Get Inspire insights from former attendees in our AMA discussion thread on Inspire Buzz. ACEs and other community members are on call all week to answer!

Alter Everything

A podcast about data science and analytics culture.
Episode Guide

Interested in a specific topic or guest? Check out the guide for a list of all our episodes!

Alteryx Alumni (Retired)

For this episode of Alter Everything, music enthusiast turned Alteryx ACE Andy Uttley takes us on a journey through the correlations between math and music.

This is our last episode of 2019, and we want to hear your thoughts on what you’d like to hear featured in 2020, so be sure to share your two cents using our survey here.

Next year, we’re rolling out some exciting new theme music submitted by our listeners, as well as new podcast artwork, and lots of extra bonus content! So stay tuned, tweet with us using #AlterEverythingPodcast, and as always, thank you for listening!






Andy Uttley - @andyuttley, LinkedIn, Twitter
Maddie Johannsen - @MaddieJ, LinkedIn, Twitter



Golden Ratio diagramGolden Ratio diagram


Photo credit: monalisa.orgPhoto credit:

Episode playlist

Check out the songs in this playlist below for examples of the concepts Andy covers.






Episode Transcription

ANDY: 00:05

[music] Anyway, this is why we need two of us.

MADDIE: 00:05

Our voices do sound kind of echoey down here, so maybe we won’t talk--

ANDY: 00:09

And should I leave the volume as it is?

MADDIE: 00:10

Yeah. The volume is fine.

ANDY: 00:11

Okay. So I'm going to have to talk to teach you, but obviously you can cut it out.

MADDIE: 00:13

Okay. Yeah, yeah.

ANDY: 00:15

[music] So these are your notes.

MADDIE: 00:20


ANDY: 00:23

[music] So if you put those fingers on.

MADDIE: 00:26

Wait. [music] What is it? [music] No.

ANDY: 00:42

Back down, that [one again?].

MADDIE: 00:42

Oh, that one. Okay.

ANDY: 00:45

Yeah, nice. Okay. So you just play that over and over. [music] Keep just playing it. [music]

MADDIE: 00:57

I'm Maddie Johannsen and this is Alter Everything, a podcast about data science and analytics culture. Today, music enthusiast turned data enthusiast, Andy Uttley, tries to teach me a simple melody on the piano, and how it all relates to analytics. [music]

MADDIE: 01:20

So you're here to talk to me about music?

ANDY: 01:24

I am, yes.

MADDIE: 01:24

And you're very knowledgeable of music because you used to be a musician?

ANDY: 01:28

Yes, once upon a time.

MADDIE: 01:30

And a music teacher?

ANDY: 01:31


MADDIE: 01:32

After studying music at Durham University, Andy wound up landing an internship with the London Symphony Orchestra. Then he began his career as a teacher through England's Teach First program.

ANDY: 01:43

So my actual classroom teaching, that was in secondary schools, which in the UK is I think age-- I should know their ages [laughter]. They were about 11, 11 to 18, I think. So all the way up to just when you go to uni, but obviously at age 11 they're just out of primary school, and I guess it's a kind of big school there. So, yeah. I'd really focus on that in lessons, and how could we use technology and start composing through computers, doing things like writing rap music, and writing pieces like that, as well as focusing on string quartets and things like that.

MADDIE: 02:22

Andy now works for Javelin Group in London and is a newly crowned Alteryx Ace.

ANDY: 02:28

If I could have chosen anything to be, it would have been a film composer. To be honest, if you ask me now, I'd probably still say the same. But I went and studied-- studied music at Durham University, and it's really training for how to be a musician in the wider world. That was far too sociable for me [laughter]. So I went down the sort of more theoretical side of things, which did lend itself well to being the film composer I'm not today. So that's much more around the theory of music.


So I still do some performing. So I'm a guitarist mostly, but like most musicians, you end up playing a lot of different instruments badly. If I had to say the one that I'm least bad at, it would be guitar because that's the one I've played since being a child. But yes. So I grew up playing music. My brother is a musician, a professional pianist. He did have a dream in music and follow it and is doing it very well. Mine didn't last quite so long, unfortunately.

MADDIE: 03:33

I don't know. This is your moment to shine.

ANDY: 03:36

Yeah. Who knows if there's any record producers listening now to whatever's about to come…

MADDIE: 03:43

We will put your full details and the show notes.

ANDY: 03:47

Please do that. And the black and white picture of me, looking just past the camera.

MADDIE: 03:48

Yeah. Exactly [laughter].

ANDY: 03:50

Anyway, so that's by the by.

MADDIE: 03:53

I feel like a lot of people do that actually. One of our old podcast guests, Andrew Derbak, was in an episode, and he also was a musician. And then he found his way to [inaudible] to use. So there seems to be overlapping themes with analytics in music.

ANDY: 04:09

I think a lot of people see that and say it because there are just so many occurrences of it where you see those kind of themes. So the A levels or two of the A levels I did were the music and maths, and everybody--

MADDIE: 04:24

What's an A level?

ANDY: 04:25

So A level, sorry, is what you do just before university, so between age 16 and 18. So you choose your favorite sort of two or three subjects--

MADDIE: 04:33

Like your majors?

ANDY: 04:35

Yeah, I guess majors. And then almost always, but not always, that's then what you'll go on to--one of those is what you'll then study at university so you can hone it down. And everybody that was in my music class--actually, everybody apart from one, and I think there was 16 of us in my music class, also was in my maths class. So there's lots of links like that. When I was teaching as well, I was asked one day, if you're ever not teaching music, is there any other subject you'd be happy for us to put you in? And whilst as a music teacher, you end up teaching drama and dance [laughter] sometimes which if anyone's ever seen my dance at Inspire conferences, will know it'd be fairly shocking to the eye. That's all that. So but when I was asked, it was like, "Math." So that was the thing that I would happily try a go teach as well. And I didn't do too much of it, to be honest, but I did cover the [inaudible], and I really enjoyed that. But I was always interested in sort of numbers, as a lot of people are that are musical. It is quite numbers and logical, a lot of music.

