ratpie

Category: technology

  • Who Owns The Codes? The Sound and Ownership of Music in the AI Age

    Preamble

    Early in 2024 I was asked to contribute a chapter to the IAEL’s book on AI in music. It was an honour to be asked, and a significant effort to deliver something that I felt would not waste the readers’ time by adding to the torrent of ill-informed texts about this curious subject. Curious, because at no point have I heard anyone complain about the paucity of supply in music; quite the opposite. As a supremely human endeavour, music sits uneasily within the framework of capital and production for profit. AI promises to remove the human, and the endeavour, so that the gains can all accrue to capital. Why we should give this idea the time of day I find baffling, and the knowledge that our elected representatives here in the UK are seriously contemplating lifting one of the great achievements of the Enlightenment – copyright in creative work – to grease the extruders is profoundly depressing.

    I set off at a clip, and by the third delayed deadline and some excellent editing had delivered a shortened version of this text. Some of the shortening took out references to interesting music produced along the timeline of music’s interaction with systems and ultimately with computing, of which generative AI is arguably a continuation. The IAEL is now no doubt working on another publication of equal importance to the music industry, and I can with confidence recommend whatever they produce as I know it is the product of a very thorough, well informed, and humanistic process. So with broader interest in music and AI probably peaking here in the UK I am taking the opportunity to evade my editors and publish my first version.

    Introduction

    AI and music is a big subject. I propose here to discuss some of the concepts and foundational methods behind using AI to generate music. I will ask how we can know about the sources that have been reworked into new music by generative AI. Finally I will consider some of the impacts we will very likely see on culture and the music industry, including how we might need to rethink some of our ideas about creativity, ownership, and attribution.

    AI is already used widely in music for other things, some of which are certainly interesting and useful. Recently AI extracted a John Lennon vocal track from an old mono cassette demo, and gave us an unexpected and unprecedented new Beatles single. And AI is smoothing out human fallibilities in pitching and phrasing on many new recordings, and helping deliver industry standard mixing and mastering. Assistive technologies raise very different questions however; we see far greater cultural, legal, and economic challenges when AI itself makes the music.

    First the usual warning! IANAL – I am not a lawyer, but I do have over 30 years broad and active experience in the digital music industry. And for another acronym – TLDR – here’s where I think we will land. I certainly don’t foresee wholesale destruction of either musicianship or the economics of making records; the value of these is not just an affective waveform, but rather a complex bundle of culture and identity alongside the traditional music making and recording skills. And that suggests that we will find ways to recognise and reward human creativity when it uses and is incorporated in new computer generated music.

    Sonification of Code

    To understand how generative AI can make music we need to start with an idea of how music can be understood in terms of codes and rules. People have been using codes to generate music probably for millennia. Anyone who been through a western music education will have learnt many sets of rules, such as those which underlie fugue and counterpoint. Some prominent composers played with their understanding of rules and chance. One thinks of Mozart’s ‘Musikalisches Würfelspiel’ (Musical Dice Game), where a roll of the dice selected precomposed bars. Outside of the western music, traditional forms such as raga and gamelan are essentially heuristics performed by musicians. 

    Instruments such as the barrel organ and player piano, which performed mechanical representations of music, were transformative, helping sophisticated music reach further than could have happened solely with trained musicians. It was not just that the machines could be operated by an untrained musician; they could also play music that was beyond human capability.  Rendering the mysteries of music as a sequence of pins or holes in a roll of card must have suggested the possibilities inherent in applying maths and geometry to expand what could be created. This characteristic was exploited to the full by Conlon Nancarrow in his Studies for Player Piano. For a fun example, track down Study for Player Piano No. 3a, which is, to my ears, remarkably similar to early generative, or AI music, featuring manic boogie woogie motifs.

    It was the introduction of the computer that blew the lid off what had until the mid 20th century seemed the rather constrained field of mechanised music. The digitisation of music in western culture, as notation, chiptunes, and later as MIDI data, made the rules inherent in music much more observable. Advances in digital signal processing during the 1990s showed that audio files containing music could be analysed and musical information extracted, starting out with pitch, onset, and duration of sounds. Early computers could be applied to algorithmic composition, picking up on the idea of music as the product of heuristic or stochastic processes.

