Is There a Formula For a Hit Song?

moveoverrover writes "What happens when two Rutgers grad students analyze 50 years of Billboard Top 10 hits with MIT offshoot Echo Nest's API and turn the data into visualizations for an assignment? Great looking visualizations for one, and a fascinating look at 50 years of Pop music at the data level. Posing the question, 'Is there a formula for a hit song?' the students write, 'What if we knew, for example, that 80% of the Billboard Hot 100 number one singles from 1960-2010 are sung in a major key with an average of 135 beats per minute, that they all follow a I-III-IV chord progression in 4/4 time signature, and that they all follow a "verse-verse-chorus-verse-chorus-bridge-chorus" sequence structure?' Using data extracted by Echo Nest on tempo, duration, time signature, musical key, as well as subjective criteria like "energy" and "danceability," the pair generated a number of visualizations with Google Motion Charts (warning: slow) and '(some) Tableau Results' for everyone to see and investigate. Curious about tempo and song duration trends in Pop music over 50 years? Correlation between record label and song tempo? Download the core data, the Tableau reader and look at it any way you want."

Hot Song Science is old news

[tepples]

Hit Song Science has been around since 2003. See previous Slashdot story [slashdot.org].

Scum

[Beelzebud]

Anyone who "writes" music based on algorithms, market research, and what a computer tells them will be a "hit", needs to be deposited in the bottom of the ocean.

Yes there is...

[DamageLabs]

... and it was written ages ago

http://en.wikipedia.org/wiki/The_Manual [wikipedia.org]

Re:So is that what we want, or the other way aroun

[hedwards]

Pretty much any genre of creation based upon personal taste is going to have some sort of a formula that's pure lowest common denominator. The more likely explanation is that it's what record execs think will sell and consequently it's what they push. Way too often the songs that get popular get popular because they're frequently played, not because they're good.

It used to be extremely unusual for songs on the radio to break out of a standard format and going beyond 2 minutes wasn't typically done.

Re:Why not more lawsuits

[Maximum Prophet]

The copied songs might be owned by the same uber holding company that owns the original.

Or the new artists got permission, but the original artist didn't want to be associated with that crap, but needed to get paid.

Or the "original" and the copy both are using a well known ancient riff. (See Red Hot Chili Peppers' "Dani California" and Tom Petty's "Mary Jane's Last Dance"

Robert Heinlein said, "Steal from the best, and file off the serial numbers".

Re:Correlation and causation

[timeOday]

That isn't a correlation v. causation problem. More likely it means the regularities found are necessary but not sufficient conditions. In other words they have identified *some* of the causes, but not enough to completely define it, as in write a hit automatically. But on that basis I agree it does not constitute a "formula" for making a hit.

Re:Correlation and causation

[Anonymous Coward]

Hey there, I'm one of the authors of this study... the media picked this up and is sensationalizing the study. You are absolutely correct that we need to look at the misses too (in the form of a control group) to make any statistical correlations, which is what we are currently working on. What we did was simply make some observations of descriptive metadata using visualization tools. They have blown this way out of proportion by mistaking our hypothetical opening paragraph as the results of the study.

From the Authors

[sdellis]

Hey everyone, just want to clear up a few things. First, we never claimed to have discovered a "hit formula". The media glommed onto our hypothetical opening paragraph and apparently didn't pay too much attention to our results. Please read the study observations, not the articles, for the full story. This was for a data visualization class and we thought it would be cool to mash up the Billboard data with the EchoNest data. There is no "control group" as we were only observing descriptive metadata from "hit songs". We are working on doing some statistical analysis to look at correlations. However, the data is available and we encourage others to play with it as well. Cheers, Shaun and Thomas

I-III-IV?

[jfengel]

Something from the summary really irked me: I doubt they'd find that the best songs use a I-III-IV progression. Pop songs practically all start with a I-IV-V progression. (Remember the lyrics to Hallelujah? "It goes like this, the fourth, the fifth...")

When the III is used, it's usually minor, though the minor vi is more common ("The minor fall.."). The I-vi-IV-V sequence has been the basis of rock and pop since the 50s. Learn those four chords, and you can play practically any top 40 hit. (You know the guy complaining about Pachelbel's Canon? Most of them are really just using the I-vi-IV-V, which happens to mesh nicely with Pachelbel's real progression: I-V-vi-iii-IV-I-IV-V.)

So I checked their data and discovered... nothing. Nowhere in their data do they talk about chord progressions. That's not really surprising, since figuring out the chord progressions is much trickier than figuring out the tempo. But they mention it in the summary. Why?

Because that progression is so universal, of course you'd see it in the top 40 hits. You're also going to see it in the songs you've never heard of. If they really had found that I-III-IV was a frequent hit, they'd actually have learned something.

This wasn't really intended as news. It's old stuff with new visualization applied. It's a student exercise passed off as research by people who don't actually know the state of the art, like the stories about "Students build 9,000 mpg car; why can't Detroit do that?"

It just irks me that they're talking a little music theory and betraying their lack of understanding of music theory in the process. What I've just talked about is something every, EVERY musician knows.

Yes, there is

[gpig]

http://en.wikipedia.org/wiki/The_Manual [wikipedia.org]