17:37
TEDxStanford

Ge Wang: The DIY orchestra of the future

Filmed:

Ge Wang makes computer music, but it isn’t all about coded bleeps and blips. With the Stanford Laptop Orchestra, he creates new instruments out of unexpected materials—like an Ikea bowl—that allow musicians to play music that’s both beautiful and expressive.

- Music technologist
Both a musician and a computer scientist, Ge Wang turns ordinary MacBooks and iPhones into complex instruments. Full bio

I want to talk to you about one thing
00:12
and just one thing only,
00:15
and this has to do with when people ask me,
00:16
what do you do?
00:20
To which I usually respond,
00:23
I do computer music.
00:25
Now, a number of people
00:28
just stop talking to me right then and there,
00:30
and the rest who are left usually have
00:33
this blank look in their eye,
00:35
as if to say, what does that mean?
00:37
And I feel like I'm actually depriving them
00:39
of information by telling them this,
00:42
at which point I usually panic
00:45
and spit out the first thing that comes to my mind,
00:47
which is, I have no idea what I'm doing.
00:48
Which is true.
00:52
That's usually followed by a second thought,
00:54
which is, whatever it is that I'm doing,
00:56
I love it.
01:00
And today, I want to, well,
01:01
share with you something I love,
01:05
and also why.
01:07
And I think we'll begin with just this question:
01:09
What is computer music?
01:13
And I'm going to try to do my
best to provide a definition,
01:14
maybe by telling you a story
01:18
that goes through some of the stuff
01:20
I've been working on.
01:21
And the first thing, I think, in our story
01:23
is going to be something called ChucK.
01:25
Now, ChucK is a programming language for music,
01:28
and it's open-source, it's freely available,
01:33
and I like to think that it crashes equally well
01:36
on all modern operating systems.
01:38
And instead of telling you more about it,
01:40
I'm just going to give you a demo.
01:43
By the way, I'm just going to nerd out
01:45
for just a few minutes here,
01:46
so I would say, don't freak out.
01:47
In fact, I would invite all of you to join me
01:50
in just geeking out.
01:52
If you've never written a line
of code before in your life,
01:54
do not worry.
01:57
I'll bet you'll be able to come along on this.
01:58
First thing I'm going to do is to make
02:01
a sine wave oscillator,
02:03
and we're going to called the sine wave generator
02:05
"Ge."
02:08
And then we're going to connect "Ge" to the DAC.
02:10
Now this is kind of the abstraction
02:12
for the sound output on my computer. Okay?
02:14
So I've connected myself into the speaker.
02:16
Next, I'm going to say my frequency
02:19
is 440 hertz,
02:21
and I'm going to let time advance
02:23
by two seconds through this operation.
02:27
All right, so if I were to play this --
02:30
(Tone) —
02:34
you would hear a sine wave
at 440 hertz for two seconds.
02:36
Okay, great. Now I'm going to copy and paste this,
02:39
and then just change some of these numbers,
02:42
220.5, 440 I shall leave it as that,
02:43
and .5 and 880.
02:47
By doubling the frequency,
02:50
we're actually going up in successive octaves,
02:51
and then we have this sequence --
02:54
(Tones) — of tones.
02:55
Okay, great, now I can imagine creating
02:57
all kinds of really horrible
02:59
single sine wave pieces of music with this,
03:01
but I'm going to do something
that computers are really good at,
03:03
which is repetition.
03:06
I'm going to put this all in a while loop,
03:08
and you actually don't need to indent,
03:09
but this is purely for aesthetic reasons.
03:11
It's good practice.
03:13
And when we do this —
03:14
(Tones) —
03:16
that's going to go on for a while.
03:20
In fact, it's probably not going to stop
03:22
until this computer disintegrates.
03:23
And I can't really empirically prove that to you,
03:25
but I hope you'll believe me when I say that.
03:28
Next, I'm going to replace this 220
03:31
by math.random2f.
03:33
