0:00:05 > 0:00:11# Oh, baby, baby, the reason I breathe is you
0:00:13 > 0:00:17# Boy, you've got me blinded... #
0:00:17 > 0:00:20We're told that ours is an age of data,
0:00:20 > 0:00:23an age in which practically everything we do,
0:00:23 > 0:00:28say and make can be reduced to flows of information by algorithms
0:00:28 > 0:00:32spinning away on server farms in remote locations.
0:00:37 > 0:00:41Data science already tells us where to go, how to get there,
0:00:41 > 0:00:44who to date and what to buy them.
0:00:44 > 0:00:46But all that's rather trivial stuff.
0:00:46 > 0:00:52I think that even the most glorious, ephemeral and marvellous
0:00:52 > 0:00:54products of the human mind
0:00:54 > 0:00:58can in fact be measured by science.
0:00:58 > 0:01:01And by that, I mean pop.
0:01:01 > 0:01:05MUSIC: I Really Like You by Carly Rae Jepsen
0:01:05 > 0:01:07Armed with just a few algorithms,
0:01:07 > 0:01:10I intend to change the way we understand pop.
0:01:10 > 0:01:13This is based upon the music -
0:01:13 > 0:01:15the wave forms, the numbers.
0:01:17 > 0:01:21I'm gathering a team of data scientists to analyse
0:01:21 > 0:01:23five decades of the UK's Top 40 hits.
0:01:24 > 0:01:27Together, we'll show the artists what science can do.
0:01:30 > 0:01:32# Doo, doo! Doo-doo-doo-doo! #
0:01:32 > 0:01:36We're teaming up with pop legend Trevor Horn.
0:01:36 > 0:01:38So, this is a thing called M Harmony PCA 4.
0:01:39 > 0:01:40Right.
0:01:40 > 0:01:44# Into the ocean with me... #
0:01:44 > 0:01:48We're going to try to do something never tried before -
0:01:48 > 0:01:52use data to give an unsigned artist a potential hit.
0:01:52 > 0:01:54I like the end. That was good.
0:01:54 > 0:01:56I can't even imagine, like,
0:01:56 > 0:01:59what the science is going to say or do to the song.
0:01:59 > 0:02:01MUSIC: Hey Ya! by OutKast
0:02:01 > 0:02:05I'll also be using my analysis to map the turning points
0:02:05 > 0:02:06of pop history.
0:02:06 > 0:02:08Love Me Do, right on the average.
0:02:08 > 0:02:11Yellow Submarine, right on the average.
0:02:11 > 0:02:14Lennon and McCartney are writing ditties for prepubescent girls.
0:02:14 > 0:02:17We don't know what the answers will be.
0:02:17 > 0:02:19Yeah, it's not exactly an advertisement for machine learning, is it?
0:02:19 > 0:02:22- No. - THEY LAUGH
0:02:22 > 0:02:24But that's what happens when science...
0:02:24 > 0:02:27- Thank you, sir.- Thank you for exposing me on national television
0:02:27 > 0:02:28- in getting it wrong! - THEY LAUGH
0:02:28 > 0:02:29Hey.
0:02:29 > 0:02:31..meets culture.
0:02:31 > 0:02:33What exactly are your qualifications?
0:02:33 > 0:02:35My PhD was on fruit flies,
0:02:35 > 0:02:38I've spent most of my professional life studying worms.
0:02:48 > 0:02:53MUSIC: You're Gonna Miss Me by 13th Floor Elevators
0:02:56 > 0:02:58Let me begin with a confession.
0:02:58 > 0:03:01I'm not much of a music fan.
0:03:01 > 0:03:04Worse, as an immigrant that's citizen of nowhere,
0:03:04 > 0:03:09what I know about British pop history is distinctly second-hand.
0:03:09 > 0:03:12It's not that I don't like the stuff, it's just that I don't
0:03:12 > 0:03:14have stacks of vinyl at home.
0:03:14 > 0:03:17And if you spent your youth
0:03:17 > 0:03:19throwing up on the King's Road or off your head in a field outside of
0:03:19 > 0:03:23Reading, then you know more about British popular music than I do.
0:03:23 > 0:03:27But to do what I want to do I don't have to be a fan.
0:03:27 > 0:03:31That's because what I want to do is science.
0:03:31 > 0:03:34You may wonder why an evolutionary biologist should
0:03:34 > 0:03:39decide to study the charts, but just as fruit flies and finches evolve,
0:03:39 > 0:03:42so too, I believe, does pop.
0:03:44 > 0:03:48Every new song comes with its own burden of mutations.
0:03:48 > 0:03:53Some of them bad, but a few of them flourish
0:03:53 > 0:03:55and get passed on to future generations.
0:03:57 > 0:04:01Listen carefully and you can hear the music evolve.
0:04:01 > 0:04:06There are countless examples, but one clear primordial ancestor
0:04:06 > 0:04:08is Kraftwerk's Autobahn.
0:04:11 > 0:04:13Bit of honking, bit of synthesiser.
0:04:16 > 0:04:19The weirdness begins right from the start.
0:04:19 > 0:04:22# Fahren, fahren, fahren auf der Autobahn... #
0:04:22 > 0:04:24"Bahn, bahn, bahn, autobahn?"
0:04:24 > 0:04:26I mean, really.
0:04:26 > 0:04:30It must've been the weirdest thing possible when people first heard it.
0:04:30 > 0:04:33# Fahren, fahren, fahren auf der Autobahn... #
0:04:33 > 0:04:37Think of its glorious weirdness as a musical mutation.
0:04:37 > 0:04:40Remember, this was 1974.
0:04:43 > 0:04:44Bizarre though it may have been,
0:04:44 > 0:04:49Kraftwerk's mutation changed the course of pop evolution.
0:04:49 > 0:04:53Three years later, record producer Giorgio Moroder heard it,
0:04:53 > 0:04:55absorbed it and put it into a song
0:04:55 > 0:04:58that he made for disco's ultimate diva.
0:04:58 > 0:05:01It begins very Kraftwerk-like.
0:05:01 > 0:05:03MUSIC: I Feel Love by Donna Summer
0:05:03 > 0:05:04Driving drum machine.
0:05:06 > 0:05:07Synthesisers coming in.
0:05:10 > 0:05:12But it's got a different feel, this is dance music.
0:05:15 > 0:05:16And then.
0:05:16 > 0:05:20# Ooh
0:05:20 > 0:05:21# Heaven knows, heaven knows... #
0:05:21 > 0:05:24Donna Summer having an orgasm, or at least faking one.
0:05:26 > 0:05:30I Feel Love was a glorious synthesis of disco,
0:05:30 > 0:05:33early electronica and pure sex.
0:05:33 > 0:05:35It was the future.
0:05:36 > 0:05:40Moroder's formula would become the basis of electronic dance music
0:05:40 > 0:05:42and its innumerable subgenres -
0:05:42 > 0:05:43house, techno,
0:05:43 > 0:05:47not to mention neurofunk, speedcore and cybergrind.
0:05:49 > 0:05:52I have no idea what I'm saying
0:05:52 > 0:05:56but I do believe that all genres only exist because musical mutations
0:05:56 > 0:06:00are passed from one generation to the next.
0:06:00 > 0:06:05And it's that inheritance, that lineage,
0:06:05 > 0:06:10which is then transmitted and recombined with other elements.
0:06:10 > 0:06:14That is the essence of evolution.
0:06:14 > 0:06:17That is how pop music evolves.
0:06:17 > 0:06:19MUSIC: Hey Ya! by OutKast
0:06:19 > 0:06:21MUSIC: Back In Black by AC/DC
0:06:21 > 0:06:24MUSIC: Fame by David Bowie
0:06:24 > 0:06:27To me, pop should be a science of diversity and change,
0:06:27 > 0:06:30competition and conflict.
0:06:30 > 0:06:31MUSIC: Heartbreak Hotel by Elvis Presley
0:06:31 > 0:06:33MUSIC: Losing My Religion BY R.E.M.
0:06:33 > 0:06:36MUSIC: Toxic by Britney Spears
0:06:36 > 0:06:41MUSIC: Royals by Lorde
0:06:41 > 0:06:43But if we're going to make it a science,
0:06:43 > 0:06:46a Darwinian science, we need to do what scientists do -
0:06:46 > 0:06:48experiments.
0:06:48 > 0:06:50MUSIC: Blue Monday by New Order
0:06:51 > 0:06:53And in my first experiment,
0:06:53 > 0:06:55I want to find the musical adaptations that define
0:06:55 > 0:07:01pop success today and put all of them into just one song.
0:07:01 > 0:07:04But to do that, I'll need a music producer who'll let science
0:07:04 > 0:07:06into his studio.
0:07:06 > 0:07:09# They took the credit for your second symphony
0:07:09 > 0:07:13# Rewritten by machine on new technology
0:07:13 > 0:07:16# And now I understand the problems you can see
0:07:16 > 0:07:18# Oh, oh... #
0:07:18 > 0:07:22Trevor Horn is the man behind some of the catchiest tunes in pop.