MADDIE: 05:40

For the first stop on this tour of exploring analytics and music, Andy introduces the golden ratio and the mathematical number, phi, spelled P-H-I.

ANDY: 05:50

So I guess I would say numbers are--people say, "Oh, numbers are everywhere and there’s kind of logic everywhere around us." And one of the key things that always comes up if you do any sort of research into that is the golden ratio, which even came up briefly in music. I know you studied art, and I'm sure it came up in that as well. You'll see it, I think as far back as the Fifth Century.

Andy narrating: It actually dates back to the Fifth Century B.C. So, I was almost right, if we allow a margin of error of just 1000 years.

It was [inaudible] so that people would see it in geometry or in nature and architecture. And whilst, I guess, we'll have some slide notes that have some examples, some presentation--what were we calling this?

MADDIE: 06:45

Well, we'll do a--

ANDY: 06:47

Podcast note [laughter].

MADDIE: 06:48

Yeah, [laughter], show notes.

ANDY: 06:48

Slide note.

MADDIE: 06:50

Slide notes [laughter].

ANDY: 06:51

Call them sliders.

MADDIE: 06:55

Yeah, we'll have some show notes on the community, on the Alteryx community.

ANDY: 07:00

So I guess we can sort of describe to anyone who isn't aware of what that is, but it's kind of a good place to start because as we go a bit deeper and talk about music and numbers in music, this theme and some of the other themes that have links to it, like some of the numeric series that we're going to speak about as well, appear all over music, today or even if we go all the way back to just how we interpret pitches and things like that, pitches in terms of notes, not pictures.

MADDIE: 07:30

It does sound like you're saying pictures, like photos.

ANDY: 07:34

Yeah, also that, it is in some photos, so it works on every level. So yeah, I guess if you take the golden rectangle--so rectangle obviously, two sides of different lengths. So if we say side A and side B, that rectangle would be a golden rectangle. If the ratio between those two sides was phi, so phi being P-H-I, we'll refer to them as like often. So like obviously pi as well

MADDIE: 08:06

You're saying that the rectangle, you can basically cut it into--

ANDY: 08:11

So if it's this perfect sort of golden rectangle, the relationship between those two sides, the side A divided by side B, would be phi which is roughly 1.618. And that number goes on and on and on. But the reason that that seemingly meaningless number makes it a golden rectangle is if we were to then chop that rectangle into a square, so it'd leave a big square let's say on the left-hand side and a narrow rectangle on the right-hand side, that rectangle would also be a golden rectangle. And we can keep--

MADDIE: 08:46

You would keep the same ratio.

ANDY: 08:46

Ratio. Yeah. So we can keep chopping it up and chopping it up and so on and so forth. So it kind of loops into itself. And whilst that's kind of an interesting concept, the thing that actually makes it relevant and most interesting, I guess, is that that concept is what appears all over in nature and in music and in art and places like that. If we're to join the corners of all those rectangles and rectangles, all of the same proportion to each other, we get the kind of golden spiral which I suspect you know a lot more about than me, so.

MADDIE: 09:21

Yeah. Yeah. They say that you can see that same sort of spiral in the composition of the Mona Lisa, for example. It's interesting because with that or in some Greek temples you can see it in there too. But there's arguments that that wasn't necessarily the intent when they were making those pieces of art but it just happens to be that because it's an inherent human feeling that we're trying to find order or trying to make it look nice or have it make sense.

ANDY: 9:56

Yeah. And it's a really interesting debate because that sort of making sense then feeds into sort of the aesthetics of actually what is nice and what isn't. And whilst we can talk about nature and say a flower that may grow in a way that is particularly efficient for its seeds to be spaced around its center point or something like that. When we talk about the Mona Lisa or-- actually there's been analysis into how beautiful we might think that a person is. You'll see people drawing these spirals on people's faces.

MADDIE: 10:25

Yeah. And they're like, "Is this person beautiful?"

ANDY: 10:27

Exactly. Yeah. And you kind of going so far that way that you're getting into aesthetics. So if we think about nature, one example that comes up a lot is the Fibonacci sequence. Which I think actually having done a bit of research was analysis about the growth of a rabbit population. So the sequence will start zero, one, one, two, three. And the point of the Fibonacci sequence is the next number in the series is the sum of the two before it. So it's a fairly simple-- you start with zero and one. Add those together you get one. Add one and one you get two. Two and one you get three. And obviously those numbers just grow and grow and it's sort of an infinite series. The reason that is interesting is those numbers appear an awful lot in nature, which I guess we'll talk about in a second. But the relationship between each number to the one before it or after it in the series is the same as what we were just discussing with the rectangle. So you will get phi or the 1.618 when you divide one by the other. Or at least it'll get closer and closer to that as we go through the series. So it'll kind of dot either side of it getting gradually closer and closer to that sort of magic number, as it were so that's one occurrence of it. But those numbers there, that sort of growth of that numeric series, is actually much more recently there's been studies into the way we see that in nature. So it's everywhere, it's all around us apparently. But can be anything from sort of how flowers grow out and how nature is obviously so efficient in the way it's evolved, I guess, over time. But if you think of, say, seeds in the middle of a sunflower, as that sunflower grows out, the sort of count of, say, petals or seeds, we'll often spot these numbers that are in this Fibonacci sequence. So not necessarily ones and twos and threes that you might think they're quite common numbers, we don't see three sheep in a field and say, "Oh that's the Fibonacci sequence [inaudible] [laughter]."