    So this approach treats music as essentially a sonification of codes, and there are some quite wild examples. The fruitful interchange between modern classical practice and the computer music labs that started to be built in the second half of the 20th century looks almost like a ‘scene’. Alvin Lucier for instance mapped brain impulses to electronic instruments, producing a percussive soundtrack to perception, thought, and feeling. Another academic composer, Xenakis, was among the most strongly identified with computer software as a compositional tool. It’s an important distinction that Xenakis was not looking for creativity in the algorithms that assisted in his composition. Instead he considered the computer as a way to express through the medium of orchestral music the mathematical concepts he was familiar with. Later he developed software that could turn his drawings into sound, enabling him to use in composing the skills he had developed as an architect.

    Beyond the Code: How Generative AI Works For Music

    Contemporary applications of AI in music are much more sophisticated in how closely they emulate what we think of as ‘human’ made music. And they generate music that aces the Turing test. Given how good humans are at making and following rules, that’s perhaps not so surprising. But when we can’t tell whether the pilot’s hands have ever touched the rudder, that seems to me a step change.

    Many aspects of our cultural and economic life have been disrupted by the rapid global development of digital and communication technologies over the last 35 years, and the academic field known as MIR – Music Information Retrieval – is no exception. The ability to make huge collections of recorded music available digitally, to apply powerful computing resources to analysis, and quickly share the resulting codified data sets, catalysed the explosion of ‘generative’ AI that we are now in the middle of.

    To understand better the impact AI will have on music, the music industry, and music culture, it helps to understand something of how those black box algorithms can churn out tunes.  With sincere apologies to the scientists and researchers whose work actually powers these extraordinary advances, here’s a very brief and sketchy overview of how generative AI works for music.

    The sounds and patterns we perceive in music can be extracted from a corpus of compositions and recordings, and expressed as a set of probabilities. Some of the basics are easy to codify – pitch, basic timing blocks such as beats and bars. Different instruments can be separated out, and their roles understood, which enables the more properly ‘musical’ features such as melody, rhythm, and harmony to be delineated. Higher level structures such as verse, chorus, bridge, etc., emerge out of further analysis. Commonalities between different works and recordings can be clustered, creating sets which are analogous to genres, styles, and modes. Hierarchical sets of features further define the differing eras and traditions, enabling automated differentiation between similar sets of instruments used in different ways.

    These models can be thought of as the vocabulary and grammar of the language of music, but of course without the meaning that natural human languages carry. In data terms the music has been separated out into tokens, the sequential relationships between the tokens have been inferred and stored, and the inference used to improve subsequent performance of the algorithms. The software machines that do this work are known as ‘transformers’. Recent advances have been driven by the discovery that adding more parameters and more transformers hugely improves the models that can then be used to generate output. In some ways this is similar to the way computers understand images, as a grid of pixels; the image gains fidelity the more gradations in colour and brightness there are, and the denser the pixels are packed. At a threshold human perception loses the ability to see the degradation, and just sees the image. Similarly with generative AI, adding vast complexity nudges the performance up to the point where we humans simply can’t differentiate between ourselves and robots.

    In AI music analysis and generation, as in other applications, this phase is known as ‘training’. Any corpus of music, which in the AI field is known as a ‘training set’, can thus be rendered as a set of descriptors, along with the probability that each musical idea or motif will happen at a given point in a piece. The result in AI terminology is called a ‘foundation model’. Prompt the foundation model with some terms that are strongly associated with a particular form or style of music and out comes what you might call a statistically probable fugue, or rockabilly song, or eurodisco track. Or a guitar part, or drum track, or string section to supply what would otherwise be provided by a musician.