0:07:22 > 0:07:24From writing songs with The Buggles...
0:07:24 > 0:07:27# Video killed the radio star... #
0:07:27 > 0:07:30..to producing Seal and Frankie Goes To Hollywood,
0:07:30 > 0:07:33he certainly knows how to make a hit,
0:07:33 > 0:07:35even if he isn't clear as to how he does it.
0:07:37 > 0:07:39Well, it's like those little puzzles, you know,
0:07:39 > 0:07:43where you've got to get, like, five balls into a hole.
0:07:43 > 0:07:45And you kind of manoeuvre it and you get one into
0:07:45 > 0:07:48a hole and then you get a second one in and then you're trying to
0:07:48 > 0:07:51get the third one in and the first two pop out.
0:07:51 > 0:07:54You know? Trying to get a hit record's a bit like that.
0:07:54 > 0:07:57Every time I've ever tried to analyse it and get any sort
0:07:57 > 0:08:00of hard and fast rule, it always changes and...
0:08:00 > 0:08:04# If I couldn't read you... #
0:08:04 > 0:08:07Trevor and I will be working with Nike Jemiyo,
0:08:07 > 0:08:09an unsigned singer.
0:08:09 > 0:08:13We're going to take one of her songs and try to turn it into a hit.
0:08:13 > 0:08:15She seems unconvinced.
0:08:16 > 0:08:21I think there's a reason why some songs last for decades.
0:08:21 > 0:08:28And I think that's more to do with heart, maybe, than science. Maybe.
0:08:30 > 0:08:33Imagine if you had a formula and all you had to do was adhere to
0:08:33 > 0:08:37this formula and you could churn out hit records, that would be so funny.
0:08:37 > 0:08:38I just can't see it, though.
0:08:38 > 0:08:41# Relax don't do it
0:08:41 > 0:08:43# When you want to suck it to it
0:08:43 > 0:08:45# Relax don't do it
0:08:46 > 0:08:49# When you want to come. #
0:08:49 > 0:08:52But I intend to bring some analytical firepower to bear
0:08:52 > 0:08:54on this problem.
0:08:54 > 0:08:58I've put together a team of analysts from the BBC R&D
0:08:58 > 0:09:01and from Queen Mary and Oxford universities
0:09:01 > 0:09:03and brought them to where I work,
0:09:03 > 0:09:07the Data Science Institute at Imperial College London.
0:09:07 > 0:09:10So if you look through, I've got, erm... Is there anything in there
0:09:10 > 0:09:12we recognise? Oh, there's Saturdays. I've heard of The Saturdays.
0:09:12 > 0:09:16We're beginning our analysis with the last six years of chart music.
0:09:16 > 0:09:19Who else is in there? Kanye West.
0:09:19 > 0:09:20Got some Katy Perry.
0:09:21 > 0:09:23So once we've got all the songs together,
0:09:23 > 0:09:27what we need to do is to extract the information from them.
0:09:27 > 0:09:30In effect, you're asking computers to listen to music.
0:09:30 > 0:09:32Erm...
0:09:32 > 0:09:34Tim Cowlishaw and Mi Tian
0:09:34 > 0:09:37have the job of reducing our songs to numbers.
0:09:37 > 0:09:42The basic idea is to turn the music, as something humans can hear, into
0:09:42 > 0:09:48machine-understandable data with meaningful information stored in it.
0:09:48 > 0:09:50MUSIC: Baby by Justin Bieber
0:09:50 > 0:09:53We're recording tempos, what instruments are present
0:09:53 > 0:09:56and what pitches are being played and when.
0:09:57 > 0:10:01We're even measuring the length and structure of each song.
0:10:01 > 0:10:04By the end, we've turned sound waves into this.
0:10:07 > 0:10:10These are the raw data of a single song,
0:10:10 > 0:10:13the information that makes it -
0:10:13 > 0:10:17the DNA of that song, if you will.
0:10:18 > 0:10:21We start with more than a million numbers per song
0:10:21 > 0:10:27and then distil them down to quantify its essence.
0:10:27 > 0:10:31I can't pretend that these numbers mean anything to me.
0:10:31 > 0:10:35Most of the features we've measured are hard for humans to interpret.
0:10:35 > 0:10:40To a scientist, numbers on this scale, on this magnitude,
0:10:40 > 0:10:42are beautiful.
0:10:42 > 0:10:44The difficulty comes in knowing what they mean.
0:10:44 > 0:10:47MUSIC: Ice Ice Baby by Vanilla Ice
0:10:47 > 0:10:50But, in truth, I don't have to know what these numbers mean or
0:10:50 > 0:10:52how they relate to what makes a song great.
0:10:52 > 0:10:56That's because we live in the age of machine learning.
0:10:56 > 0:10:59In the old way of doing science, the kind of science
0:10:59 > 0:11:02that I've done all my life,
0:11:02 > 0:11:07you look for causal associations between variables,
0:11:07 > 0:11:11the way in which one thing that you've measured in the world affects
0:11:11 > 0:11:16another, and you've got an explicit hypothesis about how that works.
0:11:16 > 0:11:19Machine learning doesn't go like that.
0:11:20 > 0:11:23Our computers know a track's success.
0:11:23 > 0:11:26They know which songs topped the charts for weeks...
0:11:29 > 0:11:32..and which scraped in at number 40.
0:11:33 > 0:11:37The machines then sift through our millions of data points.
0:11:37 > 0:11:41They're looking for those features that tend to be present in hits
0:11:41 > 0:11:43but absent in flops.
0:11:45 > 0:11:48The machine learning approach is to measure everything that you
0:11:48 > 0:11:54can possibly measure about a song and throw it into the pot
0:11:54 > 0:11:58and let the algorithm figure out what makes a hit.
0:11:58 > 0:11:59# Shake it off
0:11:59 > 0:12:01# Shake it off
0:12:01 > 0:12:02# Shake it off
0:12:02 > 0:12:04# Shake it off
0:12:04 > 0:12:05# Shake it off
0:12:05 > 0:12:07# Shake it off
0:12:07 > 0:12:08# Shake it off
0:12:08 > 0:12:10# Shake it off Oh-oh
0:12:10 > 0:12:11# Shake it off... #
0:12:11 > 0:12:15It'll take a lot of data-bashing to get any results,
0:12:15 > 0:12:18so it's a good time to introduce Trevor to my way of making music.
0:12:19 > 0:12:23For all the time that I've spent analysing pop music,
0:12:23 > 0:12:25I've never actually met a real live producer,
0:12:25 > 0:12:28never mind the man who made the '80s.
0:12:28 > 0:12:33# They send the heart police to put you under
0:12:33 > 0:12:35# Cardiac arrest... #
0:12:35 > 0:12:38- Ah, Armand.- You must be Trevor.
0:12:38 > 0:12:40I am. Come in.
0:12:40 > 0:12:42- Lovely to meet you. - Nice to meet you.
0:12:43 > 0:12:44Come through.
0:12:46 > 0:12:50The reason that this kind of study hasn't been done before is
0:12:50 > 0:12:55simply because up till now you've not been able to analyse that
0:12:55 > 0:12:59- number of songs...- Yeah. - ..in a quantitative, scientific way.
0:12:59 > 0:13:04- Right.- And I would claim that I, an evolutionary biologist,
0:13:04 > 0:13:06know more about popular music...
0:13:06 > 0:13:08- Right.- ..than anybody up till now.
0:13:08 > 0:13:10Well, what exactly are your qualifications...
0:13:10 > 0:13:13to be the person that knows the most about pop music?
0:13:13 > 0:13:15My PhD was on fruit flies,
0:13:15 > 0:13:20I've spent most of my professional life studying worms.
0:13:20 > 0:13:22Very important worms, as it so happens.
0:13:22 > 0:13:26The reason I claim that I think that I know more is because,
0:13:26 > 0:13:29in some ways, precisely because I am ignorant.
0:13:29 > 0:13:33- Right.- And because that means that I can just let the numbers talk to me.
0:13:33 > 0:13:36'But Trevor's not convinced that the high road to pop stardom
0:13:36 > 0:13:38'is paved with data.'
0:13:38 > 0:13:42Can I tell you, I have five things that I look at, right?
0:13:42 > 0:13:45If you want to be a successful artist, right?
0:13:46 > 0:13:48You have to be able to write
0:13:48 > 0:13:52or have access to the best material,
0:13:52 > 0:13:58you must have a really great voice, two octaves,
0:13:58 > 0:14:04you have to have personal charm and charisma...
0:14:04 > 0:14:07Computers can't measure personal charm and charisma.
0:14:07 > 0:14:10..you have to be physically and mentally strong. OK?
0:14:10 > 0:14:12The fifth one is you've got to want it.
0:14:17 > 0:14:20OK, forget about making her a star. THEY LAUGH
0:14:20 > 0:14:23- But that's what you're up against. - She may well have the qualities.
