SINCE opening its doors in 1865, the Massachusetts Institute of Technology (MIT) has provided shelter to many academic bright sparks. US foreign policy firebrand and linguistics heavyweight Noam Chomsky may be the best-known MIT grandee in recent times, but he is not alone.
MIT is an egghead mothership. The institute's aim is "to advance knowledge and educate students in science, technology, and other areas of scholarship that will best serve the world in the 21st century". MIT's Media Lab has been home to much bleeding-edge technological research and innovation (despite a bit of a blip when they failed to set up a European shop in Dublin), while some 61 current or former MIT folk have won the Nobel Prize.
However, MIT also attracts rather strange coves. In the last few weeks, two computer scientists, Brian Whitman and Tristan Jehan, have announced that they have developed software which can forecast the commercial appeal of a pop song. By analysing various songs and genres, the boffins claim that their Echo Nest software will be able to predict accurately if a song will be a Top 10 shoo-in or a Number 102 also-ran.
The MIT duo believe that this kind of prediction software will revolutionise the music industry. By analysing trends and doing some clever-clogs stuff with the music, the software can produce a guide to a song's chartability.
"For record company executives, this raises the tantalising possibility of knowing in advance whether their latest pop act will hit the charts at a strong position," says Whitman, in what we suspect may well have been a Dr Evil-style voice. Whitman also claims that Echo Nest predicted the US Billboard Top 10 for a couple of weeks but, strangely enough, declined to go into any further detail about this fascinating insight.
However, the Echo Nest concept sounds extremely familiar. In fact, it's a pretty close snapshot of Hit Song Science, a hit-prediction program that had a good day out in the media sun a year or so ago.
Developed by Spanish company Polyphonic HMI, HSS is a piece of music analysis technology that also claims to be able to predict if a song will be a hit or a miss. It does this by detecting certain mathematical patterns in a song, matching these against a database containing some 3.5 million tunes and arriving at a certain score. The higher the score, the higher your chances of a hit.
Polyphonic HMI claim that HSS predicted success for Norah Jones, Maroon 5 and others. Naturally, every single one of the major labels still in business has used HSS, while there is also a version of HSS aimed at songwriters and bands who want to establish the hit rating of their tunes.
In the HSS world, every song can be tweaked into a hit - as long as you can also supply the promotion.
And there's the rub. While both hit-prediction programs can tell you that your song indeed sounds like a big ol' hit (or at least a big ol' hit from yesteryear), even the geeks in their research ivory towers know that this is not enough. A lot more money has to be spent to turn that tune into a hit.
In essence, HSS and Echo Nest are set to become marketing tools which vacillating label executives will use to avoid investing in an act. Sorry pal, we're not doing a video, the tune got a low HSS score, you need to add more backing vocals and rework that killer hook of yours.
Furthermore, because these systems rely so heavily on historical data, they're unlikely to respond favourably to new kinds of music or even a genre like hip-hop which is overwhelmingly about the words rather than the music.
If labels (and songwriters) begin to rely on HSS to deliver the goods, the result will be a whole clatter of homogenous tunes that sound just like the hits of the past. And that's a scenario which seems a whole lot like what passes for pop right now.
jimcarroll@irish-times.ie