    All very interesting, but if audiences don’t rate the results the whole enterprise remains, literally, academic. Early computer music did indeed sound pretty inhuman. The micro-timings and weights in human performance, and the infinite timbral spectrum of the voice and acoustic instruments were not easy to emulate. Perhaps encouraged by the market’s reaction to his Ambient series of recordings, in the mid 1990s Brian Eno produced and released his ‘Generative Music 1’, using software created by Tim Cole & Pete Cole at their company SSEYO, and which they called ‘Koan Music Engine’. Their software emulated stochastic processes, such as the path a ball bearing takes down a chute. It lent itself naturally to contemplative, slowly developing soundscapes. Eno himself made the connection with the minimalist composers of the 1960s.

    The huge scale of today’s foundation models, well beyond what even the most powerful computers could process in the 1990s, has opened up many more styles and genres of music to convincing generative production.

    Owning the Codes

    This asks some very fundamental questions about whether the creators of the music in the training sets have an attribution or ownership interest in the output of generative AI. It is as well to know what we could be able to do about that before considering what we should do. Tracing the output back to the inputs in the sets is technically difficult, but by no means impossible. Any piece of music in a training corpus can be uniquely identified using methods we have developed and applied in the digital music industry. A combination of technological tools such as file hashes and audio fingerprints can be used to build strong associations between the music and copyright holders. The nature of the analysis required to produce a foundation model maps very well onto a requirement to store and manage the provenance of each inference stored as a token in the model. And when the model is used to generate a new piece of music, the tokens invoked, and therefore the provenance of each element, can be logged.

    This describes a situation where we could have visibility and control over each step in the process, from raw material to finished product. There are two aspects to this which would present a challenge in the real world. Firstly it is data heavy, and very dependent on trust in the parties and the systems involved. When very few entities have the resources to manage the processes involved in training foundation models this might not be insurmountably costly. As this capability spreads however it will quickly become impossible to identify who has done what with which music. And that’s the second challenge; we wouldn’t necessarily think it desirable to embed the kind of external accountability and surveillance into private computing systems that would be required to make an auditable supply chain.

    One response to this could be to decide just to select reputable sources and rely on what they tell us. That is the approach Google is taking with its SynthID product. When its own music generator, Lyria, has output a new piece of music, an identifier is also generated and embedded in the waveform. When the audio is rendered for playback this identifier can be extracted. It’s robust enough to withstand compression and transcoding. But it clearly has no meaning outside of a Google world, so doesn’t help us much when we leave the ecosystem. Auditing Google seems tough too, one of those ‘quis custodiet’ type of questions.

    Where we have no visibility or trust, either because we can’t or don’t want to embed traceability in our music production and distribution systems, there is another approach which might yield at least a reasonable degree of confidence that a source was in a training set. Just as the characteristics of sets of music are codified and stored by the foundation models in the generative AI process, the same process can be applied to the results. If the resulting models display a high degree of similarity it’s possible that the generated music was based on a highly correlated training set. That’s not quite the same as fingerprints on the murder weapon, but close.

    Given the complexity and disproportionate cost of tracking, and some of the uncertainties, it doesn’t seem viable to me to use the same private licensing and accounting for AI generated music that we have developed for sampling, for example. And with complete supply chain tracking unlikely to be achieved, we would naturally look for some kind of collectivisation of ownership, with distribution rules to deliver fairness rather than absolute accuracy. This comes with note of caution, as this is a rapidly developing field of study and much remains uncertain or unknown. But if we can train the transformers to deliver an opinion much like a trained musicologist might, the same technology that generated the music could be used to apportion revenue to the creators of the training set. We can see from this how we can maintain a set of sufficiently identified works and recordings in a repertoire pool, and a set of usage data and statistical processes with which to share out revenue pools derived from AI generated works and recordings.

    Speed and Unflagging Energy

    If AI music makes some people nervous, it’s easy to see why. We’re very attached to our Romantic notions of creativity, and the myth of the inspired and lonely genius. Indeed, artistic copyright has its foundation in individualism, and the inalienable identification of the work with its author and of the performance with the artist. Mozart’s appreciation of his own genius was ahead of its time, as his letters painfully show. His miserable daily life was that of a court servant who refused to please his master.