0:14:23 > 0:14:25- Make her a decent record. - Produce her, yeah...
0:14:25 > 0:14:30- Yeah.- ..a song that could feasibly, plausibly,
0:14:30 > 0:14:34be released and not disappear into the void of...
0:14:34 > 0:14:37- The void, the black hole of... - The black hole of YouTube.
0:14:37 > 0:14:41And if it works, we're going to go in business as a production company.
0:14:41 > 0:14:43- Right! - TREVOR LAUGHS
0:14:43 > 0:14:46And become...just a hit factory.
0:14:46 > 0:14:49MUSIC: Mrs Robinson by Simon & Garfunkel
0:14:49 > 0:14:52But we haven't just gathered modern music.
0:14:52 > 0:14:55I think we can use my techniques to see how pop
0:14:55 > 0:14:57has evolved over the years.
0:15:00 > 0:15:03The thing that I love about pop music is that it comes with its
0:15:03 > 0:15:08own meticulously-documented fossil record, the UK Official Charts.
0:15:09 > 0:15:13We've got the songs for about 50 years and we can study them all.
0:15:16 > 0:15:18Just like the modern tracks,
0:15:18 > 0:15:22these historic songs have been converted into numbers.
0:15:22 > 0:15:26But here we're looking at how the music changed over the years.
0:15:26 > 0:15:31Take 50 years of music, 17,916 songs,
0:15:31 > 0:15:33turn them into millions of numbers,
0:15:33 > 0:15:37boil those numbers down into a single variable
0:15:37 > 0:15:38and this is what you get.
0:15:40 > 0:15:43This is the rate of evolution of the UK charts
0:15:43 > 0:15:46over the last 50 years.
0:15:47 > 0:15:50This is what actually happened.
0:15:50 > 0:15:54When the red line is high, the music is evolving quickly.
0:15:54 > 0:15:55When low, slow.
0:15:55 > 0:15:58And when it crosses the yellow line,
0:15:58 > 0:16:00that's when the UK charts had a revolution.
0:16:01 > 0:16:05And it begins on a high, with a revolution.
0:16:07 > 0:16:10Our first revolution is centred around 1964,
0:16:10 > 0:16:13the year that gave us a new TV channel...
0:16:13 > 0:16:15- ARCHIVE:- BBC Two opening night.
0:16:15 > 0:16:19MUSIC: Rockin' Robin by The Hollies
0:16:19 > 0:16:21- ..pirate radio... - My name's Simon Dee,
0:16:21 > 0:16:22with you for the next two hours.
0:16:22 > 0:16:25First one off the top of the pile, The Hollies - Rockin' Robin.
0:16:26 > 0:16:30..and a pop chart in which musical evolution was in overdrive.
0:16:32 > 0:16:36We've forgotten what the sound of the early 1960s was.
0:16:36 > 0:16:39It was big, smooth orchestral numbers.
0:16:39 > 0:16:44People like Frank Sinatra, Ella Fitzgerald, Connie Francis
0:16:44 > 0:16:47were in the charts.
0:16:47 > 0:16:51It's like contemplating an age of dinosaurs
0:16:51 > 0:16:53before a mass extinction event.
0:16:53 > 0:16:56MUSIC: You Really Got Me by The Kinks
0:16:56 > 0:16:59It was music for grown-ups
0:16:59 > 0:17:00and it was doomed.
0:17:00 > 0:17:03# Girl, you really got me goin'... #
0:17:03 > 0:17:05And this is the sort of music that swept it away.
0:17:06 > 0:17:10This is The Kinks - You Really Got Me.
0:17:10 > 0:17:14Dave Davies is the man behind that crunching guitar.
0:17:14 > 0:17:15HE PLAYS GUITAR
0:17:15 > 0:17:19It's just G and F, you know.
0:17:19 > 0:17:21All the different things you can do with G and F.
0:17:21 > 0:17:24# You really got me now... #
0:17:24 > 0:17:26This was British rock and roll.
0:17:26 > 0:17:30# Oh yeah, you really got me now... #
0:17:30 > 0:17:32It was loud and sexy,
0:17:32 > 0:17:35nothing like the pretty orchestral stuff it replaced.
0:17:35 > 0:17:37# You really got me... #
0:17:39 > 0:17:42But if you play just the three bottom of notes on guitar...
0:17:46 > 0:17:48..sounds bigger
0:17:48 > 0:17:52than if it was played like a full chord, which would be...
0:17:52 > 0:17:53prettier but...
0:17:56 > 0:17:58When you really dig in...
0:17:58 > 0:18:01sounds more powerful, sexier.
0:18:03 > 0:18:04More aggressive, I guess.
0:18:06 > 0:18:08The word aggressive is key.
0:18:10 > 0:18:12These songs tend to have fewer harmonies,
0:18:12 > 0:18:16stronger rhythms and more thrashing guitars.
0:18:18 > 0:18:21So by combining these features into a single variable -
0:18:21 > 0:18:24aggression - we can see how the charts have changed.
0:18:26 > 0:18:31Aggression rises rapidly in '63, '64,
0:18:31 > 0:18:33moving through to '65,
0:18:33 > 0:18:36and we can go and look at the artists
0:18:36 > 0:18:37that are actually coming in here.
0:18:37 > 0:18:40Pretty Things, The Rolling Stones and, of course, The Who.
0:18:40 > 0:18:42# You say I've been in prison
0:18:42 > 0:18:45# You say I've got a wife... #
0:18:45 > 0:18:48The data shows the birth of British beat music,
0:18:48 > 0:18:51a musical mutation that swept all before it.
0:18:53 > 0:18:56Now, you will surely not be amazed to hear that there was
0:18:56 > 0:18:59a pop revolution around '64.
0:18:59 > 0:19:04But we've found it just by feeding the songs to a computer.
0:19:04 > 0:19:07The thing to remember is that none of this is based upon
0:19:07 > 0:19:10the standard cultural mythologies of pop,
0:19:10 > 0:19:14the hazy recollections of journalists and rock stars.
0:19:14 > 0:19:17This is based upon the music -
0:19:17 > 0:19:20the wave forms, the numbers.
0:19:20 > 0:19:24MUSIC: Love Me Do by The Beatles
0:19:24 > 0:19:26But the numbers conceal a surprise.
0:19:26 > 0:19:29# Love, love me do
0:19:29 > 0:19:31# You know I love you... #
0:19:31 > 0:19:33You may have noticed that I've not mentioned
0:19:33 > 0:19:36a certain four-piece that did quite well.
0:19:38 > 0:19:41'64 may have been the peak of Beatlemania
0:19:41 > 0:19:46but the data suggests that musically they weren't that important.
0:19:47 > 0:19:52And none of their charting singles sit high on our plot of aggression.
0:19:52 > 0:19:54Love Me Do, right on the average.
0:19:54 > 0:19:56Yellow Submarine, right on the average.
0:19:56 > 0:19:59Hey Jude, right on the average.
0:19:59 > 0:20:01Penny Lane...ah, a little bit below.
0:20:02 > 0:20:05It's the hallmark of The Beatles - average.
0:20:06 > 0:20:08Who isn't average?
0:20:08 > 0:20:11The Kinks aren't average, The Who aren't average,
0:20:11 > 0:20:14The Pretty Things aren't average, The Dave Clark Five aren't average,
0:20:14 > 0:20:18The Rolling Stones - they're certainly not average.
0:20:18 > 0:20:22The London bands are dragging mean aggression up
0:20:22 > 0:20:25and transforming the musical landscape.
0:20:25 > 0:20:27Meanwhile, Lennon and McCartney
0:20:27 > 0:20:29are writing ditties for prepubescent girls.
0:20:31 > 0:20:33Now, before you write into Points Of View,
0:20:33 > 0:20:35let me be clear.
0:20:35 > 0:20:38I'm not saying that the Fab Four weren't culturally important,
0:20:38 > 0:20:42that they didn't have winsome personalities and great haircuts.
0:20:46 > 0:20:48And I'll even concede that Sgt Pepper may well be the most
0:20:48 > 0:20:51influential album of all time.
0:20:51 > 0:20:52Or not.
0:20:52 > 0:20:53But the fact remains -
0:20:53 > 0:20:58The Beatles sat out the British revolution of 1964.
0:21:00 > 0:21:03My team are still searching for the numerical ingredients
0:21:03 > 0:21:04of a modern pop hit.
0:21:04 > 0:21:07PIANO PLAYS
0:21:10 > 0:21:13Meanwhile, Trevor's getting his first listen to the song
0:21:13 > 0:21:14we'll be working on.
0:21:20 > 0:21:22# If I couldn't read you
0:21:23 > 0:21:26# The signs say you've moved on
0:21:28 > 0:21:30# Cos when we talk and we walk
0:21:30 > 0:21:34# Down the lane of memories
0:21:36 > 0:21:42# Dive into the ocean with me
0:21:42 > 0:21:45# Cos if you stand still... #
0:21:45 > 0:21:47The song's called Dive.
