1 key modulation + 1 stripped chorus + 1 glory note = HIT!

I’m not a big listener of mainstream radio, but when I do tune in, there are definitely songs I gravitate towards more than others. Interestingly enough, I’ve always noticed a definite similarity between pop songs I enjoy, but I could never quite place my finger on what it was, much less identify some kind of underlying musical structure or formula that was captivating my attention. All I knew was that I liked what I liked. Maybe that made me picky. Maybe it made me a sucker for cheesy music. Maybe certain hit songs were embedded with subliminal coding that was hypnotizing me into liking songs that, in reality, I really didn’t like.

What's on his iPod?  Math, apparently.Maybe all three are right, because apparently, we are hard-wired to respond favorably to certain tunes, beats and drumrolls. And if you’re a record exec, one company wants to help you make sure that your songs reach their true hit potential. The folks behind Platinum Blue, a fairly recently developed computer application, claim to be able to break down a song into mathematical patterns, then to overlay that pattern onto a new song and predict, with 80% accuracy, whether that song will become a hit. The software uses a process called “spectral deconvolution” (kay…try saying that after a few cups of eggnog) to extract bits of information about a song–chord progressions, instruments, cadence and the like. And by analyzing almost every song to peak on the charts since 1960 in this manner, the Platinum Blue folks say that they have discovered “clusters,” or groups of a small set of characteristics shared by almost 80% of all pop hits over the past few decades.

Well, okay, that sounds simple enough, so why should someone like Clive Davis need to shell out major moola for a computer program? The answer is because identifying the “clusters” isn’t the least bit intuitive (or so Platinum Blue claims). For example, songs by Norah Jones, Vanessa Carlton, and Van Halen belong to the same cluster, while U2 and Beethoven are also grouped together, as are Brahms and Hall & Oates.

The software isn’t without its detractors, however. Well-known British songwriter Guy Chambers deems Platinum Blue “evil,” while philosopher Jaron Lanier claims that the product creates a “meaningless sensibility.” Lanier also points out that because Platinum Blue relies on the success of past hits, it could never predict the success of a truly innovative artistic breakthrough. “If our purpose is to please ourselves in the most average way possible, without caring what anything means, we might as well just kill ourselves,” Lanier laments. “We’ve lost the moral authority to want anything.” Platinum Blue, hardly oblivious to the criticism directed its way, posts this “disclaimer” on its website:

Hmm, what cluster does Gorillaz fall into?“Our technology does not influence creativity or artistic integrity rather it simply helps the industry place more intelligent investment behind songs with a better chance of market success. As a tool it can be used to find the next mainstream pop hit. It can also be used to help labels feel more confident when promoting “risky” or innovative sounding music by showing them that there is a market for it and how to reach that market. Additionally, the technology can be used by music producers to help push the envelope of creativity.”

Aside from the fact that the disclaimer could definitely benefit from the insertion of a semicolon or two (Grammar Bitch strikes again!), both truth and falsity can be found in their statement. Yes, if the logic behind Platinum Blue holds true, than Clive Davis shouldn’t have to worry about making Taylor Hicks sound more like mainstream pop and less like a soul singer–“clusters” of elements aren’t genre-specific, so as long as a Hicks song contains a few of the proper elements, it should be a hit regardless of whether it sounds more like Otis Redding or more like Rob Thomas. In that vein, Platinum Blue can be used as a tool for creativity. At the same time, though, the formula can be incredibly constricting. For example, the Gnarls Barkley song “Crazy” scored extremely high with Platinum Blue’s hit prediction software; if it hadn’t, Platinum’s engineers aren’t shy about admitting that they would have suggested how to change the song to make it fit within a pre-existing cluster.

The code, she is cracked.So, is Platinum Blue good or bad? Do we really want to enter a realm in which the songs we’re fed via ClearChannel and MTV have all been pre-selected to conform to an established pattern? Or is Platinum Blue simply telling record execs what they’d eventually find out the hard way–that people are likely to respond well to certain types of songs but not others? In other words, is Platinum Blue really making our choices for us, as Lanier seems to fear, or is the software merely a harmless predictor of what we are likely to choose anyway when all the options are set out in front of us?

The answer, I suppose, lies in whether we, as musical consumers, really do have all the options set out in front of us. Surely we don’t on the radio–just a quick spin on your car tuner reveals that you’ve got a fairly lean menu to select from, especially if your tasts fall into an off-genre such as soul, metal, bluegrass, jazz (and not Kenny G-type jazz, either). Perhaps programs like iTunes, Rhapsody, and Pandora (which uses a similar elemental breakdown to recommend music based on pre-existing tastes) are overcoming that hurdle, but at least a program like Pandora makes recommendations based upon your input, not upon some universal lowest common denominator of pleasure stimulation.