    It seems possible that our Romantic notions of creativity follow, rather than lead our investment of copyright in the results of creative efforts. The myth of the lonely genius seems to emerge alongside the introduction of copyright, perhaps to manage the risk of being accused of plagiarism. In an inversion of the political economy that made a pauper of Mozart, the most successful copyright creators have ended up with the wealth and status of their erstwhile aristocratic patrons, while somehow preserving their bohemian freedoms. Exclusive ownership in music turned out to be very much worth having.

    Even the loneliest and most inspired Romantic composer leaned on formal systems, such as scales, harmony, regular rhythms and time structures, and larger musical forms such as the symphony. Most have studied music theory, covering systems and rules as well as precedent. Innovators such as Satie and Debussy were contemporaries at the Paris Conservatoire, and today there is a Conservatoire named after Debussy, which tells you which of the two was more of an iconoclast!

    Contrast the most formal approach to composing and performing music with the speed and unflagging energy of AI, processing tune after tune in seconds. Then add the notion that the AI could at some point be able to call on an almost perfect and complete set of music theory, and millions of previous examples of human music. It’s a classic B movie plot – the machines are taking over, and all we can do is build more electricity generating plants and data centres to feed their insatiable maws. And, thanks to the combination of cloud computing and the internet, the music machines are universally accessible, at low cost, and already connected to global music distribution platforms.

    A Synthetic Economy of Music?

    It’s here, at the intersection of culture and the economy of music, that some anticipate an AI-induced tectonic shift. There’s a dystopian way of looking at this; synthetic artists flooding synthetic markets with synthetic music. To me this seems to misunderstand the nature and meaning of value in music, severing it from its cultural and human context. That today’s markets might not be great at rewarding the value we find in music is not an argument for abandoning them!

    Digital music platforms have a history of intervening in the supply side of the market; indeed the digital market was kickstarted by two big interventions, unbundling the album, and enabling mass ripping of CDs to fill iPods. Platforms have encouraged mass piracy, and more recently incubated and supported DIY music upload businesses. AI is the latest intervention and is already fueling a massive increase in supply. Just one company, Boomy, claims to have generated nearly 20m songs.

    Success in the industry equates, broadly, to high popularity over time. Music is a cultural good with very low barriers to consumption, and a typical consumer’s usage remains strongly influenced by her awareness that others like her are also enjoying the same music. Such a competitive dynamic fuels a very concentrated market; a few tracks earn almost all the revenue, while most get little or none. Even the crudest division of the market into winners and losers asks the pertinent question: which cohort is most vulnerable to competition from AI-generated music?

    I suggested above that current AI-generated music aces the Turing test; this assertion needs heavy qualification. There is a big difference between ambient trance music, and a complex piece involving traditional acoustic instruments, singers, and sophisticated production techniques. AI is much better at trance. A big movie soundtrack remains a big beast of a creative challenge (although one can prompt an AI music generator with ‘epic cinematic soundtrack’ – the result sounds like the band Dragonforce but without the humour). So the strength of AI music generators is their effectively infinite capacity to increase the supply of generic music.

    But our collective preference is demonstrably for hits and innovation over time, rather than just more generic music with less of the cultural and human value in the bundle. Even the recommendation and curation interventions struggle to compete with audience preference. The prized Spotify editorial playlist placements have a follower to playcount ratio of between 1% and 5%. Unless we get to the point where we prefer AI generated music, its impact on success for artists will be minimal.

    Assimilation

    Like so often in culture and markets, the infiltration of generative AI into music turns out to be a question of values as much as of technology. A dystopian outlook for a human economy of music making seems to me to be founded on a view that the value in music is solely in any anonymous affect it has when heard. This is ahistorical. Music might not have the cosmological significance it once had, but it remains an experience and activity loaded with personal and societal significance. We might find comfort in familiar or unchallenging music, but the norms expressed in our copyright laws incentivise innovation and originality, and markets agree. It’s the Queen catalogue that is the Koh-i-Noor diamond for the music rights hoarders, not the tens of thousands of contemporaneous rock plodders.

    Here then is a positive view of a future in which generative AI makes music. We will surely find ways to harness AI in generating therapeutic or performance enhancing music for health and sport. Generative AI can iterate and shorten the feedback loop in ways that humans just can’t. The value of this music will be immense, and there is already a lot of literature supporting its effectiveness, even before we could adapt and optimise the music itself. 

