The Science of Music - Part Two
Tuesday, January 22nd, 2008Recently in an email discussion regarding the musical career of an up and coming song writer, it was made known that he (the song writer) had uploaded his tunes to hitsongscience to have his tracks appraised for pop success…
He writes:
My scores were at an average of 6.2 with a few 7s and 7.15s. The site says that a score higher than 7.3 has a high hit potential.
Which is great, but then after a few other replies in the discussion an email from another contributor to the list, Nicholas McRoberts, bounded in with a very elegant reply…
In full:
This is an interesting and long standing debate about what makes music pleasing! And it’s not just with internet that these questions have arisen. In 1722, Rameau wrote “Traité de l’harmonie”, discussing in detail how to make a hit song for the 18th century and this was by no means the first discussion of the subject – there are traces of similar philosophical debates going back at least to Greek antiquity.However, despite what they teach us at the conservatorium – “Pop music is easy to write. It’s all formular based” – the reality is quite the opposite and the music that becomes legendary tends to change the way we hear all music by shifting subtlely AWAY from actual trends. For example:
Nowegian Wood – The Beatles
Stairway to Heaven – Led Zep
Nothing else matters – Metallica
Come away with me – Norah Jones
Yellow – Coldplay
The Bends (the whole album) – Radiohead
Smile – Lily Allen
Rehab – Amy Winehouse
Relax, take it easy – Mika
Whenever, wherever – Shakiraetc. etc.
And while there is an element of structure that is often similar (or the same) this only lets us guess what will happen when and not whether or not it will be exciting, agreeable etc.
I think of this as a recipe. If you set out to make a complicated french dish (like Fillet de lotte enrobé de noix et contisé à l’andouille de Guéméné) and you don’t heed a recipe, you may succeed or fail. If you use a recipe, you may get closer but you may also burn, overcook, undercook, oversalt, overspice, mash or generally mess up the dish. If you don’t have fresh produce to begin with it’ll never taste any good!
Regarding the hit song science site – this has been the holy grail of music researchers since IT allowed us to dabble. Algorithms that allow us to understand and predict why and how we like music. It sounds sexy but the state of the art is actually pretty rudimentary. No website offers to “read” your novel and tell you whether it will be a best seller or not because it’s not possible – simply because computers can’t “understand” what they’re reading. The same is true of music.
I (naughtily) uploaded a real hit song to HSS to see how it scored.
In 2005, the single “Vertigo” by U2 won in all three categories in which it was nominated: Best Rock Song, Best Rock Performance by a Duo or Group with Vocal, and Best Short Form Music Video. The single, which topped the charts in several countries, including the UK. It reached #31 on the US Billboard Hot 100 and topped the Modern Rock Tracks chart, and it topped the digital downloads chart in both the US and the UK, becoming U2’s best-selling single ever in the US with 2x Platinum status.
Vertigo scored only 7.12 on HSS (pdf attached). Take heart Betinho! You’re right up there!
This is so great on so many levels…
Personally I think the service is great, but as we know there is far more to a hit song than just musical ability…
Over the last three years the TOTC claim that they have been downloading various newly released music, overdubbing new parts to the track in various guises such as news instruments and additional production techniques AND then re-seeding them back to the torrent sites as the same track.