I'm going to generate a random number
03:36
between 30 and 1,000 and send that
03:37
to the frequency of me.
03:40
And I'm going to do this every half a second.
03:42
(Tones)
03:45
Let's do this every 200 milliseconds.
03:48
(Tones)
03:50
One hundred.
03:55
(Tones)
03:56
All right.
03:58
At this point, we've reached something
03:59
that I would like to think of as
04:01
the canonical computer music.
04:02
This is, to me, the sound that mainframes
04:06
are supposed to be making
04:10
when they're thinking really hard.
04:11
It's this sound, it's like,
04:14
the square root of five million.
04:16
So is this computer music?
04:22
Yeah, I guess by definition,
04:26
it's kind of computer music.
04:28
It's probably not the kind of music you would listen to
04:29
cruising down the highway,
04:32
but it's a foundation of computer-generated music,
04:33
and using ChucK,
04:38
we've actually been building instruments
04:41
in the Stanford Laptop Orchestra,
04:42
based right here at Stanford Center for
Computer Research in Music and Acoustics.
04:44
Now the Laptop Orchestra is an ensemble of laptops,
04:48
humans and special hemispherical speaker arrays.
04:51
Now the reason we have these
04:54
is so that for the instruments that we create
04:55
out of the laptop, we want the sound to come
04:58
out of somewhere near the instrument
05:00
and the performer,
05:03
kind of much like a traditional, acoustic instrument.
05:04
Like, if I were to play a violin here,
05:07
the sound would naturally not come out of
05:09
the P.A. system, but from the artifact itself.
05:11
So these speakers are meant to emulate that.
05:13
In fact, I'm going to show you
05:17
how we actually built them.
05:18
The first step is to go to IKEA
05:21
and buy a salad bowl.
05:23
This is an 11-inch Blanda Matt.
05:24
That's the actual name,
05:27
and I actually use one of these
05:29
to make salad at home as well, I kid you not.
05:30
And the first step is you turn it upside down,
05:33
and then you drill holes in them,
05:36
six holes per hemi,
05:38
and then make a base plate,
05:41
put car speaker drivers in them
05:42
along with amplifiers in the enclosure,
05:45
and you put that all together and you have
05:48
these hemispherical speaker arrays.
05:49
Add people, add laptops,
05:51
you have a laptop orchestra.
05:52
And what might a laptop orchestra sound like?
05:55
Well, let me give you a demonstration
05:59
of about 200 instruments we've created so far
06:01
for the Laptop Orchestra.
06:04
And what I'm going to do is
actually come over to this thing.
06:07
This thing I have in front of me
06:09
actually used to be a commodity gaming controller
06:12
called a Gametrak.
06:15
This thing actually has a glove
you can put on your hands.
06:17
It's tethered to the base,
06:19
and this will track the position of your hands
06:20
in real time.
06:23
It was originally designed as a golfing controller
06:24
to detect the motion of your swing.
06:27
That turned out to be a rather large
06:29
commercial non-success,
06:31
at which point they slashed prices to 10 dollars,
06:33
at which point computer music researchers
06:37
said, "This is awesome!
06:39
We can prototype instruments out of this."
06:41
So let me show you one instrument we've created,
06:44
one of many, and this instrument
06:47
is called "Twilight,"
06:49
and it's meant to go with this metaphor
06:50
of pulling a sound out of the ground.
06:53
So let me see if this will work.
06:56
(Music)
06:59
And put it back.
07:08
And then if you go to the left,
07:10
right,
07:14
it sounds like an elephant in pain.
07:21
This is a slightly metallic sound.
07:25
Turn it just a bit.
07:28
(Music)
07:31
It's like a hovering car.
07:37
Okay.
07:44
This third one is a ratchet-like interaction, so
07:45
let me turn it up.
07:51
(Music)
07:56
So it's a slightly different interaction.
08:02
The fourth one is a drone.
08:03
(Music)
08:07
And finally, let's see,
08:20
this is a totally different interaction,
08:23
and I think you have to imagine that there's
08:25
this giant invisible drum sitting right here on stage,
08:26
and I'm going to bang it.
08:29
(Drum)
08:34
(Laughter)
08:36
So there we go, so that's one of many instruments
08:40
in the Laptop Orchestra.
08:42
(Applause)
08:44
Thank you.
08:48
And when you put that together,
08:50
you get something that sounds like this.
08:52
(Music)
08:56
Okay, and so, I think from the experience
09:45
of building a lot of instruments
for the Laptop Orchestra,
09:47
and I think from the curiosity of wondering,
09:50
what if we took these
09:52
hopefully expressive instruments
09:54
and we brought it to a lot of people,
09:55
plus then a healthy bout of insanity —
09:58
put those three things together —
10:00
led to me actually co-founding a startup company
10:02
in 2008 called Smule.
10:05
Now Smule's mission is to create
10:08
expressive, mobile music things,
10:10
and one of the first musical instruments
10:14
we created is called Ocarina.
10:18
And I'm going to just demo this for you real quick.
10:20
So Ocarina —
10:27
(Music) —
10:29
is based on this ancient flute-like instrument
10:33
called the ocarina,
10:36
and this one is the four-hole
English pendant configuration,
10:37
and you're literally blowing into the microphone
10:41
to make the sound.
10:44
And there's actually a little ChucK script
10:47
running in here that's detecting
10:49
the strength of your blowing
10:50
and also synthesizing the sound.
10:51
(Music)
10:53
And vibrato is mapped to the accelerometer,
10:57
so you can get —
11:00
(Music)
11:02
All right. So let me play a little ditty for you,
11:05
a little Bach.
11:11
And here, you'll hear a little
accompaniment with the melody.
11:13
The accompaniment actually follows the melody,
11:15
not the other way around.
11:18
(Music)
11:22
And this was designed
11:35
to let you take your time
11:36
and figure out where your expressive space is,
11:38
and you can just hang out here
11:41
for a while, for a really
dramatic effect, if you want,
11:43
and whenever you're ready —
11:46
(Music)
11:49
And on these longer notes,
12:04
I'm going to use more vibrato
12:06
towards the end of the notes
12:08
to give it a little bit more of an expressive quality.
12:09
(Music)
12:12
Huh, that's a nice chord to end this excerpt on.
12:19
(Applause)
12:23
Thank you.
12:26
So I think a good question to ask about Ocarina is,
12:31
is this a toy or it an instrument? Maybe it's both,
12:35
but for me, I think the more important question is,
12:40
is it expressive?
12:42
And at the same time, I think
12:44
creating these types of instruments
12:46
asks a question about the role of technology,
12:47
and its place for how we make music.
12:50
Apparently, for example,
12:52
not that long ago, like only a hundred years ago —
12:54
that's not that long in the course of human history —
12:56
families back then
12:57
used to make music together
12:59
as a common form of entertainment.
13:01
I don't think that's really happening
13:04
that much anymore.
13:05
You know, this is before radio, before recording.
13:07
In the last hundred years, with all this technology,
13:09
we now have more access to music
13:11
as listeners and consumers,
13:13
but somehow, I think we're making less music
13:15
than ever before.
13:18
I'm not sure why that would be.
13:19
Maybe it's because it's too easy just to hit play.
13:21
And while listening to music is wonderful,
13:24
there's a special joy to making music
13:26
that's all its own.
13:29
And I think that's one part
13:31
of the goal of why I do what I do
13:32
is kind of to take us back to the past a little bit. Right?
13:34
Now, if that's one goal, the other goal
13:39
is to look to the future and think about
13:41
what kind of new musical things can we make
13:42
that we don't perhaps yet have names for
13:45
that's enabled by technology, but ultimately
13:47
might change the way that humans make music.
13:50
And I'll just give you one example here,
13:53
and this is Ocarina's other feature.
13:55
This is a globe,