MADDIE: 12:39

It's the [golden?] sheep.

ANDY: 12:41

Yeah. Yes [laughter]. The golden sheep, yeah. That's just some farmer's put three sheep in a field.

MADDIE: 12:45


ANDY: 12:46

But we could [inaudible]--

MADDIE: 12:47

I don't know, yeah.

ANDY: 12:48

--who knows--

MADDIE: 12:48

Who knows?

ANDY: 12:49

--there might be something at play here that we have just uncovered with the golden sheep. But much more sorts with the higher number. So you get up to 21, 34, 52, they're really common numbers in nature. There is this underlying thing that our minds are always kind of searching for this order, for example. And that's completely true, like I was saying, about beauty with people [that are?] in music as well. And we were saying the other day, you and I might share similar tastes in certain bands or something like that, but equally, we might have completely opposite tastes about something else. And it's perfectly fine that people don't like certain bits or parts of music or even certain genres. And as a teacher I was, hopefully, very clear to stress that. There's no right or wrong to liking something, in my opinion. Maybe the golden ratio would say otherwise [laughter].

MADDIE: 13:46

The golden sheep, they determine [laughter].

ANDY: 13:49

Yeah. They decide all, and they decided Beethoven was good. But yeah, so I was very careful, hopefully, to say that there's no right or wrong, however, I think it probably would be agreed - you might divide it up into, maybe, western music here, as opposed to globally, the reasons for which I can explain in a second - that whilst we might differ on genres or bands or anything like that, there are certain combinations and orders of notes, either in a melody or together, which would be a cord, so multiple notes at once, that, I would say, unquestionably, you and I would think, "Oh. That sounds nice." And, "Oh. That doesn't sound so--

ANDY: 14:36

That sounds weird, yeah.

MADDIE: 14:38

--nice." And it's the kind of playing around with that that is what composers do so well. Whether they know it or not, and I would guess the professional composers do know that--

MADDIE: 14:49

So if composers are looking at their work in an analytical way, [they are?] trying to bring logic into their pieces. I wanted to hear more about famous musicians who have worked in math and visa versa.

ANDY: 15:00

--if you look into it, Einstein, obviously, unbelievable mind and mathematician. His sister was actually quoted as saying, "Nothing gives him more pleasure than music." Actually, that's wrong – he said that. [Rewind sound effect and laughter]. So Einstein-- I already misquoted Einstein [laughter]. So Einstein said, "Nothing gives me more pleasure than music." And Einstein, quite famously, would use music almost as a way to solve his problems. He's not writing it out on a piano and then suddenly the solution will come to him through a cord, but he will use music in a way that it can kind of let his mind wander as he plays. But he's from a family of musicians [laughter]. I'm not comparing my family to Einstein's family [laughter]

MADDIE: 15:51

you're like, "I just like Einstein."

ANDY: 15:53

Just FYI.

MADDIE: 15:56

You share a lot of similarities.

ANDY: 15:58

Yeah, I'm ahead of my time [laughter]. Those golden sheep are really going to [laughter]-- they're going to listen to this one now. But so he learned violin as a child, and like a lot of kids, he hated music. He really hated it until he discovered Mozart, age 13. And that doesn't really surprise me that it was Mozart that was the composer that kind of got him into music properly. It's kind of interesting. Now, teenagers don't rebel in the same way. "I'm going to listen to Mozart [laughter]." But Mozart was incredibly mathematical as well, so a lot of his music is based around that. There's a lot of rules in his music. He was, according to his sister, obsessed with figures. And I guess not figures in figurine style, but numbers and figures and things like that. And he would write pieces sometimes using a dice, for example. Not regularly, but I mean, it just shows his sort of interest in the field. And there were certain sequences that his mother could play on the piano to get him out of bed in the morning. So like a lot of kids, he'd be stuck in bed, couldn't get out. And rather than, I don't know, pour water on him or whatever most parents do to--

MADDIE: 17:20

Yeah, like take the blankets.

ANDY: 17:21

That's not what my-- yeah. That's not what my parents did, just FYI.

MADDIE: 17:23

I hope not, yeah [laughter].

ANDY: 17:25

Very aggressive [crosstalk].

MADDIE: 17:27

Bucket of water in the morning [laughter].

ANDY: 17:28

Oh, morning, mom [laughter]. So rather than do that, his mom would be downstairs, or so rumour has it, anyway, and would play a particular chord sequence and he would be so frustrated to not finish that particular sequence. So if I just--

MADDIE: 17:46

She would leave it unfinished.

ANDY: 17:47

Yes. So I've got a guitar here, by the way,

Andy narrating: We actually found this guitar in the place where we were recording which was pretty lucky. It’s also quite… rustic. So, apologies if it seems a bit out of tune in certain registers.

So if we have-- so she'd play a series of chords like this. [music] And that almost leaving him hanging. That would frustrate him so much, he'd run downstairs and go, [music] "You got me again."

MADDIE: 18:21

Good one, mom.

ANDY: 18:21

Yeah, cheers, mom. Yeah, so that's Mozart for you. Other quick examples, Brian May, guitarist in Queen has a PhD in astrophysics. Art Garfunkel has a master's in maths. There's all sorts of links between-- I mean, I guess by the law of big numbers and being so many musicians, some of them are going to have studied maths. There is evidence that there is links between them. So we were talking about perceived beauty and how that lends itself really well to music and how we might interpret things in different ways. And there is a lot of numbers behind that, and again, we can start to think about the file, the sort of numeric infinite sequences.

