How Tech in Music Is Changing Our Tune
For many of us music is an escape but we also know of course that it’s a business. It always has been and as such it has a history of skulduggery and nefariousness in which the creator is often the exploited victim and the intermediary (manager, label, promoter) the exploiter. Historically, much of this has been the fact that artists were not business-like and so they kept money issues at arm’s length; transactions happened out of sight and in the shadows.
It wasn’t always that way. Back in the 50s and 60s in Jamaica before a commercial market existed for music (there were no record shops because homes owned no record decks), Sound Systems held the reins of power. They bought the only copy of a record to ensure that those punters who wanted to hear it would have to come back again to them and to them only. They also hedged their bets by first playing the unknown song to an audience — only if people danced would they then shell out some cash to the musician. For the buyer it was all upside and no risk. For the artist seller it was the opposite.
Things changed as a more traditional music industry took shape everywhere in which record companies, music magazines, radio stations then TV stations all acted as intermediaries between artist creator and the music fan consumer. In this environment, artists may not have been at the narrow behest of the Sound Systems but they had a LOT of navigation to do to get their songs to the ears of potential buyers and they had a lot of ‘help’ from intermediaries to do it. Formats may have changed over the years — from vinyl and tape to CD — but the pecking order remained much the same. Sources of income — record, ticket and merch sales — were generally understood, the economics stable and the cuts recognised.
Then along came the internet and it all changed. Today the advent of a broad range of technologies in both creation and distribution has flipped many traditions on their head leading some to see the creator today as being better protected, more in control and less likely to be a victim. Maybe.
Certainly today artists can create their music and make it available wth no reliance on DJs, radio stations or traditional retailers. They have no physical format restrictions — it’s out there in the ether. For consumers, we can hear pretty much any song from any artist at any time — in one click.
So is this current world some sort of nirvana for the artist released from servitude? Hardly. As Douglas Rushkoff says in his book ‘Team Human’, “while digital music platforms make space for many more performers to sell their music, their architecture and recommendation engines end up promoting many fewer artists than a diverse ecosystem of record stores and FM radio did. One or two superstars get all the plays and everyone else sells almost nothing.” In other words it’s easy to be there but much tougher to be seen.
Even if an artist is seen that doesn’t amount to the same thing as being paid. For a whole generation of listening consumers, the very idea of parting with money for music has become a rather quaint oddity of history. We have lost the connection between human creativity in originating a song and the digital technology used to deliver it. The thing itself just becomes noughts and ones available for free; the thing itself must have no value.
For the music business, this was not on its horizon when it accepted the inevitability of streaming (such acceptance was in stark contrast to its losing fight against the download revolution.) The reality though is that as Christian Castle — expert in music technology and policy — said in 2018 “per-stream royalties do not come close to making up for the CD and download royalties they cannibalise”
Never is that better illustrated than by a quick comparison of sales today with what has traditionally been the arbiter of musical success. In the mid 70s a Gold Disc represented sales of a million singles and would earn an artist — subject to intermediary cuts — a place at the top table of fame and fortune. Today the same volume of streams on Spotify would earn the artist the princely sum of around £3300.
Or rather it probably wouldn’t. An arcane structure of pay-outs underpins these measly streaming rates and means that any money the listener shells out does not necessarily go to the artists to which they listen. As an example, Spotify retains some 30 to 35% of all income and then splits the remainder into a pool for paying royalties. This pool is then split by the overall percentage of plays achieved by a particular song so if band X gets 3% of overall plays then they get 3% of the pool — even if you as a listener have never heard band X (or have but hate them!)
Whilst on the face of it this looks a bit mad from both the creator’s and the listener’s viewpoint, a change to reflect the actual streams of a user would not necessarily help the smaller artists. As Stuart Dredge of musically.com says “It’s no Robin Hood-style ‘rob the majors to feed the indies’ dynamic — the majors’ big back catalogues would benefit from the change” so smaller indie artists’ chances of winning out depend on “how intensely their fans stream them.”
At least artists have more direct control over getting their music out there and being paid for it. That’s something but the control isn’t quite as direct as it might appear. Spotify for example doesn’t pay the artist but rather the label, distributor, publisher or collecting society. Middlemen therefore still prevail. In 1970, Ray Davies wrote in the Kinks’ “Moneygoround” (as accurate a reflection of the biz as has likely ever been written) –
One stop on the moneygoround that might be expected to have disappeared these days is the record label. Traditionally it was the cash source that enabled creators to record and release so it understood the need to invest over time. In return it took the lion’s share of income. Nowadays, technology has made the artist so much more self-sufficient in recording and distribution that many see no point in even having a label. Yes, one may help wth these things as well as contacts/expertise, marketing and PR but they no longer invest in the long term (if at all.)