0:21:47 > 0:21:50Nike wrote it about taking risks to achieve your dreams.
0:21:50 > 0:21:55# What we could be. #
0:22:06 > 0:22:08Good.
0:22:08 > 0:22:12I was going to say, you probably started singing in church, right?
0:22:12 > 0:22:15- Yes, I did.- Yeah, I can tell from the sound of the way you're singing.
0:22:15 > 0:22:17But, I mean, there's nothing wrong with that,
0:22:17 > 0:22:19that's where Dionne Warwick started to sing, church.
0:22:19 > 0:22:22You know? It's a great place to learn.
0:22:22 > 0:22:25I mean, great place to learn music, anyway.
0:22:25 > 0:22:27Yeah, it's a pleasant song, so...
0:22:27 > 0:22:31There's loads of people with pleasant songs,
0:22:31 > 0:22:33you've got to find some way of getting through all of them,
0:22:33 > 0:22:37- you know? But there's a couple of things that we can try.- OK.
0:22:37 > 0:22:39And the first thing would be to get the song out of you
0:22:39 > 0:22:41with a piano and a click.
0:22:41 > 0:22:42OK.
0:22:42 > 0:22:45METRONOME CLICKS
0:22:45 > 0:22:48PIANO PLAYS
0:22:48 > 0:22:52Science can't yet direct a singer to produce a perfect vocal,
0:22:52 > 0:22:54so I'm leaving this to Trevor.
0:22:54 > 0:22:56# If I couldn't read you
0:22:58 > 0:23:01# The signs say you've moved on... #
0:23:01 > 0:23:04Stop there. Just one more time.
0:23:04 > 0:23:07I think a teeny bit more edge than that, if you can do it.
0:23:07 > 0:23:10# Packed your past in a box on a ship
0:23:10 > 0:23:13# Sail till dawn. #
0:23:13 > 0:23:15Sorry to be a pain, to keep you doing it, but just try...
0:23:15 > 0:23:17Can you try it at 65?
0:23:20 > 0:23:22# Now you've had your freedom
0:23:23 > 0:23:27# You want to stay out in the cold
0:23:27 > 0:23:29# You think it's easier
0:23:29 > 0:23:32# When you're given much to hold... #
0:23:34 > 0:23:36That's good. That's what I meant.
0:23:38 > 0:23:41I can't even imagine, like, what the science is going to say
0:23:41 > 0:23:44or do to the song but it will definitely be interesting.
0:23:44 > 0:23:49# Into the ocean with me
0:23:49 > 0:23:53# Cos if you stand still... #
0:23:53 > 0:23:55I can't wait to hear what happens when I come back
0:23:55 > 0:23:57and what they've done with it.
0:23:57 > 0:24:02# What we could be. #
0:24:02 > 0:24:05I like the end. That was good.
0:24:05 > 0:24:08That was kind of a bit of Minnie Riperton there.
0:24:08 > 0:24:10Good. You got through it.
0:24:10 > 0:24:11Come here and take five.
0:24:13 > 0:24:15Yeah, I think you slowed it down.
0:24:15 > 0:24:18Oh, did I? Where do I start to slow down?
0:24:18 > 0:24:19At the top.
0:24:19 > 0:24:21THEY LAUGH
0:24:21 > 0:24:25MUSIC: Can't Get You Out Of My Head by Kylie Minogue
0:24:27 > 0:24:30The next job is to see if our analysis can turn Dive
0:24:30 > 0:24:32into a chart-topper.
0:24:36 > 0:24:40So, what we want to know is, what is the magic ingredient that
0:24:40 > 0:24:43makes a pop song a hit, right?
0:24:43 > 0:24:45As opposed to a non-hit.
0:24:46 > 0:24:49Ben Lambert's been going through the data.
0:24:49 > 0:24:52The hope was that through machine learning we'd find musical
0:24:52 > 0:24:55features that help us distinguish hits from flops.
0:24:56 > 0:24:59But the results are not very promising.
0:25:00 > 0:25:03So, 50-50 would be just picking randomly.
0:25:03 > 0:25:08And we get an accuracy of about 52% or 53%.
0:25:08 > 0:25:11- Right.- So, slightly better than just randomly picking
0:25:11 > 0:25:12but, basically, not.
0:25:14 > 0:25:17Yeah, it's not exactly an advertisement for machine learning, is it?
0:25:17 > 0:25:19- No. - THEY LAUGH
0:25:19 > 0:25:23If we can't identify the features that predict the relative
0:25:23 > 0:25:27success of a song, then it's really not clear what to tell Trevor.
0:25:27 > 0:25:31We run all the models we can think of, we tweak them in all
0:25:31 > 0:25:35the different ways we can, we subset the data...
0:25:35 > 0:25:37in all kinds of ways...
0:25:38 > 0:25:40..and we get nothing.
0:25:40 > 0:25:41We can't predict it.
0:25:41 > 0:25:44We just don't know what makes a hit.
0:25:44 > 0:25:47But there's a faint glimmer of hope,
0:25:47 > 0:25:50for we have identified one correlation.
0:25:50 > 0:25:53The one thing which we did find, and it's not a very,
0:25:53 > 0:25:57very strong signal, but it's a statistically significant signal,
0:25:57 > 0:26:00so that's something, is that there's an association
0:26:00 > 0:26:05between success in the charts and
0:26:05 > 0:26:09how close a song is to the average.
0:26:09 > 0:26:13MUSIC: Style by Taylor Swift
0:26:13 > 0:26:16Imagine a perfectly average song,
0:26:16 > 0:26:20one whose every feature sits at the centre of our distributions.
0:26:22 > 0:26:26Our analysis shows that such a song should do better than most.
0:26:27 > 0:26:31Of course, none of our songs actually hit that statistical
0:26:31 > 0:26:34sweet spot, but we can measure how close they are to it,
0:26:34 > 0:26:37and the closer a song gets, the better it seems to do.
0:26:41 > 0:26:43Perhaps, then, what we have to tell Trevor
0:26:43 > 0:26:46is to simply make Dive really average.
0:26:47 > 0:26:52I guess I would say that in all my years as a scientist,
0:26:52 > 0:27:00the discovery that the most average song tends to be the most successful
0:27:00 > 0:27:05song is one of the more depressing results that I have ever found.
0:27:07 > 0:27:11And I fear that Trevor won't like the news.
0:27:13 > 0:27:16The thing that I've found that is predictive of success
0:27:16 > 0:27:18is how average the music is.
0:27:18 > 0:27:25This is a force which is sort of driving music at any given time
0:27:25 > 0:27:27to some sort of...
0:27:27 > 0:27:30I don't want to call it a lowest common denominator, but sort of
0:27:30 > 0:27:35the centre of the distribution, a kind of a homogenising force.
0:27:35 > 0:27:39Do you think that there's a certain inevitability about that?
0:27:39 > 0:27:41Because we're in a unique position at the moment.
0:27:41 > 0:27:44We're in a position that no-one's ever been in before,
0:27:44 > 0:27:49where we have at least 50 years of, 60 years,
0:27:49 > 0:27:5270 years of recorded music.
0:27:52 > 0:27:54We can go back and we can listen to all of this music
0:27:54 > 0:27:56- from the '60s, the '50s.- Yes.
0:27:56 > 0:27:58Never been able to do that before.
0:28:00 > 0:28:02I suspect that Trevor's right.
0:28:02 > 0:28:06Today, in a few finger taps, you can hear almost any song
0:28:06 > 0:28:08in recorded history.
0:28:08 > 0:28:12Perhaps this explains why our algorithms have struggled here.
0:28:12 > 0:28:15If modern artists are combining genres promiscuously,
0:28:15 > 0:28:19the result will be the songs that are neither one thing nor the other.
0:28:22 > 0:28:26But this hasn't always been the case.
0:28:26 > 0:28:28Our pop history shows that the charts were once
0:28:28 > 0:28:32a bloody battleground in which genres vied for supremacy.
0:28:32 > 0:28:36We are back at our rate of evolution plot and you can
0:28:36 > 0:28:40see that the rate at which the music is changing in the charts
0:28:40 > 0:28:43begins to pick up in the 1970s.
0:28:43 > 0:28:47Around 1975 it crosses the line, we are into a revolution.
0:28:47 > 0:28:51It peaks in the late 1970s,
0:28:51 > 0:28:54when the music is changing with maximum speed.
0:28:54 > 0:28:57This revolution is going to become
0:28:57 > 0:29:01one of the most important in the history of British pop.
0:29:07 > 0:29:11When this revolution came, the country seemed half asleep.
0:29:11 > 0:29:14The UK was a sea of brown, orange and mustard.
0:29:15 > 0:29:18But something was stirring in the pop charts.
0:29:18 > 0:29:22# I am an antichrist
0:29:22 > 0:29:25# And I am an anarchist... #
0:29:25 > 0:29:29Ask a Brit of a certain age what happened in the late 1970s
0:29:29 > 0:29:34and chances are he'll say, "Well, mate, it was all about punk."