MADDIE: 19:11

The ratio numbers, yeah.

ANDY: 19:11

Exactly. So it gets a little bit technical as you sort of look into the math behind it. But there are more-- before we talk about that, there are more sort of glaring examples of that that people have spotted in music. And to your point about Mona Lisa, you kind of take these with a pinch of salt whether these were intentional things or not. Who are we to say?

MADDIE: 19:32

Yeah, exactly.

ANDY: 19:32

So one example would be a composer called Béla Bartók, a fantastic composer I studied a long time ago at our university. And my interest in film music, I would come across him from time to time. Stanley Kubrick was a massive fan of Béla Bartók and used him in films like The Shining. And actually, this piece here is in The Shining. He was really interested in the Fibonacci sequence and how he could sort of use that in music. And it does appear in all sorts of places in his work and in his pieces. There's a particular snippet, which I guess we could play now, which is a glaring example of the Fibonacci sequence in music.

MADDIE: 20:22

Yeah, let's play it.

ANDY: 20:22

And this is one I think is intentional.

MADDIE: 20:25

Yeah, I want to hear it.

ANDY: 20:26

So if we think back to that sequence, so that sequence being one, one, two, three, five, eight, he works his way up that sequence and back down it just with the xylophone on a single note, just hitting it, and divides the numbers up evenly within a section.

MADDIE: 20:46

Got it.

ANDY: 20:47

Very haunting.

MADDIE: 20:47

Yeah, it sounds haunting [laughter]

ANDY: 20:51

[music] So that's a really glaring example, and I would say unquestionably that is intentional. Beethoven was obviously a massive pioneer at the time, unbelievable composer, well, in terms of great music. Some of the, I guess, what people would say were obvious examples of the Fibonacci sequence are in what is probably the piece that people know the most by Beethoven, which is his Fifth Symphony, the da-da-da – we’re sticking with the haunting theme.

ANDY: 21:25

Some people have argued that he's really referencing the Fibonacci sequence in that. The reasons being, that famous motif, last five bars. Oh, okay. Five bars. That's a number from the sequence, and we're kind of getting back to the sheep example.

MADDIE: 21:45

Does that matter--?

ANDY: 21:45

Yeah. Exactly. Is that that strange? Well, kind of. Often some melodic phrases would've been 4 bars or 8 bars or 16 bars, so maybe that's a bit of an odd to it. That famous theme happens at the start and at the end. It's referenced throughout, but more recently, people are now looking for the golden mean in a piece, and that's looking for that sort of ratio like we described between A and B in the middle of a piece. So if we take the golden mean which would be 1.618, so this piece is 601 bars long. I've pre-done the maths [laughter].

MADDIE: 22:20


ANDY: 22:22

I'm doing this live.

MADDIE: 22:23

You're very good at this.

ANDY: 22:24

Yeah. If it’s not in twos, I'm fine [laughter]. So it's 601 bars long. So if you think about 1.618 through that, it's going to be, I guess, it’s a little over half -way which is actually bar 372, and that's where it really comes back, that theme, at the end. So people would argue that he's sort of referencing that there. I'm less sure that Beethoven is doing that, but it's a theme that's 5 bars long, and the whole da-da-da-da-da is a third, like a doorbell.

ANDY: 22:58

But that's a third which is another Fibonacci number. Yes. So maybe Beethoven was intentional in doing that, and I kind of feel like you can wade your way through enough music now, and you can say any artist is doing it if you find a particularly interesting part of the piece. So one example I saw the other day, which I think is almost certainly ridiculous, is Drake's “In My Feelings”. I don't know if you know the song. Is that even how you say it?

MADDIE: 23:25


ANDY: 23:28

Okay. I'm quite new to some of his work.

MADDIE: 23:28

What would you have said?

ANDY: 23:30

I thought it was “In My Feeling”.

MADDIE: 23:31

I think it's “In My Feelings”.

ANDY: 23:34

He's got plural feelings, well, [crosstalk].

MADDIE: 23:36

Because you have feelings. Yeah [laughter].

ANDY: 23:37

Wow. I don't know how that feels [laughter]. So we can do the same thing without looking at the bars. We just look at the length of the song. So that song is 3 minutes 37 on Spotify. So it’s 217 seconds, and the golden mean point in that three times 217 by 1.618. We're looking at about 134 second into the piece. It would be this so-called golden moment where the whole thing comes together Again, I’m completely unsure about that as a principle.

MADDIE: 24:14

But even if it wasn't like, okay right at this second let's do this big moment in the song, it goes back to the same thing where it's like, oh it's nice and it makes sense here.

ANDY: 24:23

Yeah, true. But I think there are much better examples if we start to talk and think about how music works. And my example kind of nearer the start where I was saying, "These notes sound nice together and these don't," there's a lot of numbers at play in how we interpret that and how notes together sound nice. And that all comes down to sort of the fundamentals of harmony and even pitch. If we start to think about what is a pitch? And there's a difference, obviously, between sounds and pitches. Those frequencies are going to be consistent and you're going to get a note. And, I guess, once we kind of dig deeper into how that works because that at heart is numbers-- so it's Hertz. So 64 Hertz for example is a C. So if I was to play a C on a guitar [inaudible] how many notes are we hearing there?

MADDIE: 25:21


ANDY: 25:22

Exactly. It's one note. However, the way the sound waves work, there's actually harmonic overtones. We're actually hearing lots of different pitches on top of that, as well, in a particular order that is defined. So if this is 64 Hertz, the first note that actually in theory is in play to give us the richness of that sort of low C there, is the C above it. The C above it would be 128 Hertz [inaudible]. Those two notes sound nice together, I think, if the guitar is in tune. And that gives us an octave. It's the same note but an octave higher. Oct being eight. Just eight notes higher. Note there I spoke about the Hertz. If a C is 64 Hertz, that octave above we've added 64 Hertz to it. 128 in the first instance. These notes that I'm playing actually because I'm on a guitar-- we should really have done this on a piano, but they're a bit lower than this but the same rules apply. This harmonic series, or this kind of order of notes, it is just this infinite sequence like the Fibonacci sequence we spoke about before. If the first note that we're going to pick, and it doesn't matter what we pick, so we'll pick a C, 64 Hertz. If we add 64 every single time, the order of which the notes that produces almost is like an order that feels most familiar to us. So the first note we'll hear after that one [music] is an octave above [music]. The note that we'll hear after that is a fifth above it, which is a G [music]. So these two notes [music] or if I play all three of those notes [music] we're kind of getting the basis for a chord here. That's like a nice chord. Does this feel [music] aggressive on your ears [laughter] at this point?

MADDIE: 27:26

Yeah. It's very soothing.

ANDY: 27:27

Exactly. But it's all just numbers at play there. So that's just 64 [music], add 64 [music], add 64 [music]. But what's interesting here is the relationship, therefore, that all of these notes have to that root note. So the root note is the first one I play [music]. So it's relationship to the note afterwards, Hertz-wise, is it's half of that. It's now a third of that note so again, this infinite sequence,

ANDY: 27:59

So if we start to think of chords in the most basic part of just being, whatever the [tonic?] is the first note. If we take the first third, fifth, we've got the chord then, it doesn't matter where that is, we can move it. We're going to get chords that sound nice basically. What is really relevant, and this is why we're talking about-- what have I just said there? Notice 1, 3, 5, and 8. Where have you seen that before?

MADDIE: 28:26

Fibonacci numbers.

ANDY: 28:26

Here it is, it's back. So 1, 3, 5, and 8 Fibonacci numbers-- it's the perfect sequence for our [ears?], our minds to interpret. Yeah, I guess. So people say, "I woke up and had this melody in my head." and you think, "Okay, well, how did you put chords to that?". Well, there are certain chords that will work well with certain notes. And if we're thinking ahead to later in our conversations, where we'll talk a bit about how machines and algorithms could start to write music. We're already talking about certain rules. And if I choose a chord that doesn't have a [unit?], it might not sound as nice. There's nothing wrong with that necessarily. Sometimes breaking the rules is a good thing. Because it's that we don't want the whole thing to sound 'nice' because that's--

MADDIE: 29:16

--what we heard before.

ANDY: 29:16

Exactly. So it's nothing new. And it's the kind of tension and resolution that sometimes is so powerful in music, right? This is where we said you might compose a piece.

MADDIE: 29:30

Andy came up with this idea of putting what we just learned to the test. So he asked if I could find a deck of cards lying around, and of course, created a soundtrack for me looking for the cards.

MADDIE: 29:46

Sounds like I need a rest now.

ANDY: 30:14

Yes, good quick [music] We actually only need number one to eight. So we could just draw--

MADDIE: 30:15

As luck would have it, I found some.

ANDY: 30:19

Oh, nice. Perfect. Right. We need to just find numbers one to eight basically.

MADDIE: 30:28

This is quite a nice deck of cards.

ANDY: 30:28

Actually, let's just pick out numbers one, three, four, and five.

MADDIE: 30:36

Okay. All of them?

ANDY: 30:37


MADDIE: 30:40

Okay. One, three, four, and five?

ANDY: 30:46

Yeah. Are we doing aces as one?

MADDIE: 30:46


ANDY: 30:47

Well what the hell else would we use [laughter]? Should we use nines as ones? [laughter]. Okay. And I'm going to throw in a six.

MADDIE: 31:01

Just to spice it up.

ANDY: 31:01

Yeah. Make it real spicy. Okay. Are we still recording?

MADDIE: 31:01


ANDY: 31:07

Let's move that. Let's Just close it.

MADDIE: 31:15

Should we combine them?

ANDY: 31:16

Yes. You can shuffle, I can not shuffle.

MADDIE: 31:17

You want to shuffle? Okay.

ANDY: 31:20

So Maddie while you shuffle. What we're going to do now is through randomness write a piece of pop music. The reason we're choosing pop is to my point before it's actually very quick to put something together that will sound nice. This is very much a fast food situation that we're dining on though [laughter]. Well this isn't going to be like--

MADDIE: 31:43

We're about to embark on.

ANDY: 31:44

Yeah. They're not going to look back on this in a 100 years time and think, "Wow, that was good."

MADDIE: 31:48

You don't know that [laughter].

ANDY: 31:50

Only the sheep will tell us [laughter]. So the only rules I'm going to put in place for this-- and bear in mind I've never tried this by the way. This is just an idea I had on the way here. I think it would be good to finish on the root. So that means the final note in our melody needs to be one. So could you find an ace and put an ace at the end here? So we've decided that the final note is going to be the root. And we're going to play in C major because we're about to go to the piano and C major is the easiest key to play in. That goes without saying. So how many notes do you want in this melody? Bear in mind this is a melody you're going to just play over and over. So anything between I'd say--

MADDIE: 32:41

So like four.

ANDY: 32:42

--four and seven.

MADDIE: 32:43

Let's do five.

ANDY: 32:44

Yeah. Five's a good one. Bear in mind that you've already got a note at the end there.

MADDIE: 32:45

Right. So we're going to pick four more cards?