Even if labels fade away, other intermediaries appear. Today, the independent musician is offered a whole raft of services to help them get their music in front of listeners — from advertising and marketing to PR and promotion. All at a cost and none with any guarantees of music sales. In fact none with any expectation of sales.
Much of this is down to the flipping in importance of recorded music and live performance. It used to be that the former subsidised the latter but now the reverse is true. Nowadays the idea of generating money from music sales seems fanciful to many — it has become all about the stage. The pandemic has therefore been a hammer blow — with no gigs there’s no income.
Once again though there is a technology answer of sorts. Here it is livestreams. Facebook and YouTube are just the best-known tip of the iceberg of platform options. Although neither are built to generate money for artists, it’s easy enough to use PayPal or similar to set up a tip jar — sort of like a virtual busker. There are also platforms designed to take payments via paywalls. One way or another streaming is here to stay — it’s unlikely to replace live gigs but it will find a place alongside them.
During these days of lockdown, most of us will have recognised that it’s much easier to set up a livestream performance than it is to deliver an engaging one. It’s creativity rather than technology that is usually the difference. Fortunately, the role of the machine (tech) is to support the central role of humans (creativity.) The line is never crossed. Or is it?
In 2017 in ‘The Secret Science of Pop’, BBC R&D researchers and academics from Oxford and Queen Mary University London worked with Imperial College Evolutionary Biologist Armand Leroi to try to discover the secret formula of a hit record. Armed with machine learning analysis tools and 5 years of chart successes, Leroi set out to identify the shared attributes of these hits and then use them to create a ‘sure fire’ hit of his own. After enlisting the help of none other than Trevor Horn to produce, the team ‘created’ a song written, arranged and recorded to order. They then awaited inevitable chart success. It didn’t come. Despite representing a scenario to get the music biz suits purring — how to get rid of the pesky and petulant ‘talent’ — it failed because they couldn’t.
The attempt to master the human creativity behind song composition lies somewhere between the narrow form of AI mostly being employed now (creating a distinct, and often narrowly defined, action) and Artificial General Intelligence (AGI) which would deliver what Nick Heath of ZD Net describes as “understanding the world as well as any human, and with the same capacity to learn how to carry out a huge range of tasks.” AGI is a distant prospect; recreating the spark of creativity is not much less so.
Technologists don’t stop trying though. Earlier this year a Dutch broadcaster ran a competition for the creation of a hit pop song by AI. The winner was an Australian team called Uncanny Valley comprising both musicians and AI engineers; it used a dataset of Eurovision songs and — would you believe — koala and kookaburra cries. The project illustrated the importance of humans working with AI; the models didn’t make all the decisions but rather the humans produced iterations and combinations of models to achieve what they wanted. Sandra Uitdengogerd of RMIT University in Melbourne and one of the Uncanny Valley team told Will Douglas Heaven of MIT Tech Review that “honestly, I think that humans could have done it equally well.” Maybe the learning was that it could be done rather than why it would be. In the end too, deciding on a winner was the job of a good old human judging panel.
The Uncanny Valley experience suggests that any shift towards technology becoming the creative process rather than playing a role in supporting it is a way off. Some people however are in a hurry.
Jukebox is an AI based system that generates music in the style of specific artists by analysing their own work to identify attributes that a listener would associate with them. One of its projects was a Christmas Song in the style of Frank Sinatra. In principle of course — and as an exercise in innovation — this is all well and good. In practice though it creates obvious potential problems. In May, Dave Gershgorn reported on Medium that a developer had created a track on Soundcloud that so convincingly mimicked Jay-Z that the rapper’s agency took action to have it removed. Neither of these cases involved money being generated but it is hard to believe that the technology couldn’t and wouldn’t be used to create dubious material that would impact an artist’s reputation and their income; from Fake News to Fake Tunes. In that case it wouldn’t be a simple case of applying copyright laws but rather laws of impersonation which are very different indeed.
Some might say that the threat of technology has hung over musicians for decades — drum machines promised to make drummers redundant and sampling to do the same to pretty much everyone — and yet they are still here because music sells and music generally can’t happen without them. The more pertinent point however is that their life gets tougher and the range of threats wider; technology is the main driver behind that.
The question that we listeners might ask ourselves is do we care?