0:29:35 > 0:29:41There's no doubt that punk rock's cultural impact was immense,
0:29:41 > 0:29:49so much so that it's easy to forget just how tiny it all was.
0:29:49 > 0:29:53MUSIC: No Future by Sex Pistols
0:29:53 > 0:29:59Safety pins, spiky hair, spittle and swearing on national TV.
0:29:59 > 0:30:01Punk grabbed all of the headlines.
0:30:01 > 0:30:04Go on, you've got another five seconds, says something outrageous.
0:30:04 > 0:30:06- You dirty- BLEEP.
0:30:06 > 0:30:08- Go on, again. - You dirty- BLEEP.
0:30:08 > 0:30:09- What a clever boy(!)- BLEEP.
0:30:09 > 0:30:12But that's not the same as making music that mattered.
0:30:16 > 0:30:19This is a network of musical relationships
0:30:19 > 0:30:22between some 800 artists,
0:30:22 > 0:30:26pretty much everybody who charted in the UK in the 1970s.
0:30:26 > 0:30:31And it's based upon the music as measured by our computer
0:30:31 > 0:30:35and you can see that the relationships it gives make sense.
0:30:36 > 0:30:38The more songs of a particular genre,
0:30:38 > 0:30:40the bigger the block of colour.
0:30:40 > 0:30:44Up here, for example, we have funk and disco,
0:30:44 > 0:30:46they are all grouping together.
0:30:46 > 0:30:49Up here, we have vast swathes of soft pop.
0:30:49 > 0:30:54James Taylor, Joan Baez, Gordon Lightfoot, people like that.
0:30:54 > 0:30:59And down here, in yellow, we've got punk.
0:30:59 > 0:31:02And there's not very much of it.
0:31:02 > 0:31:07In the entire 1970s, only about 68 songs that could be called punk
0:31:07 > 0:31:10by any reasonable definition charted.
0:31:10 > 0:31:12MUSIC: Hersham Boys by Sham 69
0:31:12 > 0:31:15There just wasn't enough punk to have had a significant impact
0:31:15 > 0:31:18upon the evolution of the UK charts.
0:31:18 > 0:31:20So what was changing pop?
0:31:22 > 0:31:24In the '70s, we see songs becoming faster.
0:31:24 > 0:31:25MUSIC: We Are Family by Sister Sledge
0:31:25 > 0:31:27And more percusso.
0:31:27 > 0:31:29Again, by combining features
0:31:29 > 0:31:32we create a new variable, rhythmic intensity.
0:31:32 > 0:31:36It's the thing that changes in the 1970s.
0:31:36 > 0:31:38It begins to increase in 1972,
0:31:38 > 0:31:42climbs rapidly, peaks in 1979.
0:31:42 > 0:31:48Whatever's changing in the 1970s isn't punk, it's not rock,
0:31:48 > 0:31:50it's something else and it's coming from America.
0:31:50 > 0:31:54# We are family
0:31:54 > 0:31:56# I've got all my sisters with me... #
0:31:56 > 0:31:58It started as a funk invasion
0:31:58 > 0:32:02but quickly morphed into the music of glitter balls and flairs.
0:32:02 > 0:32:05# Get up everybody and sing... #
0:32:05 > 0:32:08Disco may have started in the black and gay clubs of New York City...
0:32:08 > 0:32:11MUSIC: Stayin' Alive by Bee Gees
0:32:11 > 0:32:13..but in the UK, it was Saturday Night Fever
0:32:13 > 0:32:16that sent it stratospheric.
0:32:16 > 0:32:20And the irresistible grooves of the Bee Gees ruled the charts.
0:32:22 > 0:32:27You can hear, you can feel, you can see
0:32:27 > 0:32:32and we can measure that rhythmic intensity, that driving beat.
0:32:34 > 0:32:36The data are unambiguous -
0:32:36 > 0:32:38a tidal wave of disco flooded the charts
0:32:38 > 0:32:42with pulsating four-on-the-floor rhythms.
0:32:42 > 0:32:46Punk arose, flourished and vanished in almost an instant.
0:32:48 > 0:32:51So why does punk rock seem to matter so much?
0:32:51 > 0:32:56I think it's a combination of British chauvinism, nostalgia,
0:32:56 > 0:33:00Johnny Lydon's charisma and Vivienne Westwood's clothes.
0:33:00 > 0:33:06But the fact of the matter is that as far as the music is concerned,
0:33:06 > 0:33:07it was never that special.
0:33:07 > 0:33:10# If I choose to believe... #
0:33:10 > 0:33:12- MUSIC REWINDS - # If I choose to believe you... #
0:33:12 > 0:33:13MUSIC REWINDS
0:33:13 > 0:33:15# If I choose to believe... #
0:33:15 > 0:33:18Trevor's working with arranger Julian Hinton.
0:33:18 > 0:33:19# Dive into the ocean... #
0:33:19 > 0:33:23Their first job is to produce Nike's song as they normally would.
0:33:23 > 0:33:28This is more of an organic feel so, because of the nature of the song,
0:33:28 > 0:33:30it's got a lot more emotion to it,
0:33:30 > 0:33:32it needs to have an ebb and flow.
0:33:32 > 0:33:35That's why I am being more painstaking.
0:33:36 > 0:33:38I'm going into a lot more detail
0:33:38 > 0:33:41so that hopefully it retains its performance
0:33:41 > 0:33:44and be the best version of what is essentially there.
0:33:44 > 0:33:46I'm not trying to...
0:33:46 > 0:33:48fix or change the character of it.
0:33:50 > 0:33:53The song is a ballad, so they're putting luscious strings
0:33:53 > 0:33:54underneath it.
0:33:54 > 0:33:58Whether they be from a keyboard or from an orchestra
0:33:58 > 0:34:03they're a very, very warm and expressive sound,
0:34:03 > 0:34:04probably one of the...
0:34:04 > 0:34:07Strings are probably one of the most expressive things
0:34:07 > 0:34:08other than the voice.
0:34:14 > 0:34:20# Dive into the ocean with me
0:34:20 > 0:34:23# I know there's danger
0:34:23 > 0:34:27# But this time I'm braver
0:34:29 > 0:34:34# Dive into the ocean with me... #
0:34:34 > 0:34:37- See just that bit there? - MUSIC STOPS
0:34:37 > 0:34:40- The dive, can you just make it dive at the right time?- Mm.
0:34:40 > 0:34:42He's hitting the water too late.
0:34:42 > 0:34:47# Dive into the ocean with me
0:34:48 > 0:34:52# I know there's danger
0:34:52 > 0:34:56# But this time I'm braver... #
0:34:57 > 0:35:00'With Trevor happy, it's time for me to see if I can get
0:35:00 > 0:35:03'Dive's features as close to the average as possible.'
0:35:03 > 0:35:04- Hi, guys.- Hi.
0:35:04 > 0:35:07'But to do that, we need a point of reference.
0:35:07 > 0:35:12'We need to hear some songs that our data show really are average.'
0:35:12 > 0:35:15- Oh, yeah, it's G-Eazy featuring... - Featuring Bebe Rexha.
0:35:15 > 0:35:17- Featuring Bebe Rexha.- Yeah.
0:35:17 > 0:35:20MUSIC: Me, Myself & I by G-Eazy and featuring Bebe Rexha
0:35:20 > 0:35:22Yeah.
0:35:22 > 0:35:24There's some sonic things going on,
0:35:24 > 0:35:27- there's a very electronic percussion.- Yep.
0:35:27 > 0:35:31And a very crisp electronic high-end as well, which is copied
0:35:31 > 0:35:37from one track to the next and is a really classic pop structure.
0:35:37 > 0:35:39MUSIC: You Don't Own Me by Grace
0:35:39 > 0:35:42Next up, Grace's You Don't Own Me.
0:35:42 > 0:35:44- Exactly.- Now here we've got a blossoming,
0:35:44 > 0:35:49very overtly uplifting sonic here.
0:35:49 > 0:35:52- It's got a '50s sound and it's got an '80s, '90s sound...- Yep.
0:35:52 > 0:35:55..and all kinds of other things in between in terms of,
0:35:55 > 0:35:58"Ooh, I can hear '70s influenced drums
0:35:58 > 0:36:01"or some strings from the 1950s going on in there."
0:36:01 > 0:36:05As Trevor predicted, our average songs seem to be a mix of everything
0:36:05 > 0:36:07that's gone before.
0:36:07 > 0:36:10In order for that track to become successful,
0:36:10 > 0:36:13people have got to be OK with all those sounds and textures.
0:36:13 > 0:36:16We come to a place of unbelievably...
0:36:16 > 0:36:18believable open-mindedness, actually.
0:36:18 > 0:36:22# Dive into the ocean with me... #
0:36:23 > 0:36:28'The current charts are a homogenised blend of earlier genres.
0:36:28 > 0:36:32'That suggests that we need to make Nike's song...'
0:36:32 > 0:36:33Nice.