Aside from the moral quandries surrounding the program’s potential (or actual–hardly anyone admits to using it, but you just know they all are), some research actually suggests that the program itself is meaningless. One study asked teenagers to listen to music by (then-) unknown artists, then to rate the songs on a scale of 1 to 5. The teenagers were each directed to one of eight identical-appearing sites. Data at the top of each site informed the teenagers of the most popular songs within that site. Based upon Platinum Blue’s logic, that there is an inherent musical formula that speaks to all of us, one would expect the same cluster of songs to become breakout hits on each site. But au contraire; instead, the researchers found that different songs became hits on different sites–that the teenagers were rating songs based on what was already popular on that site, not based upon what appealed to them. Now, of course, teenagers are uniquely predisposed to go with the flow, and perhaps the study would have been better performed on adults, but the results certainly call Platinum Blue’s methods (or relevancy) into question, even if they are disturbing on an entirely other level.

All this debate raises some interesting questions. As to Lanier’s point about artistic breakthroughs…well, is there really such a thing as an “artistic breakthrough,” or is all past, present and future music all composed of certain inherent mathematical patterns? After all, the old adage “everything old is new again” isn’t without some merit. And as to whether Platinum Blue stifles or encourages creativity…well, in a way it’s a good thing, because Clive Davis can feel more confident marketing a blues tune that contains six or seven “clusters” of hit elements, instead of force-feeding us another Jessica Simpson. On the other hand, though, what if that blues tune, for artistic reasons, contains a soft, understated base line, and the Platinum Blue people feed it through their hitmaker and tell Clive that the base line really needs to be ramped up for optimum hit potential? Clive will win and the artist will lose, of course, and the artistic cohesion of the song will have been sacrificed in the name of hitmaking.

These questions, of course, can’t be answered in a single blog post. But perhaps it’s fitting to end with yet another one of Platinum Blue’s disclaimers: “By using our technology hit rates can be increased to over 80%; four times the accuracy rate the industry currently has, effectively providing a risk management capability that helps yield massive improvements in return on investment.” Wait…does that mean that the record companies still have to make an investment? Do they still have to dump a lot of time, money and energy into promoting a potential hit single? Why not just quietly release it to radio and let the mathematical formula do the rest? After all, if Platinum Blue is right, then we should all automatically respond to the song on some instinctual level, without all the marketing and the hoopla. Maybe the answer lies somewhere between Platinum Blue and the above-mentioned research that seemed to discount its logic. Maybe Platinum Blue identifies what a consumer is likely to respond to, but maybe like a fickle teenager, the consumer will only select it if he or she can be assured that all the cool kids have chosen the same product.

Information from this post was gleaned from the following sources:
Oliver Burkeman, How Many Hits? The Guardian, Nov. 11, 2001.
Malcom Gladwell, The Formula, The New Yorker, Oct. 16, 2006.
Platinum Blue Official Website


3 Responses to “1 key modulation + 1 stripped chorus + 1 glory note = HIT!”

  1. 1 Chris December 25, 2006 at 4:42 pm

    Bad news is that they aren’t the only company doing it. Polyphonic HMI has been around for years.

  2. 2 David December 27, 2006 at 3:38 am

    Yeah, but Polyphonic HMI was founded by the same people.

  3. 3 PayTheDevil January 6, 2007 at 4:27 am

    How did we ever purchase records? Ah, that’s right, we had “payola”

Comments are currently closed.

December 2006
« Nov   Jan »

What the kids are sayin’

"I hate them all. The judges, TPTB, the blatant manipulation, the songs, the contestants, everything. I'm a die-hard Cook fan, but for the love of god, at least try to look like you're enjoying yourself up there! Please? Syesha was awesome but she ruined it by being completely shameless and disgusting. Yes, being on American Idol is exactly like the civil rights movement, except for the part where you're fighting to make the world a better place."

"All I can say after the disgusting display tonight of favoritism towards the mediocrity that is David A. - good luck trying to market and make money off of that kid, American Idol. (Not to mention good luck dealing with his father.) All the teeny boppers may buy up his American Idol coronation single, but they will quickly forget about him before the album comes out. And I shudder to think of a David A. album - song after song of unrelenting sameness and heavy breathing. Why they are pimping him for the win is beyond me."

"This show was simply a hot buttered mess tonight. And Jason "needs to be arrested for what he did to I Shot the Sheriff. But I hope he stays. He amuses me. "

%d bloggers like this: