Science Interviews


Mon, 11th May 2015

How has music evolved in 50 years?

Matthias Mauch, Queen Mary University of London

Listen Now    Download as mp3 from the show Safety at 40,000 Feet

Pretty much everyone enjoys listening to music, but are you the kind of person who believes pop music went completely downhill after The Beatles, or do you love bopping along to a bit of Playing guitarRihanna or rap? One person who's taking a more scientific approach to discovering how musical tastes have changed over the years is Matthias Mauch, from Queen Mary University of London, who has been analysing thousands of songs from the US Billboard hot 100 chart to track the evolution of popular music. Kat Arney went to meet him...

Matthias - We were interested in analysing the evolution of popular music, so we needed a dataset. I was working in Last FM so we had access to a lot of recordings and we also had access to the Billboard Charts. So, we put the two things together and we had this beautiful dataset of 50 years of US Charts, Singles Charts, Hot 100. 17,000 songs that we simply had on our hard drives. Now, what do you do with this? Luckily enough, I specialise, as some other people around the world do, in music informatics and this is exactly our business. So, we are scientists who develop new algorithms to extract information from music and thatís exactly what we did.

Kat - So, you werenít listening to 17,000 songs. You basically made a computer listen to it and tell you, what does this sound like, what's in here?

Matthias - Thatís exactly right. Trust me, we listen to a lot of them, but not nearly 17,000. This sort of study is only possible if you use computers.

Kat - What sort of things were you looking for? What were you analysing the songs for?

Matthias - So, we were analysing the songs mainly for harmony and timbre. Harmony, we analysed chord changes. So, I was really interested in trying to understand how the song writing in terms of chord changes are changed. In terms of timbre, this is merely like sound colour. So, we could identify dark or aggressive or percussive sounds.

Kat - What sort of patterns do you see in the data?

Matthias - We do have some beautiful waves. There are some other patterns weíve got like a camel-like, two humped, pattern that is aggressive guitaring music. And that was once prevalent in the early Ď70s, or around the early Ď70s, and once around the late Ď80s. so, that was kind of the early rock, and then it comes again as stadium rock. Yeah, we've got beautiful patterns. Well I've talked to you now about these timbre patterns, so letís have a look at some harmonic patterns. So, to me, the fascinating one is the decline of the dominant 7th chord. Now, for your listeners who are not musically inclined, dominant 7th chords are very important in blues and jazz. In blues particularly, they give the dissonance in the dominant 7th chord, it gives it a sort of dirty colour. And these chords seem to be vanishing in the charts. So, even by 1975 or something, they'd almost completely gone and never came back.

Kat - Were there any chords that replaced them?

Matthias - Yes, there certainly were. So then we can see around 1970, suddenly, there was this new kind of chord. It was not exactly new. They have been used before, but not very much. And then suddenly, as funk came into the charts and slowly turned into disco, these minus 7th chords were suddenly prevalent. Indeed, if you listen to Night Fever by the Bee Gees, you'll hear lots of these dangly minus 7th chords. So, they came in and stayed pretty much for a long time. Later, around in the Ď90s, there came this new kind of harmonic style that came in to the charts and we looked at it and we were wondering what it was. It was no harmonic. So, that was quite exciting to us and we saw that suddenly, rap and hip-hop, these songs dared essentially to do away with harmony altogether pretty much and focus on something completely different.

Kat - Did you see music evolving or were there just step changes where it became very different?

Matthias - This is one of the fascinating things, that to me are fascinating, but donít really make the news very much because it seems so normal. But actually, what you say is kind of the reason why we call it evolution because what we see is that charts donít change just like that from one year to the next. They donít change. They might change rapidly, but there's always some lag to it. What happens is that people usually take something that exists, recombine it and change it, and then make new music. So, itís very rare that music is created from a vacuum, if itís possible at all.

Kat - People basically like listening to things that are kind of familiar. They might know something, but then itís got something new in it thatís coming from somewhere else.

Matthias - Obviously, composers and songwriters donít write the same song again - except sometimes theyíve tried. But if they were to write exactly the same song then obviously, no one would buy it once more, right?

Kat - Some people would say, well, we donít make music like the old days and itís all rubbish now. Was there any algorithm that could tell you about the quality of the songs whether they were actually any good?

Matthias - Yeah, we wish we could. So, we donít actually have that. So, we have to admit that our algorithms - some musicologists can analyse few songs really well and we can analyse very many songs not quite so well - our measures are more crude. But what we can say is more generally perhaps about the whole charts. So, we can't really say how good one song is. But we can say how diverse, how exciting the chart as a whole is. So, would you see many different sounds in the charts? So, if you randomly turned on the radio in a chart show, would you get nice, excitingly different ones? That turned out to be quite an interesting exercise because we realised that in the midĎ80s, that was much less the case than in other eras.

Kat - So, the charts that we have now and the charts before the Ď80s were a lot more different types of music in there.

Matthias - That seems so. Of course, our study went on only to 2010, but yes. So, it seems as if there was quite a lot of diversity in the charts then it dipped in the Ď80s and then it came back. Partly because rap added these fresh notes and you had a new kind of style in the charts that made the charts as a whole more diverse and perhaps interesting.

Kat - Do you have a favourite decade or did you even find a favourite year where you're like, ďYeah, I love all those songsĒ?

Matthias - Personally, I quite like the Ď70s actually. I'm a bit of a Ď70s person I think.


Subscribe Free

Related Content


Make a comment

See the whole discussion | Make a comment

Not working please enable javascript
Powered by UKfast
Genetics Society