ANDY: 32:47

Yep. So you're going to pick four cards here. And I think in the interest in you choosing the melody you'd prefer do that twice. And I'll play both of them and you can choose which melody you'd prefer.

MADDIE: 32:56

Got it. Okay. Okay. So I'm flipping it over.

ANDY: 33:00

First out of the hat is a four. Excellent. A three. A one. And we need one more. And a five. Very interesting. I'm going to move these the other way round because we're sat opposite. So yeah. Okay. So four three four five one. Interesting. So note four is going to be one, two, three four. [music]. Oh, isn't that lovely [laughter]. That's your first piece of music.

MADDIE: 33:31

Oh, wow. [inaudible]. Kind of sounds like I just won something. [music]. Yeah [laughter].

ANDY: 33:41

Yeah. What kind of rhythm would you want? [music] or like a [music].

MADDIE: 33:51

Oh, nice. Yeah. That or like the faster one that you did whether or-- yeah.

ANDY: 33:57

Okay, that's your first melody. Let's try it one more time.

MADDIE: 33:58

So four more cards.

ANDY: 33:59

This time no rules. Yeah. Four, one, four, five and--

MADDIE: 34:07

Then the one, right?

ANDY: 34:07

No, let's get rid of that rule. That's four, one, four, five, five. [music] It's more like you're asking a question now.

MADDIE: 34:23

It is, yeah.

ANDY: 34:28

And then you got the question right and you win the prize. [music] So that's your melody.

MADDIE: 34:27

It's perfect, yeah.

ANDY: 34:28

Which do you prefer?

MADDIE: 34:29

I think I like the first one. [music] The second--

ANDY: 34:32

It's good. It's kind of simple because you're going to have to learn it on piano now. Should we go try it?

MADDIE: 34:37

Then we shuffled over to the piano.

ANDY: 34:42

We don't need the cards to [make it work?]. Oh, there's two seats, ready and waiting.

MADDIE: 34:46

Yes, it's nice, right?

ANDY: 34:48

So you need the seat on the right.

MADDIE: 34:49

It's cold down here. Turn on a light… there we go. [music] Nice.

ANDY: 35:05

Your melody.

MADDIE: 35:06

Nice. [music]

ANDY: 35:15

Yeah, nice.

MADDIE: 35:18

I don't know if this is right. [music]

ANDY: 35:32

Very sinister. It's your piece! [music].

MADDIE: 35:42

It's awesome. [music]

ANDY: 35:47

Alright anyways, this is why we need two of us

MADDIE: 35:49

Our voices do sound kind of echoey down here, so maybe we won't talk.

ANDY: 35:52

And should I leave the volume as it is?

MADDIE: 35:53

Yeah, the volume is fine.

ANDY: 35:56

Okay, so I'm going to have to talk to teach you though, otherwise, you can't [crosstalk]. So the note. [music] So these are you notes. [music] So if you put those things on--

MADDIE: 36:10

Wait. [music] What is it? [music] No. [music] Oh, that, okay.

ANDY: 36:29

That one then. [music] Yeah, nice. Okay, so you just play that over and over. Just play the note. [music]

MADDIE: 36:42

Oh, I messed up.

ANDY: 36:43

So the idea is you could choose any of these notes now in any order, and it will sound like-- so you can play any of these notes [inaudible], so just play them randomly and it will sound nice.

MADDIE: 36:55

Wait, of the same ones?

ANDY: 36:57

Yeah, but in any order. [music]

ANDY: 37:21

There you go, that was your first pop song you wrote with cards. Do you think that sounded good?

MADDIE: 37:26

I thought that was great.

ANDY: 37:26

I thought that sounded really good.

MADDIE: 37:28

I think it did too.

ANDY: 37:30

So in theory, actually I should wait till that mic is on.

MADDIE: 37:33

Yeah, yeah. Well, it is on.

ANDY: 37:34

But I’ll still wait

MADDIE: 37:41

I didn’t stop it. Okay.

ANDY: 37:41

We good?

MADDIE: 37:42


ANDY: 37:43

That was nice.

MADDIE: 37:43

It was nice.

ANDY: 37:50

What did you think of your first pop song?

MADDIE: 37:51

I loved it.

ANDY: 37:53

I thought it was good.

MADDIE: 37:55

Yeah. If I do say so myself. I was really great during the cards and putting that together.

ANDY: 37:58

You learned it pretty quickly.

MADDIE: 38:00

No, it wasn't that quick.

ANDY: 38:03

So that is obviously a really simple example of how we can use randomness and numbers to put together a piece very, very quickly. And really, it's all about this sort of-- as we said before, where you start to make it a bit more interesting is where you introduce imperfections or you can start to use the numbers to tell you a bit of a story.

MADDIE: 38:26

Now that we've created a melody using randomness within a set of rules and a deck of cards, I wanted to hear how we could apply this using a computer or even Alteryx.

ANDY: 38:36

So as we've discussed and seen when we talked about nature, talked about sequences and things like that, algorithms have existed way longer than computers. But there is a big focus now on how can computers be used and algorithms and things like that, or even like, say, machine learning to write pieces. You and I have just written a pop song in 30 seconds using a pack of cards.