    Alongside this the processes involved in generative AI are producing new insights into what music is, leading to an expanded commons of ideas and understanding available to the foundation models as well as to human artists and composers. Our copyright framework recognises the expression of an idea as worthy of protection, not the idea itself. This common stock from which musicians draw is an important cultural seedbank. We use tunings and structures in contemporary music that were fresh in the Baroque period; generative AI will perhaps help keep this stock fresh so that future musicians have their own go at making the perfect break up song or pastoral tone poem.

    Our veneration for artists and creators is unlikely to be eroded by the presence of artificially produced music in our public spaces and on music platforms. We will continue to want to see the artist’s presence in and behind the music, and to do so we will surely need to strengthen the way we manage identity and authenticity in creation and performance. So with an awareness that I am writing for an audience of entertainment lawyers, here is my best guess for what the near future holds for music in the age of generative AI. We will add an AI value for the use of music and recordings to our collecting society mandates, and find inexpensive ways to distribute revenue, probably using similar technology to that which made the AI music in the first place. And we will support the deep cultural and social value our artists represent with stronger attribution rights, and technology to identify and authenticate their work.

  • AI Will Help and Hurt Music. Here’s How.

    Computers have been helping us make and store music for many years. Music AI is not so new either; already by the early 1990s there was widely available software runing on consumer grade desktop computers which could generate and regenerate music-like patterns of sound. The last 30 years have seen incremental improvements in the way computers analyse, categorise, and parameterise music.

    This body of knowledge is increasingly used to make new music, but it’s hard to find a basis on which to judge the output of these robot composers. Should we fully decontextualise the results, risking the rather nihilistic conclusion that everything produced is a perfect version of itself? Or should we continue to base our judgements on our own human history, achieving perfection only when we can’t tell the fuguebot from the Bach? Or perhaps we should be simple hedonists, and ask the artificial intelligence to iterate for our greater pleasure.

    The digital music market gives us another new way to evaluate the robots – commercial productivity. Is their output merchantable alongside music made by humans, and at a competitive cost to produce and supply? Early indications suggest that it is, and that in many ways robots are a much better fit than musicians with the market conditions created by music subscription services.

    So what is it that makes a robot the perfect modern artist? Consider the contours of the digital market: each month is a contest for the plays that are claims on a finite pot of money, paid in by subscribers who can only play one track at a time. Increasing the play count across the whole service does not increase the money available, so one track’s success comes at the expense of all the rest. If instead of being an artist a supplier could be the owner of a fleet of them, each turning out the optimal amount of the right kind of music at the right time, sheer bulk in each monthly contest would tilt the odds slightly, even without any algorithmically refined fitness.

    An entirely digital supply chain means that new tracks can go from AI algo to audience without being touched by human hands, and on an unprecedented scale. Here is Boomy, a fairly recent start-up taking advantage of these capabilities:

    How does Boomy work?
    Boomy uses music automation technology powered by artificial intelligence, which you can use to create and save original songs in seconds for free. You can also create Releases and distribute them to all major streaming services and digital music retailers worldwide, and earn a share of royalties when your songs are played on networks like Spotify, Apple Music, TikTok, and YouTube.

    https://boomy.com/about

    Choose a style and a few options, click a button, and it will make a track. Type in a title and ‘artist’ name, and it is ready to be delivered. Boomy retains the ownership ‘for convenience’, but pays whatever pennies the music generates on Spotify and many other music services. Somewhat disingenuously Boomy claims, “Boomy users have created 9,007,441 songs, around 9.1% of the world’s recorded music”, all without a single shake of a tambourine. They are conveniently short too, ideal for increasing the collective playcount when playlisted.

    It’s surely not controversial to suggest that capital likes indefatigable and fatigue-free robots. So to be able to piggyback on 300 years of hard won creators’ copyright law and soak up some music subscription money seems like an offshorer’s dream scenario. They should enjoy it while it lasts, because there are many reasons why that won’t be very long.