0:36:33 > 0:36:36'..into a mishmash of them.'
0:36:36 > 0:36:38If I take that down to the original...
0:36:38 > 0:36:43- DRUMS PLAY - # Dive into the ocean with me... #
0:36:43 > 0:36:46It doesn't work...
0:36:46 > 0:36:47in that style.
0:36:47 > 0:36:49Let's take it up even more.
0:36:49 > 0:36:51- TRACK PLAYS FASTER - # This time I'm braver... #
0:36:51 > 0:36:54Now we're into sort of... It feels more like a remix.
0:36:54 > 0:36:57'But although we can estimate an average,
0:36:57 > 0:37:00'it's quite hard to define it musically.'
0:37:00 > 0:37:02TRACK PLAYS
0:37:02 > 0:37:04So that's the other groove I have.
0:37:04 > 0:37:06# Dive into the ocean with me... #
0:37:06 > 0:37:10I'm hearing, sort of, mid-Madonna, a little bit there.
0:37:10 > 0:37:12'The features themselves are hard to interpret,
0:37:12 > 0:37:16'so to find the centre of the charts, we have to experiment.'
0:37:16 > 0:37:22One of the things that a lot of the average tracks have,
0:37:22 > 0:37:25and I think this is precisely why they're average,
0:37:25 > 0:37:28is that they have this combination of a big melodic segment
0:37:28 > 0:37:32and then you've got a... Let's say pop dude, sort of, who comes in,
0:37:32 > 0:37:34does a bunch of rap.
0:37:34 > 0:37:36Can you give Nike's song that?
0:37:36 > 0:37:37In other words...
0:37:37 > 0:37:38can we get a rap in there?
0:37:40 > 0:37:41Jul?
0:37:41 > 0:37:43HE LAUGHS
0:37:43 > 0:37:44I'm sorry,
0:37:44 > 0:37:47are you just googling "rap a cappella edify bmp"?
0:37:47 > 0:37:51And I have 118,000 hits.
0:37:51 > 0:37:54RAP PLAYS ON COMPUTER
0:37:54 > 0:37:56Er...
0:37:56 > 0:37:58- RAPS:- # So many people give their religion
0:37:58 > 0:38:01# The music and church let you know how you're livin'... #
0:38:01 > 0:38:06'Now this is Crash DDZ, a rapper from Kentucky.'
0:38:07 > 0:38:09DRUMS PLAY
0:38:09 > 0:38:13# Everything happens for a reason... #
0:38:13 > 0:38:15HE LAUGHS
0:38:17 > 0:38:20Who's controlling this, by the way? Is it you or you?
0:38:20 > 0:38:23- That was a collaboration, actually. - That was a collaboration.
0:38:23 > 0:38:25- Seriously?- A classic production collaboration.
0:38:25 > 0:38:30# We can sense our defences come down so easily
0:38:30 > 0:38:33- RAPS:- # Man would stand, guitar in his hand
0:38:33 > 0:38:36# Recording artists, the history, tell us, travelling bands
0:38:36 > 0:38:39# Travelling bands, Johnny Cash and clones... #
0:38:39 > 0:38:42'In a few mouse clicks, and for 1,
0:38:42 > 0:38:44'we had a rap to pair with Nike's vocals.'
0:38:46 > 0:38:49This is going to be one happy rapper.
0:38:49 > 0:38:53Until we measure Dive, we won't know how close we've got to our goal
0:38:53 > 0:38:56of making its features average.
0:38:56 > 0:39:00So it's not clear how useful my analysis will be.
0:39:00 > 0:39:02# Dive into the ocean with me... #
0:39:02 > 0:39:04'But what is clear is that music production
0:39:04 > 0:39:07'is already very reliant upon technology.'
0:39:09 > 0:39:13The tools that these guys are using, high-end computers,
0:39:13 > 0:39:16various kinds of programmes, different kinds of algorithms,
0:39:16 > 0:39:19they're utterly familiar to me and yet their product,
0:39:19 > 0:39:24the things that come out of this stuff, it's...
0:39:24 > 0:39:25Well, it's like magic.
0:39:27 > 0:39:30And there's no doubt that the influence of technology
0:39:30 > 0:39:34on the last three decades of the pop charts has been immense.
0:39:34 > 0:39:38But it doesn't show up in our data where you might expect.
0:39:38 > 0:39:42We've seen a revolution in the 1960s, British rock and roll.
0:39:42 > 0:39:45We've seen another revolution in the 1970s,
0:39:45 > 0:39:49the rise of disco and funk. But what about the 1980s?
0:39:49 > 0:39:53The music's changing, it's always changing,
0:39:53 > 0:39:55it's not just changing very fast
0:39:55 > 0:39:58and there's nothing resembling a revolution.
0:39:58 > 0:40:01MUSIC: Take On Me by A-ha
0:40:01 > 0:40:05This may be a surprise - the 1980s was the decade of A-ha...
0:40:06 > 0:40:09MUSIC: Karma Chameleon by Culture Club
0:40:09 > 0:40:11..Culture Club...
0:40:11 > 0:40:15MUSIC: Never Gonna Give You Up by Rick Astley
0:40:15 > 0:40:18..and, of course, Rick Astley.
0:40:20 > 0:40:23But while there was no pop revolution in the '80s,
0:40:23 > 0:40:25I do think that it contained the seeds
0:40:25 > 0:40:28of the greatest revolution of them all...
0:40:30 > 0:40:32..the rise of the machines.
0:40:35 > 0:40:38As soon as technology came along, you could dream something,
0:40:38 > 0:40:41you could make it happen, you know?
0:40:41 > 0:40:43If you sort of dreamt something
0:40:43 > 0:40:46and you tried to make it happen in the '70s, it was harder
0:40:46 > 0:40:47because you had to get people to play it.
0:40:51 > 0:40:54In the wake of Kraftwerk's pioneering electronica,
0:40:54 > 0:40:58the '80s and '90s brought the arrival of increasingly versatile
0:40:58 > 0:41:00drum machines and synthesisers.
0:41:02 > 0:41:05In a few clicks, producers could make any sound they wanted.
0:41:06 > 0:41:09So where is tech's influence on the evolution of pop?
0:41:11 > 0:41:16We just need to plot our data in a different way.
0:41:16 > 0:41:21If we take 17,916 songs - all of the songs in our data
0:41:21 > 0:41:24over the history of the UK charts - and plot them
0:41:24 > 0:41:27on a single graph of rhythmic intensity,
0:41:27 > 0:41:30what we see is something that looks like this.
0:41:30 > 0:41:33The key to this plot is the vertical range of music
0:41:33 > 0:41:35at any given point in time.
0:41:36 > 0:41:39That tells us how much rhythmic variety is in the charts.
0:41:39 > 0:41:44We begin in the 1960s. Down here we've got low-intensity stuff,
0:41:44 > 0:41:45Shirley Bassey.
0:41:45 > 0:41:48Up here, Chubby Checker.
0:41:48 > 0:41:53They sound very different but the range is relatively small.
0:41:53 > 0:41:54We move along.
0:41:54 > 0:41:59The average is changing but that's not actually the big story here.
0:41:59 > 0:42:03As we progress into the early '80s,
0:42:03 > 0:42:08we start seeing a new world of high-intensity music.
0:42:08 > 0:42:10This is electronica, Kraftwerk,
0:42:10 > 0:42:14Cabaret Voltaire, Bam Bam,
0:42:14 > 0:42:17dance music, house, techno,
0:42:17 > 0:42:21all this stuff is coming in and it expands even further.
0:42:25 > 0:42:29This expanding space is the third great story in our history of pop.
0:42:30 > 0:42:34It's the relentless advance of electronic dance music.
0:42:35 > 0:42:39DJ Annie Mac explains why she thinks it's so successful.
0:42:40 > 0:42:44It's the collective experience of 100,
0:42:44 > 0:42:471,000 people standing on a floor...
0:42:47 > 0:42:49# The weekend is coming up... #
0:42:49 > 0:42:51..all experiencing the same thing.
0:42:51 > 0:42:55That kind of physical collective experience
0:42:55 > 0:42:57is a very beautiful thing.
0:42:58 > 0:42:59If you say so.
0:43:03 > 0:43:08But the most striking thing is how many sub genres dance has spawned.
0:43:08 > 0:43:12House, techno, funky, grime,
0:43:12 > 0:43:17drum and bass, jungle, hard-core, break beat, big beat.
0:43:19 > 0:43:22Liquid drum and bass, progressive house.
0:43:22 > 0:43:25And it's easy to see why this diversity flourished.
0:43:25 > 0:43:29I would say technology is definitely a massive reason for
0:43:29 > 0:43:34why dance music, electronic music has been so prone to expansion
0:43:34 > 0:43:35and evolution and fragmentation.
0:43:35 > 0:43:38It's one of the most exciting things, I think,
0:43:38 > 0:43:39about electronic music.
