Hit Song Science?

Thought you might all find this interesting.

http://www.hitsongscience.com/

I'm not exactly what one would call a Luddite, but this is rubbish. Yes, there are going to be patterns in hit songs, and yes, I believe a computer can, with some measure of accuracy, figure out certain elements and characteristics that are more pleasing to the ear or more likely to sell a lot of records. But something like this seems very evil to me. It's promoting recycling and rehashing of pre-existing patterns in music, which already exists WITHOUT a computer telling us what to recycle and rehash. Why don't we just leave the writing, performance, and production up to computers entirely and take humans out of the picture entirely? (Yes, I know that's also been done already.)

Besides, we all know that it's record company marketing muscle that makes hit records, not catchy melodies.

Think I read about this like... 6 months ago or something... Didn't know that you could actually get your song evaluated online for hitworthyness though.

Besides, we all know that it's record company marketing muscle that makes hit records, not catchy melodies.

There's a reason I don't listen to the radio ;-)

From the article:

Historically, what is pleasing to the human ear has not changed since man began writing music.

[rant]
This is just flat wrong. In a very real sense we are pleased by what we are used to. I don’t believe that our 7 note system is universal or always been the one used. I can’t be certain, but I’ve understood that other cultures have very different systems of music. Even if not, there are certainly different ways of putting music together other than our standard system of music theory. I had a professor in college who’d spent a lot of time listening to music of other cultures growing up and didn’t really care for our western system of music. He wrote some music I considered truly awful, but he seemed to like it quite a bit. I have no question he was a brilliant musician, but just in a different context from how my ear had been “trained”. We are (or most of us) are used to a pretty standard system where 7 notes are arranged into a key and utilized to make music. This is true for rock & roll as well as Motzart, but it wasn’t true for gregorian chant. We might like gregorian chant, but there are other systems (like my professor’s) that just sound “off” to use because of our musical predispositions.
[/rant]

The other thing of note is that a good melody can only take you so far. Which is why I have loads of music on my empeg that I like, but I always come back to Rush. Their music offers more under the surface than just a “pretty face”.

Historically, what is pleasing to the human ear has not changed since man began writing music.

This is just flat wrong.

Yes, it is. However if you replace "human" with "brainwashed American consumer" you have a true statement. Your "average music listener" (least common denominator) wants to hear the same little poppy catchy hooks every time.

I didn't read the article (what is this, Slashdot? ) but there are certain contexts in which

what is pleasing to the human ear has not changed since man began writing music

is true.

Our 7/12 note system, equal temperament, is a well-designed approximation of what sounds good to the human ear, but it's not the only one available. It and the others that exist approximate the ratios of frequencies that sound good. Most of them exist to provide the most common good ratios and allow the ability to play in different keys without retuning the instrument. But few, if any, of the notes in equal temperament are 100% correct ratios from the root.

Then again, there are notes that sound awful played together without context, but can sound good in the right context. The most prominent of them is the flatted fifth, which almost always sounds awful, but can sound interesting in the right context. The flatted 2nd and sharp 7th tend to sound pretty bad all the time.

So I think that most musical theories, natural or invented, try toward the same goal -- those frequency ratios that sound good, and that those ratios are universal. There may be outlying stuff about the music you don't like, but there's a lot of stuff written in equal temperament that I don't like, so I wouldn't blame the scale. But, then, there are some things that sound good only under certain circumstances, and those may well be culture-specific. (Don't you like the way I include new information in my conclusion? An F from the composition teachers amongst you.)

Good point, I hadn't thought of the statment as not being all-inclusive. I'd say your conclusion is very accurate.

And amusingly, certain ratios which really do not work on an acoustic guitar sound brilliant on an electric with heavy distortion as the extra notes generated by the distortion build a more pleasing sound.

Mmmm - love that fat distortion sound

what is pleasing to the human ear has not changed since man began writing music

is true.

No, it's not. I understand your point about the diatonic scale. So yeah, it's true that off-pitch notes that aren't in the proper scale will sound bad to our ear. Or that even certain legal intervals within that scale will sound dissonant.

But you said it yourself: even-tempered scale. Before we settled on the even-tempered scale, we were tuning the instruments to scales that weren't even-tempered. Some of the classical music we listen to today was composed for non-even-tempered scales, and would sound slightly out-of-tune to our ears if played that way today. (You can even find modern recordings of such performances just for reference.)

And then you have to think of how someone from the 14th century would react to a rap song or a distorted guitar. Or heck, even a piece of improvisational jazz.

I agree completely with your point about songs needing to be in tune and not too dissonant to be hits. But does a record company really need a computer program to tell the A&R people whether or not a song is in tune? If that's true, the company needs to fire its A&R people.

You're arguing something that I'm not. I was just arguing the one point about certain things universally sounding pleasant, by stating that certain things are inherently assonant, not whether or not that has anything to do with the song being a hit.

Ah, OK, cool.

This whole discussion is quite interesting to me. I love music theory in composition, and if I could do anything I wanted for a living it’d be to write music.

The truth of the matter is that music is an intersection between math* and art. You CAN analyze the math part and based on people’s predispositions have a good idea of how they will react to a song. In fact, a person knowing music theory intimately can read a piece of music (even if he or she doesn’t posses the ability to “hear” it in his or her head) and be impressed or bored, if only on the logical level. This part of music could very easily be analyzed in this manner and you could do a decent job of picking out good tunes. However, you might not hit the great ones. Nevertheless, the math part is essential and all musicians must adhere to it. Some musicians have been trained and have learned the rules; other’s just have good ears and have learned the rules by practice, but all good musicians have some feel for them.

The second part- what I consider the “art”- is the intuitive prerogative of the creator. While a good musician must know the mathematical rules (or at the very least follow them intuitively), there is a step beyond the math that is quite immeasurable. If a musician doesn’t know or follow the rules, however, most often the music comes out as either un-listenable or uninteresting. Often artist have to break the rules, but in these cases they generally know exactly what they are doing. A theory professor once likened this to circus performers who ride unicycles. They wobble all over the place like they have no idea how to ride, but the reality is they could never do it if they didn’t know the “correct” way to ride in the first place. Is their knowledge of how to “break” the rules just right that makes it interesting.

To carry the analogy further, an evaluation program like the one linked would promote everyone following the “rules.” That might be a good starting point, but if no one ever gets outside of the box and into the “art” area of music, all we’ll have is a bunch of people riding unicycles very properly around a tent. Not very interesting in the end, is it?

*I say “math” because underlying all music is a system of numbers, and not a simple one either.