    Music seems to be fundamental to who and what we are as humans. Musical instruments are among the earliest artefacts that we recognise as setting us apart from our fellow primates. We know that flutes have a continuous history of more than 40,000 years; it would not be surprising that tuned percussion has been even longer with us. And of course our voices and bodies precede anything we made in prehistory. Mechanisation came to music surprisingly early too. There are references to instruments resembling the hurdy-gurdy from very early mediaeval times.

    With its brief periods of fashionability, the hurdy gurdy is an illuminating glass through which to see AI in music. For a while it was an innovation, playing a role in churches before the mighty organs were developed. Then a smaller version brought more music to more people as a reliable way to get a dance going at fairs and gatherings. Rediscovered by the fops of the French court it made an arch appearance in high society for a while, before resuming its place with the peasants and itinerant beggars who were music’s first democratic professionals. It is now perpetuated as a heritage instrument, maintained as a reflection of our inability to give up our cultural past.

    Despite its presence at so many events over the ages there is not one single piece of hurdy gurdy music among the highest achievements of our musical traditions. Contrast the piano, also an innovation in the mechanics of musical performance, but not designed for efficient production or portability. Instead the hammered strings brought the potential of infinite subtlety to the rather limited prior pluckers of the keyboard family. Initially a sceptic, the great Bach was later so impressed he became a sales agent for the second wave of pianofortes.

    So there are two contrasting innovations, one to increase the productivity of less skilled music workers, and one to extend the expressiveness of the best musicians of the age. And perhaps that is how we can understand today’s emergence of AI in mass production of music, while keeping open our hope for new possibilities in human creativity.

    The value of music is the humanity we invest in it, and find in it. As an unashamed elitist I see no value in the millions of hours of audio generated from the parameterisation of musical concepts; and no inherent additional value in the particular instruments and tools that musicians deploy. But I anticipate with joy the new worlds that will be created when the best musicians harness the generative power of AI.

  • Net Neutrality Won’t Help Music. Here’s Why…

    With the best intentions, some in the music industry are adding their voices to protest the US FCC’s rollback of net neutrality regulations. Keeping them won’t help the music industry; pretending it will means the real threats to open and fair digital markets will remain unaddressed.

    To recap, net neutrality, as it is framed in the regulations, stops your consumer broadband provider from making either prioritisation or access to your devices conditional on fees or other non-charging policies. It was motivated partly by some clumsy attempts by telecommunications providers to protect their voice traffic by blocking VoIP and other peer to peer protocols (without net neutrality rules Skype would have been a far less attractive acquisition for Microsoft). Small content providers saw similar threats from huge media businesses, and feared being priced out of their audiences’ home Internet connections.

    All this seems reasonable, but the concept of net neutrality – invented by a lawyer, not a technologist – is based on a very much oversimplified understanding of how the Internet works. It’s easy to assume that your consumer broadband provider connects to the Internet with everything else somehow magically taken care of. Indeed, many network diagrams just show the Internet as a cloud. Global connectivity is complex however and has changed a lot over the last 25 years.

    So how has the Internet changed, and why does it matter? The discrimination that many fear is already happening in many ways and many places, to the advantage or disadvantage of many parties. Please pardon a somewhat detailed history lesson, but it is important to understand just why net neutrality is not capable of addressing the concerns it pretends to.

    In the early days the Internet was made up of a fairly neat hierarchy of local, national, and global networks, referred to as tiers 3, 2 and 1. Lower tiers bought traffic capacity – known as IP transit – from higher tiers, and the tier 1 networks mostly just swapped data with each other at no charge. This business model drove investment in fibre and datacentres with relative agnosticism about what content and services were being carried.

    Telehouse in London's Docklands
    Telehouse, London

    One of the oddities of this arrangement however was that data sent between two near neighbours sometimes took a huge round trip to get to its destination. You don’t need to be a network engineer to see that exchanging traffic between UK networks in the UK should be quicker and cheaper than sending it across the Atlantic and back; in fact that was exactly the problem that led a group of UK ISPs to establish LINX, the London Internet Exchange, in 1994. Connection facilities such as LINX are known as Internet exchange points (IXPs). Initially membership of exchanges was limited to networks with roughly similar in/out traffic ratios, and costs were mutualised rather than individually billed between network owners.  The exchange of traffic is known as ‘peering’ – giving a sense of its cooperative heritage – and can be done privately between two or more networks, or publicly by plugging into shared data switches. The global network of IXP centres continues to develop.

