
00:57:42.250 --> 00:57:44.530 A:middle L:90%
You know, maybe the same reason people like watching
786
00:57:44.539 --> 00:57:46.159 A:middle L:90%
terminator, you know, there's, you know,
787
00:57:46.170 --> 00:57:50.619 A:middle L:90%
the humanness of it has a certain kind of appeal
788
00:57:50.630 --> 00:57:58.159 A:middle L:90%
to but and I think we're probably going is really
789
00:57:58.530 --> 00:58:01.239 A:middle L:90%
bridging that gap. And yeah, I mean the
790
00:58:01.239 --> 00:58:06.050 A:middle L:90%
second you put on glasses, your cybernetic organism sense
791
00:58:06.090 --> 00:58:08.199 A:middle L:90%
. And so I think that kind of integration of
792
00:58:08.210 --> 00:58:12.639 A:middle L:90%
technology is just going to keep on, keep on
793
00:58:12.650 --> 00:58:14.940 A:middle L:90%
going. And of course, yes, there are
794
00:58:14.940 --> 00:58:17.280 A:middle L:90%
lots of programs that you can program feel into the
795
00:58:17.280 --> 00:58:30.789 A:middle L:90%
drum machines do other questions, speaker speaker