0:43:42 > 0:43:47# Dive into the ocean with me
0:43:47 > 0:43:49# I know there's danger... #
0:43:49 > 0:43:54Back in the studio and we're still trying to make Dive truly average,
0:43:54 > 0:43:55a song that mashes everything together
0:43:55 > 0:43:59to sit at the statistical centre of the charts.
0:43:59 > 0:44:02And we think we need a fair chunk of dance intensity.
0:44:04 > 0:44:07- How much bass have you guys put on this thing?- How much bass?
0:44:07 > 0:44:09Yeah.
0:44:09 > 0:44:11I mean...
0:44:11 > 0:44:15is it the case that we can push this towards more of a dance song?
0:44:15 > 0:44:17That's really the nub of the problem.
0:44:17 > 0:44:22This is a ballad and ballads are a completely different thing and,
0:44:22 > 0:44:26you know, if you think about the best dance songs of all time,
0:44:26 > 0:44:30- say something like Boogie Wonderland by Earth, Wind & Fire...- Yep.
0:44:30 > 0:44:33..the tune in the verse is all off the beat.
0:44:33 > 0:44:36HE SINGS THE MELODY OF BOOGIE WONDERLAND
0:44:36 > 0:44:40It's all off the beat and it dances on top of the track.
0:44:40 > 0:44:44Most dance songs don't start out life as ballads, you know?
0:44:44 > 0:44:48We're going to try one last push to make Nike's song really,
0:44:48 > 0:44:50really average.
0:44:50 > 0:44:54# Dive into the ocean with me... #
0:44:54 > 0:44:57We've even brought her back to do a fresh, faster vocal.
0:44:57 > 0:45:01# But this time I'm braver... #
0:45:01 > 0:45:07- Given that you have a math degree from Imperial College...- Yes.
0:45:07 > 0:45:08..can you imagine just...
0:45:10 > 0:45:13..setting yourself up as a music analyst from now on?
0:45:13 > 0:45:16You know, every song you write, it's just going to be,
0:45:16 > 0:45:19"Hmm, a little standard deviation away from the mean,
0:45:19 > 0:45:22"got to move in there."
0:45:22 > 0:45:24Well, maybe we'll see how this works.
0:45:24 > 0:45:27Yes, I think perhaps we should.
0:45:27 > 0:45:29If it works then maybe.
0:45:29 > 0:45:31DANCE VERSION OF DIVE PLAYS
0:45:31 > 0:45:34'We've created a few versions of Dive in the hope
0:45:34 > 0:45:37'that one will stick closely to the chart average,
0:45:37 > 0:45:39'our key to pop success.
0:45:39 > 0:45:42'But some versions just don't work.'
0:45:42 > 0:45:43All right...
0:45:43 > 0:45:45- MUSIC STOPS - Enough.
0:45:45 > 0:45:49Yeah, we can only apologise to Nike, for doing that to her song.
0:45:49 > 0:45:51No, but I would...
0:45:51 > 0:45:54'Using data to make a hit is proving to be a challenge.'
0:45:56 > 0:45:57But I've got another idea.
0:45:59 > 0:46:01Thanks to the rise of technology,
0:46:01 > 0:46:04we now live in a world of bedroom producers.
0:46:04 > 0:46:07There's an ocean of undiscovered artists out there,
0:46:07 > 0:46:10and some of them might even be competent.
0:46:15 > 0:46:18Can algorithms find the stars of the future?
0:46:18 > 0:46:20It's time for another experiment.
0:46:20 > 0:46:22And to conduct it, I don't need to go far.
0:46:23 > 0:46:26That's because the BBC is home to Introducing,
0:46:26 > 0:46:30a website to which unsigned artists can upload their music,
0:46:30 > 0:46:32music that we can analyse.
0:46:35 > 0:46:39What I have here is a hard drive containing 1,786 songs.
0:46:39 > 0:46:43This is the raw material of evolution,
0:46:43 > 0:46:47unfiltered by any company or broadcaster
0:46:47 > 0:46:50or any consideration of taste other than the musicians' own.
0:46:51 > 0:46:56And what we want to know is, is any of it any good?
0:46:56 > 0:46:59Normally each track is vetted by a human being
0:46:59 > 0:47:03but I suspect that machines can do their job just as well.
0:47:03 > 0:47:08What we do is we teach the computer, using a machine-learning algorithm,
0:47:08 > 0:47:11what the charts are now
0:47:11 > 0:47:15and then we apply that model to the Introducing data
0:47:15 > 0:47:20and we ask which of the Introducing songs are most chart-like,
0:47:20 > 0:47:24which does the computer think are most likely to go into the charts?
0:47:27 > 0:47:30As before, our computers measure each song.
0:47:30 > 0:47:32HE MUTTERS
0:47:32 > 0:47:34'This is the same process we went through
0:47:34 > 0:47:36'for all the chart music earlier on,'
0:47:36 > 0:47:38it's just sort of MP3 files go in one end
0:47:38 > 0:47:41and spreadsheets come out the other.
0:47:41 > 0:47:46The result, a simple list with the most chart-like songs at the top.
0:47:46 > 0:47:52Out of all those songs, our algorithm picked one.
0:47:52 > 0:47:58And it's a song called Margarita by a group called The Modern Strangers.
0:48:01 > 0:48:05The Modern Strangers were actually soon to play a gig in London.
0:48:06 > 0:48:07So, two weeks later,
0:48:07 > 0:48:11I found myself heading to a dingy club to hear them.
0:48:13 > 0:48:16Margarita was the finale.
0:48:16 > 0:48:19# Sit back, Margarita
0:48:19 > 0:48:22# Nice and slow
0:48:22 > 0:48:24# Oooh
0:48:24 > 0:48:31# Baby, all I ever wanted was your love
0:48:31 > 0:48:32# Ooh... #
0:48:32 > 0:48:35Catchy. It certainly had people dancing.
0:48:43 > 0:48:47# Sit back, Margarita... #
0:48:47 > 0:48:50It's the first time I've really listened to the song.
0:48:50 > 0:48:53The computer just gave us a name.
0:48:53 > 0:48:57None of us had actually heard the thing.
0:48:57 > 0:49:01And I've got to say, it's amazingly convincing.
0:49:07 > 0:49:11Margarita is almost an old school disco track -
0:49:11 > 0:49:15it comes straight in with a big beat and melodical.
0:49:20 > 0:49:25The lyrics may be lacking but it's very, very danceable.
0:49:28 > 0:49:29The computer is dumb.
0:49:31 > 0:49:35It doesn't have a sophisticated model of beauty
0:49:35 > 0:49:38or danceability.
0:49:38 > 0:49:41But here's the thing - this song is great.
0:49:41 > 0:49:44And people think it's great.
0:49:44 > 0:49:48It seems to have bottled some musical magic.
0:49:49 > 0:49:56And our computer algorithm has found that same magic, it's...
0:49:58 > 0:50:01It shows that it can be bottled by math.
0:50:01 > 0:50:03And that's rather amazing.
0:50:03 > 0:50:09# Baby, all I ever wanted was your love
0:50:09 > 0:50:13# Ooh. #
0:50:13 > 0:50:16Thank you very much for having us! Have a good evening.
0:50:18 > 0:50:21But every experiment needs a control.
0:50:21 > 0:50:26BBC Music Introducing in Kent with Abbie McCarthy.
0:50:26 > 0:50:28Good evening, it's after eight o'clock...
0:50:28 > 0:50:31I want to pit my algorithm against some human competition.
0:50:31 > 0:50:33..right here in Kent.
0:50:34 > 0:50:39Abbie McCarthy is the Introducing DJ at BBC Radio Kent.
0:50:39 > 0:50:42We get sent probably about 300 tracks a week,
0:50:42 > 0:50:44so there's lots of music to listen through,
0:50:44 > 0:50:46and then we're just looking for a song that really stands out,
0:50:46 > 0:50:48whether that it's really well produced,
0:50:48 > 0:50:50it's got a really good beat to it.
0:50:50 > 0:50:51# I am scared... #
0:50:51 > 0:50:54Then sometimes we have a moment where we're like,
0:50:54 > 0:50:57"Wow, this song's really, really incredible."
0:50:57 > 0:51:00# Oh-oh-oh... #
0:51:00 > 0:51:04Abbie's also picked a song - Gold by singer-songwriter Shells.
0:51:07 > 0:51:10I'm going to play both her choice and mine to Rhys Hughes,
0:51:10 > 0:51:13the head of programming at Radio 1.
0:51:13 > 0:51:15Now, it's not going to be a proper experiment,
0:51:15 > 0:51:17but if I played you two songs, would you warrant that you could
0:51:17 > 0:51:21pick out the one that Abbie picked and the one that the machine picked?
0:51:21 > 0:51:23I've got a 50% chance, haven't I?
0:51:23 > 0:51:25You do, that's why it's not a very good experiment but...
0:51:25 > 0:51:27Song number one.