    Internet architecture is designed to support peer to peer traffic, and each device can theoretically send to and receive from any other connected device. But as the net developed its commercial model, the traffic flows dramatically changed shape. More and more data was coming from specialised networks, such as CDNs (content distribution networks) and then from very large content distributors. The IXPs resisted at first, but then allowed these very asymmetric networks to join. In their quest to get ever closer to the consumer CDNs and content distributors quickly moved from public peering to private, investing in large capacity gateways to the bigger consumer ISPs. Then, as the consumer ISP market consolidated, the big content distributors started to install their equipment actually inside the ISP network, with private fibre connections to their own datacentres, a method known as colocation.

    This change was being discussed, as a ‘massive disruption’, in the network operator community by 2007. Here’s a paper on it by one of the real experts, Willliam B. Norton, and here he is talking at a conference:

    My own perhaps romantic take on this from 2012 is here: How Art Changes the Architecture of the Internet

    Today these three modes of connectivity co-exist:

    • Public peering – relatively low cost, mutualised and operated by membership of exchanges
    • Private peering – organised ad hoc between network operators, and often used for specialised needs in business to business traffic exchange
    • Colocation – extremely expensive and only available to a small club of the biggest content distributors

    To get an idea of the thresholds which drive traffic from one mode to another here’s a policy published by the UK’s national broadcaster, the BBC:

    Any peer exchanging more than 1Gbps of traffic in either direction at any public peering location will be requested to migrate to private peering, at our discretion.

    It should not be surprising that Google, Amazon, and Netflix are some of the biggest users of colocation, given the amount of video they ship to consumers. Here’s what Google says about their service:

    With our edge nodes, network operators and internet service providers deploy Google-supplied servers inside their network.

    Google has even installed its servers inside the Cuba national telecoms company, with now former chairman Eric Schmidt following close behind the Obama administration’s political normalisation efforts:

    The Empresa de Telecomunicaciones de Cuba SA ( ETECSA ) and the US company Google have concluded the talks with the aim of signing an agreement Google Global Cache (GGC).

    And here’s Netflix explaining just how powerful the economic forces are behind colocation:

    Globally, close to 90% of our traffic is delivered via direct connections between Open Connect and the residential Internet Service Providers (ISPs) our members use to access the internet. Most of these connections are localized to the regional point of interconnection that’s geographically closest to the member who’s watching. Because connections to the Netflix Open Connect network are always free and our traffic delivery is highly localized, thousands of ISPs around the world enthusiastically participate.

    We also give qualifying ISPs the same Open Connect Appliances (OCAs) that we use in our internet interconnection locations. After these appliances are installed in an ISP’s data center, almost all Netflix content is served from the local OCAs rather than “upstream” from the internet.

    So to understand today’s Internet it is important to have these three different models for how data gets into the consumer’s home; and to understand how the economics and incentives that drive their respective capacities and capabilities differ from each other; and most important of all, the balance between them, how they compete with each other, and the investment model that supports each of them. Because as traffic moves from one model to another – from IP transit to public, then private peering, then to colocation, the investment case moves too.

    The simple fact is that a colocated content distributor needs less private peering, and little or no public peering in order to reach its consumers. Traffic might still be growing across the internet, but with public peering growing slower there’s less money to invest in IXPs, and more money for what are effectively private overlays occupying the same conceptual space as the public Internet. It’s obvious which companies are involved; Netflix for instance, and Google, for its YouTube traffic. Netflix is not investing in public peering capacity; to do so would only benefit its competitors.

    But here’s where it starts to get a bit more interesting (sorry it took so long). Cloud providers too are colocating alongside their peering arrangements. This club includes Amazon and Microsoft, and Alibaba now entering the global market. Music companies are big users of the cloud, and if they are buying AWS, Azure, or Google Cloud each of them, no matter how concerned they might be about net neutrality, is benefiting from the advantages of getting out of peering and into colocation.