MADDIE: 39:01


ANDY: 39:03

Yeah. And obviously, we applied a few rules. In theory now, we could write programs that write that kind of music. I'm not saying their going to be anywhere near as good, but some people have begun to explore that kind of thing. So there's examples that we'll put in the session notes, from things like counterpoint programs, so these kind of rules that people have written into bits of software, and we'll see in this example things like if the melodic leap is greater than an octave, then try something else. And that's the if statement that they've written. These are things that are now becoming really familiar to our day to day in Alteryx or whatever other tool we're using. It's just an if statement. The idea being that you can only break certain rules so many times, and if you got to the end of the piece and you've broken too many rules, it would say, "Just loop around and start again." And that's sort of loop around and start again is just one massive iterative macro. It's saying, "Write and write and write a piece, just randomly picking notes." In theory, because once we pick this next note, does it match the rules? Is it fifth above the note we've just had? Yes, it is. Okay. Was the not before it also a fifth above? No, it wasn't. Okay, so it's fine. What are the notes we should have underneath it? The harmonies? So like what I was playing on the piano. Well, we know if that note's a fifth, we know what chords will work nice underneath it, and then we can jut keep sort of randomly picking cards or using random functions and formulas to just keep going and going and going, and effectively it will start to just write us a piece for us. What's really interesting is obviously then when you get into machine learning models where we have sort of feedback loops that say, "Oh, I like this piece. Oh, that one was rubbish. Don't do that again." And the more and more we train it, the better and better it gets. That becomes an interesting debate then over is there any use to that? Because who's training that model? So there's all sorts of examples we see day to day where-- we all know the examples, like where you log into a new website or something, and it says, "Oh, click on the images that are road signs." And here you're effectively training am image recognition model there. It wants to check that it was right some time. But we couldn't do that with music and say, "Okay, so which of these sound like good Bach fugues?" Not because people don't necessarily know of fugues, but it's just like there is always that element of subjectivity to music or to art.

MADDIE: 41:31

Yeah. Is it good or not?

ANDY: 41:31

Exactly. What is good to me might sound horrible to you and vice versa. Everything we discussed at the start, there's still this air of subjectability, which I think might be a made-up word.

MADDIE: 41:44

No, I think that's right. Subjectability.

MADDIE: 41:46

Actually, I checked, and it's definitely not a word.

ANDY: 41:50

Nice. It's the longest word I've ever used. Okay. So this is a good example. So Pierre Boulez, he at one point would write a lot of sort of serialist music. So, serialism. and one focus of that-- I'm not going to go into too much detail of what that is, because we'll be here all day, but he said, "When I wrote these pieces, the responsibility of the composer is practically absent. Had computers existed at the time, I would have just put the data through them and made the pieces that way, but I had to do it by hand." And he's sort of created this piece kind of like we were doing there, but with cards. In theory, we just did that by hand. His process is obviously a million times more complex than what we just did, and I've got an example which we can play, or you can, I guess, add later through Spotify. It's on the playlist. And the idea is we can use these sort of algorithms to write these pieces. And then that's when you get into really strange sort of territory of just because we can, doesn't mean we should.

MADDIE: 42:56


ANDY: 42:56

It's more an exercise in the sort of, I guess, like what is possible as opposed to like, what is nice. I keep using the word nice, and really bad word to use throughout this podcast because nice itself is subjective.

MADDIE: 43:13

Or some say, [subjectible?].

ANDY: 43:15

It shouldn't have to be nice, that's a ridiculous thing to keep saying, "Well, that's nice." Always be nice. But [Boulez?] himself said, once you finish writing it, he said, "I'm not terribly eager to listen to it. It was just an experiment that was absolutely necessary." It's interesting to see how these things sound.

ANDY: 43:39

But anyway, right, let's talk about composers that are now using technology to write music, and of course, all pop music and everything now does use technology. We have Garage Band on our phones now. That's not really what we're talking about here. We're talking about much more complex algorithm-based software that can create sounds, and write the pieces for you under some guidance from a human. My brother's a pianist musician. He's very interested in this kind of stuff. And so I guess modern classical music, and he was telling me about composer called Xenakis is, which I think is how you say it, I've been practising that. And he would write sort of stochastic compositions, stochastic is itself term in maths which designates a process in which sort of sequence of values is drawn from a corresponding sequence of jointly distributed random variables. Got it?

MADDIE: 44:13


ANDY: 44:15

So he's got this software that can do that, or is using high-speed computations to calculate various probability theories to aid compositions, like “Atrées” which we can hear now. Yeah, we'll put it in after okay. So that programme he's got sort of takes a score from a list of notes. And then you can weight it by the programmer-- so the composer is now the programmer, I guess, effectively, the software is writing the music for him. But we can already start thinking about how we're doing things like that in our day to day work with clients. So the composer or programmer is dictating the score in the system so you can choose what are your sort of weightings, or your preferences and that comes from say, an analytic app. It will randomly generate certain functions and work out the probability of certain things appearing. From there it can start to put together a piece. Does it meet the rules that the user has set at the start? No? Loop around and do it again and keep going and going until I've got the piece that I want. So there's some interesting experiments into at the moment, it's not something I know loads about, as might have been apparent there. But it's really, the thing that's going to really interest me about this and as it grows, is sort of the Jurassic Park quote, is that "Just because we can doesn't mean we should." and I don't think we're going to end up in the Jurassic Park style situation. But like the whole point, in my opinion of music is this creative outlet whether I compose [or we're?] listening to it, and it can invoke certain feelings. And it will always, in my opinion, have that human element to it. I think computers will enhance the ways in which we can create music, and we can have a piece of software that will write a pop song like we just had there. But kind of so what? We can do that in two seconds ourselves, and we're not in a position where we need a million pop songs tomorrow.

MADDIE: 46:46

Yeah. Exactly.

ANDY: 46:46

So I just think it can only enhance music, but I don't think it's ever going to, in my opinion, replace the human composer.