14:01
and here you're actually listening
14:03
to other users of Ocarina
14:06
blow into their iPhones to play something.
14:08
This is "G.I.R." from Texas,
14:12
"R.I.K." I don't know why it's these
three-letter names today, Los Angeles.
14:17
They're all playing pretty,
14:23
somewhat minimal music here.
14:28
(Music)
14:31
And the idea with this is that, well,
14:37
technology should not be foregrounded here,
14:39
and — (Laughter) —
14:43
we've actually opened this up.
14:45
The first thought is that, hey, you know
14:48
there's somebody somewhere out there
14:50
playing some music,
14:53
and this is a small but I think important
14:54
human connection to make
14:56
that perhaps the technology affords.
14:57
As a final example,
15:00
and perhaps my favorite example,
15:03
is that in the wake of the 2011 earthquake
15:04
and tsunami disaster in Japan,
15:07
a woman reached out in one of our singing apps
15:10
to try to get people to join in to sing with her
15:13
on a version of "Lean on Me."
15:16
Now, in these apps, there's this thing that allows
15:17
any user to add their voice
15:20
to an existing performance by any other user
15:22
or group of users,
15:25
so in some sense, she's created this kind of
15:26
global ad hoc corral of strangers,
15:28
and within weeks, thousands of people
15:31
joined in on this,
15:33
and you can kind of see people
coming from all around the world
15:35
and all these lines converging on the origin
15:37
where the first rendition of the song was sung,
15:40
and that's in Tokyo.
15:42
And this is what it sounds like
when there's 1,000 people.
15:44
This is 1,000 voices.
15:47
(Recording) ♪ Sometimes in our lives ♪
15:49
♪ We all have pain, we all have sorrow ♪
15:53
♪ But if we are wise ♪
16:01
♪ We know that there's always tomorrow ♪
16:05
♪ Lean on me ♪
16:10
♪ When you're not strong ♪
16:13
♪ And I'll be your friend ♪
16:16
♪ I'll help you carry on ♪
16:18
♪ For it won't be long ♪
16:23
♪ Till I'm gonna need ♪
16:27
♪ Somebody to lean on ♪
16:29
♪ Just lean on — ♪
16:33
Is this computer music?
16:35
(Applause)
16:37
Was that computer music?
16:42
Yeah, I guess so; it's something that you really
16:44
couldn't have done without computers.
16:46
But at the same time, it's also just human,
16:49
and I think what I've essentially answered so far
16:53
is maybe why I do the stuff that I do,
16:56
and let's just finally return to the first question:
16:59
What is computer music?
17:02
And I think that the catch here is that,
17:04
at least to me, computer music
17:06
isn't really about computers.
17:07
It is about people.
17:09
It's about how we can use technology
17:11
to change the way we think
17:12
and do and make music,
17:14
and maybe even add to how we can
17:16
connect with each other through music.
17:19
And with that, I want to say,
17:22
this is computer music, and thank you for listening.
17:24
(Applause)
17:27

▲Back to top

About the Speaker:

Ge Wang - Music technologist
Both a musician and a computer scientist, Ge Wang turns ordinary MacBooks and iPhones into complex instruments.

Why you should listen

Ge Wang explores the intersection of technology and music, researching how programming languages and interactive software systems can push computer music from coded beeps and tones to something that musicians can actively play in the moment. An assistant professor at Stanford University, Wang is the founding director of both the Stanford Laptop Orchestra (SLOrk) and the Stanford Mobile Phone Orchestra (MoPho). 

Wang is also the man behind ChucK, a programming language built specifically for sound. He also co-founded Smule, a startup dedicated to musical apps that let people around the world make beautiful music … and connect with each other in the process. Their app Ocarina turns your iPhone into an ancient flute, for example, and Guitar! lets you strum your phone while a karaokist somewhere in the world provides lyrical support.

More profile about the speaker
Ge Wang | Speaker | TED.com