Will Hit Prediction Software Make It Harder For Creativity To Be Heard?

In recent years, a lot of research and development has been applied to the idea that hit songs can be scientifically identified before they become popular. Given the fact that every musician on the planet can now record and distribute their songs worldwide from the comfort of their own bedroom, the sheer volume of music being produced does need a proportionately more advanced filtering system than was necessary even ten years ago. But can a machine really do it for us?
(from Music Technology)

Songs By The Numbers

There are several companies which claim to have produced music analysis software that can indicate a song’s hit potential. Amongst these is Hit Song Science, which is currently targeting its product at music creators – independent artists and small labels.

The idea is that you first pay a fee to upload your songs. The software will analyse these songs against their database of hits and provide each one with a score between one and ten – which indicates their mathematical probabilty of being likely to succeed in the marketplace. Anything over 7.3 is deemed to be of mainstream appeal, and a score of between 6.75 and 7.29 may also be likely to succeed with a little more marketing creativity and promotional push behind it.

Where Mathematics and Music Meet

Analysing sound is something that sofware has been doing well for quite a long time, and spectrum analysis has been improving steadily alongside advances in modern technology. Although there is a substantial difference between ‘sound’ and ‘music’, there are certain aspects of songs that can be quite easily reduced to manageable data. Things which the HSS algo considers include “melody, harmony, tempo, pitch, octave, beat, rhythm, fullness of sound, noise, brilliance, and chord progression”. Song structure can also be derived by the software, and once all the information has been extracted, it can be compared to the existing database of over a million songs.

Will This System Reinforce Formulaic Regurgitation?

The idea of analysing songs for quality control certainly has some merit; in fact, producers and mastering engineers do this for a living. An automated filtering system would also be of benefit to A&R types; they may go by a different name in the future, but the record industry will continue to search for potential hit singles in which to invest their promotional budget. Anything that can reduce the amount of songs they have to listen to, and increase the chances that what they do listen to may actually become that hit, must be worth trying.

Although the industry is changing, the money will always go where the money is; distribution and production are now in the hands of the artists, so the majors will turn their attention to promotion. If they start using software to help them decide what artists they want to adopt, then the chances of anything truly original being picked up will decrease even further; however, it has always been the case that anything truly original was unlikely to have mainstream appeal anyway.

In terms of earnings, the biggest hits must have the broadest appeal – and this sets musical boundaries in itself. It is entirely possible that a software program can predict the likelihood of a particular song becoming a hit; what it will never be able to do, however, is tell you whether it is any good.


Home | Canabrism | Guides | All Music Technology Posts | XML Sitemap