MADDIE: 46:57

[music] Thanks for tuning in to Alter Everything. Check out our show notes on for behind the scenes photos, bonus content, and a playlist featuring the music we chatted about on today's episode. This was our last episode of 2019, and I want to hear your thoughts on what you'd like to hear featured in 2020. So if you have a great topic to share or you just want to share your two cents, fill out our survey on community linked in the show notes. Next year, we're rolling out some exciting new theme music submitted by you, our listeners, as well as new podcast artwork, and lots of extra bonus content so stay tuned. Tweet with us using #altereverythingpodcast. And, as always, thanks for listening.

ANDY: 47:46

So if we think about chord one, chord five, chord six, and chord four, that is the basis of loads and loads of pop songs. So with year seven, who are the youngest year that you teach in a secondary school, this was always pretty much the first lesson I'd do with them. This sounds very campfire Kumbaya-esque but we'd sit in a big circle--

MADDIE: 48:15


ANDY: 48:19

--I'd have a guitar-- well, that's not the word they use, but, guitar, and I'd just play these chords over and over. And the whole point was we needed to try get the room singing in different parts or different melodic lines, which, in theory, is quite a difficult challenge because that's a tricky skill to be able to sing something at the same time that is different to what someone else is singing. But because it's the basis of loads of pop music, and there's a great video that kind of explains this by a band called-- I think they're called Axis of Awesome something. We'll put a link in the notes. But they call it the 4 Chord song, which it is--

MADDIE: 48:52


ANDY: 48:52

--and their point is, "[Oh, no?]. We hear it in this pop song. You hear it in this pop song. It's just those chords over and over." Right. This is where you might need to do some editing [crosstalk] [music]

ANDY: 49:11

So what's the song?

MADDIE: 49:11


ANDY: 49:12

Very good.

MADDIE: 49:23

Yeah. Where is the Love?.

ANDY: 49:25

Black Eyed Peas. Very good.

MADDIE: 49:26

I want to be forever young. Forever Young, yeah.

ANDY: 49:30

Very good. James Blunt. Is he big in the US?

MADDIE: 49:31


ANDY: 49:31

My singing's going to be horrible but, “My life is brilliant. My love is pure.”

MADDIE: 49:40

Oh, wow. You sound just like him.

ANDY: 49:41

Oh, he's here in the room with us. Is this the way you left me. I'm not pretending-- No happy ending. I won't hesitate no more, no more. It cannot [be?]. Honestly, there's so many more. U2, With or Without You. Maroon 5. She will be loved. Even Disney. Can you feel the love tonight. [music], "Oh, I love that song." It's like you've heard that song a million times elsewhere although that was a bad one to use as an example because that's probably the best of all the ones [laughter]. But the idea is so because it's all over the same chords, the notes of what I sing, the melody what I sing, whilst it might be different to what you're singing, we'll work together. But obviously the point there is different melodies there work together because it's all over the same core progression. The notes in these melodies are all just using notes from those chords, which is chord one, chord five, chord six, chord four. And we'll hear that in loads of other songs. I think a capo actually let's just see if we can work it out. What was it? Leonard Cohen

MADDIE: 50:59

Leonard Cohen. Hallelujah.

ANDY: 51:01

Yeah. I don't think I can sing it at this pitch. But he references the chord numbers. Shall I put my capo on?

MADDIE: 51:16

Yeah. [music] Nice.

ANDY: 51:17

I think this is a better fit. It goes like this, the fourth, the fifth. He then says minor fall, but it actually goes up to the sixth chord. Minor fall, because the sixth is minor, and then back to the fourth, the major lift, then back to the fifth.

MADDIE: 51:37

Yeah, there you go.

ANDY: 51:39

But he doesn't. He goes-- and this is where when we mentioned before, you can start to use notes that aren't necessarily in the scale. Instead of going to resolve it, he goes from this chord to a really, on its own, less nice sounding chord. And that's because you've got this note in here. This note is not part of the Hallelujah scale. So the scale is when the key here, B flat. This note does not belong in that. But, to my point before, it's breaking the rules that make music sound nice, basically.

MADDIE: 52:12

Stands out. Yeah.

ANDY: 52:14

So the reason why this kind of works is going to-- is our ear is desperate or our mind is desperate for it to kind of resolve, so the note would typically fall or raise by one. More often than not fall, but we can go up here. So we can go from here to here, so we have these two notes. If you put them both up so like on a piano, we just literally move them up one note each. We go from there to-- and then suddenly it's kind of resolved. And then we get into the chorus of-- and that's note we were looking at before. We can make it sound nice Hallelujah if we use a chord that's a major chord with that note in it, or we can use a minor chord that has that in it. So it's like a sad Hallelujah now. And then he just kind of plays on that for a while.

ANDY: 53:12

[music] And it's just going between those two chords, basically. Jeff Buckley's version, which is slightly more dark, he takes a bit further, and he starts putting this note in it as well, so where we get-- which is quite a nice sound there--

MADDIE: 53:28

Yeah, that's really nice.

ANDY: 53:28

Jeff Buckley starts his, the intro of the piece is-- he gets kind of aggressive, like-- that's because these two notes are a semi tone apart, which is not very nice. So then he resolves it, and you think, "Okay, nice. We're back." And then he puts another one in. And then again, we're desperate for that to resolve. So that's only going up by a semi tone. So it's actually where our mind interprets these notes, which are really, I guess, intervals in the scale. If one of them feels a bit out of place, it's kind of okay with it if it then resolves to something nicer. So like the-- that's just going from there to there. And pop music, they play on that all the time.

(Closing theme music)




Episode Transcription
Nice job! You found a secret bonus after show, only available here on the Community show notes. Enjoy, and don't forget to download Steve Reich's "Clapping Music" app, linked above.

This episode of Alter Everything was produced by Maddie Johannsen (@MaddieJ).