0:51:27 > 0:51:30MUSIC: Margarita by The Modern Strangers
0:51:30 > 0:51:33# Sit back, Margarita
0:51:33 > 0:51:34# Nice and slow... #
0:51:34 > 0:51:36Get the idea?
0:51:36 > 0:51:38- Get the idea?- Yeah.
0:51:38 > 0:51:39Next up, Abbie's song.
0:51:39 > 0:51:43MUSIC: Gold by Shells
0:51:43 > 0:51:46# Oh, oh, oh
0:51:46 > 0:51:50# Oh, everybody's made of gold
0:51:50 > 0:51:53# So put me in black and white... #
0:51:53 > 0:51:54MUSIC STOPS
0:51:54 > 0:51:58One of those two songs, Margarita and Shells, was chosen by computer,
0:51:58 > 0:52:01the other one by DJ.
0:52:01 > 0:52:02Which was which?
0:52:04 > 0:52:10I would say the computer picked the second one you've played.
0:52:10 > 0:52:13Wrong. The computer picked Modern Strangers, it picked the first one.
0:52:13 > 0:52:15- Right, OK.- Yeah?
0:52:15 > 0:52:18So you have to concede that at least our algorithm
0:52:18 > 0:52:20is doing as well as a DJ.
0:52:20 > 0:52:23What do you say to just ditching them all and replacing them
0:52:23 > 0:52:24with a computer?
0:52:24 > 0:52:27I don't think an algorithm would've picked out a Bob Dylan,
0:52:27 > 0:52:30I don't think an algorithm would've picked out a David Bowie.
0:52:30 > 0:52:35I don't think a computer can understand the emotional response
0:52:35 > 0:52:37that you get to a record.
0:52:37 > 0:52:42You know, we all have records that we know that make us happy,
0:52:42 > 0:52:45records that make us incredibly sad,
0:52:45 > 0:52:48records that make, you know, you want to jump around your bedroom
0:52:48 > 0:52:51and, you know, throw shapes in the mirror.
0:52:51 > 0:52:53And I don't think an algorithm can do that.
0:52:53 > 0:52:56Isn't it really just a matter of time before
0:52:56 > 0:52:58this rather romantic vision buckles?
0:52:58 > 0:53:01- I think...- And you will have machine-learning algorithms
0:53:01 > 0:53:05- and you are going to become as teched up as Google.- Yeah.
0:53:05 > 0:53:08We are, I mean, but I'm an incurable romantic
0:53:08 > 0:53:13and I think that... I think, you know, that the human voice
0:53:13 > 0:53:18and the human passion will always, will always win through.
0:53:18 > 0:53:19Thank you so much.
0:53:19 > 0:53:21Oh, thank you for exposing me on national television
0:53:21 > 0:53:22in getting it wrong!
0:53:24 > 0:53:27I think I've shown that we can indeed pick fantastic songs
0:53:27 > 0:53:29without listening to a note.
0:53:29 > 0:53:33But it doesn't look as though I'll be selling my algorithms to Radio 1
0:53:33 > 0:53:35any time soon.
0:53:35 > 0:53:39As for helping Trevor, well, my team have finished analysing
0:53:39 > 0:53:44all the versions of Dive and the results are frankly unimpressive.
0:53:44 > 0:53:49Here we've plotted the distance of every one of the songs
0:53:49 > 0:53:54in the charts over the last year or so from the centre of the charts,
0:53:54 > 0:53:56from their average.
0:53:56 > 0:54:00Right at the centre is Taylor Swift's Style.
0:54:00 > 0:54:02Maybe that's why she's so successful.
0:54:02 > 0:54:04And here you can see the problem,
0:54:04 > 0:54:07here we have ever-increasing distance from the centre.
0:54:07 > 0:54:11Dive is out here at the edge of musical space,
0:54:11 > 0:54:15and it's not as though we didn't try to push it down here somewhere -
0:54:15 > 0:54:19we gave it a bit of oomph, we upped the tempo,
0:54:19 > 0:54:21we thought, "Maybe it needs more bass,"
0:54:21 > 0:54:24so we gave it more of that. But whatever we tried,
0:54:24 > 0:54:27it just pushed it further out into musical space.
0:54:27 > 0:54:30We're not even getting near where we need to be.
0:54:31 > 0:54:34And here's the thing, we can measure them,
0:54:34 > 0:54:38we can plot them, but so far,
0:54:38 > 0:54:39we don't know how to move them.
0:54:39 > 0:54:42# If you stand still... #
0:54:42 > 0:54:46All that's left is to break the news to the team.
0:54:46 > 0:54:48- RAPS:- # So many people give to religion
0:54:48 > 0:54:50# The music and church let you know how you're livin'
0:54:50 > 0:54:53# Save a part of life and give till it hurts
0:54:53 > 0:54:55# Creativity... #
0:54:55 > 0:54:57Well, erm...
0:54:58 > 0:55:00So, er...
0:55:01 > 0:55:03I like the song.
0:55:03 > 0:55:07It's really nice and, erm, the question that we've asked is,
0:55:07 > 0:55:13how far are those songs from the centre of the charts?
0:55:13 > 0:55:15And the answer is...
0:55:20 > 0:55:23..that they're all really far away.
0:55:23 > 0:55:25HE LAUGHS
0:55:25 > 0:55:27- So my thinking is...- Yeah.
0:55:27 > 0:55:32..that to the degree that we can tell you anything useful...
0:55:34 > 0:55:40..it should be to ignore what we are saying.
0:55:40 > 0:55:45'It seems that the recipe for pop success will remain hidden for now.
0:55:45 > 0:55:49'Even hit-makers sometimes don't know it until they hear it.'
0:55:49 > 0:55:53But sometimes when you're going through the process, things happen.
0:55:53 > 0:55:55Just the way the singer sings that line
0:55:55 > 0:55:58that you didn't think sounded that great,
0:55:58 > 0:56:01but suddenly when this person sings it it just sounds amazing
0:56:01 > 0:56:04and then suddenly, man, you have something, you know?
0:56:04 > 0:56:09And you might have to work for days to get something really special.
0:56:09 > 0:56:12It doesn't happen all that often, it's hard to find.
0:56:12 > 0:56:16- We'd call it non-linear effect. - Non-linear effect, yeah.
0:56:16 > 0:56:19And non-linear effects,
0:56:19 > 0:56:23- they're the hardest for us to get at nowadays.- Yeah.
0:56:23 > 0:56:25No, I can imagine.
0:56:25 > 0:56:32We had hoped to show how science could help
0:56:32 > 0:56:39a few really talented musicians how to make a great hit song.
0:56:39 > 0:56:42And I have to say that we have singularly failed.
0:56:42 > 0:56:44DANCE VERSION OF DIVE PLAYS
0:56:47 > 0:56:50That's not to say that Dive isn't lovely,
0:56:50 > 0:56:52and in exploring the musical space
0:56:52 > 0:56:57I suppose that science did play a part, but only a modest one.
0:57:00 > 0:57:04If Dive is any good, and I think it's great,
0:57:04 > 0:57:07it's because Nike, Trevor and his team made it so.
0:57:07 > 0:57:11# Defences come down so easily
0:57:11 > 0:57:15# Dive into the ocean with me... #
0:57:15 > 0:57:16Wow.
0:57:18 > 0:57:20Wow. That's very different.
0:57:21 > 0:57:24I'd never imagined as a dance track, no, but I like it,
0:57:24 > 0:57:29I think it's good. I really liked the synths, I liked the rhythms,
0:57:29 > 0:57:31I liked some of the chord changes.
0:57:31 > 0:57:34My head was kind of bopping to it.
0:57:35 > 0:57:40You may doubt that we can capture creativity,
0:57:40 > 0:57:43that we can bottle art. And I have to agree,
0:57:43 > 0:57:47it hasn't been very convincing. And yet, I still think we can.
0:57:48 > 0:57:51After all, we've seen the power of data.
0:57:51 > 0:57:54It's shown us how pop evolved...
0:57:57 > 0:58:00..and how to find great new music.
0:58:00 > 0:58:03There's nothing intrinsically mysterious about this,
0:58:03 > 0:58:08it is after all just physics and neurobiology.
0:58:08 > 0:58:10# Dive into the ocean with me... #
0:58:10 > 0:58:14Given enough computational power, given enough data,
0:58:14 > 0:58:16we will work it out.
0:58:16 > 0:58:17When?
0:58:17 > 0:58:20Not by the end of this programme, but we will.
0:58:20 > 0:58:24# Dive into the ocean with me
0:58:24 > 0:58:27- # I'm braver - Yes, I'm braver
0:58:27 > 0:58:31- # Braver - Yes, I'm braver
0:58:31 > 0:58:36# Dive into the ocean with me
0:58:36 > 0:58:38# Cos if you stand still
0:58:38 > 0:58:42# You'll never, ever, ever see what we could be
0:58:42 > 0:58:46RAPPING
0:58:46 > 0:58:48# What we could be
0:58:48 > 0:58:53RAPPING
0:58:53 > 0:58:55# What we could be. #