    Now consider this from the point of view of a consumer broadband provider. Your colocation partners are already on your network, and on very high capacity connections; they need it for their video streams. Your private peering partners are also connected to an extension of your network in a datacentre somewhere, and have as much capacity as they need, but are further away in network terms so are less reliable and efficient. Your public peering is now very much the poor relation; you have no control over the quality of the route that the publicly peered content takes before it arrives at your network, and while you need to maintain the connection in order to be able to deliver the whole of the rest of the Internet to your customers, it’s impossible to anticipate which connection a large file or video stream might come from, nor help it arrive smoothly and in one piece.

    Raphus cucullatus, Roelandt Savery (1626)
    Raphus cucullatus, Roelandt Savery (1626)

    So, in brief, many music companies are now providing additional revenue to the cloud providers, who are abandoning as fast as they can the mechanisms by which the Internet remains open and fair. And that does not just include record labels; many consumer music services have moved their systems to the big cloud providers.

    Here’s an ironic thought – an artist, a record label, and a distributor might already each be paying Amazon to store and process a track, in order to send that music to be sold and streamed by… Amazon!! There are more than twenty different ways that Amazon can extract revenue from music (sometimes in scalable and repeatable ways), and all but a few result in no revenue for the creators. It seems that the music industry collectively is doing as much damage as it can, as fast as it can, to the benefits that net neutrality is supposed to preserve.

    Business is littered with situations where the collective interests of a class of companies are in conflict with the individual interests of most of them. Usually the result is the development of monopolistic or monopsonistic markets; no industry benefits from the dominance of Facebook and Google in digital advertising for instance, but many businesses benefit hugely from the reach and capabilities of the data and marketing platforms. And the answer to monopoly is usually regulation, but for better regulation to work it is essential that the intended beneficiaries have access to the skills and knowledge to move profitably into any fresh opportunities that the new rules might bring.

    Can music make that move? As an industry we’re suffering from a brain drain out of the supply side and into the demand side, and there’s a lot of evidence that our trade associations and collective bodies are under-resourced to help such a change. But the game is far from over. In net neutrality, and in technology, the best approach seems to me to be the same as in so many other areas where we are looking at erosion of our freedom and agency. And that approach is a fierce and supremely well informed dedication to independence.

  • Three Conditions for Independent Music to Survive

    A little over two years ago I set out three reasons why I thought indie music was poised for a period of cultural and commercial success. Those reasons seem still valid now, but a series of events in the indie music world has shown that I overlooked some important factors.

    Floral Matryoshka set

    My biggest mistake was to assume that major labels would not want to buy margin-shrinking and inherently weak independent distribution companies. Since 2015, several ‘insider’ distributors (the kind whose founders and owners sit on trade association boards) have been sold to major labels, or major owned distributors. It’s impossible to understand these acquisitions any way other than by the damage they do to market access for successful small record labels; these deals must be strategic.

    The second mistake I made was in over-optimism that the ‘digital dividend’ would materialise for small businesses. Online platforms have revolutionised how teams can operate around the world, but indie labels seem to have missed out somehow. Perhaps their needs are so specialised that more generic tools – productivity suites, online accounting – just don’t have the same impact as in other industries.

    Time then for a rethink. And it is one from the same fundamental optimism of a few years ago, but with some caution and new caveats. So instead of offering reasons why indie music will thrive, here are three conditions for survival as we head towards the end of Q1 of the 21st century:

    Efficient Collective Rights Management

    The indie music world seems to me to have fully accepted the need for intermediaries, aggregation, and where it’s necessary, collectivisation. What’s more, Merlin has set standards in rights management that have lifted the quality of licensing across the whole market.

    A Thriving Marketplace for Services

    Small businesses buy in more and use in-house teams less. Our community has not generated an effective market in the services we need, from sales and royalty systems to production, design, marketing, business information, and intelligence. Instead we have churn in providers, and risk for investors that customers will disappear at a moment’s notice.

    An Indie Scale Connected Digital Platform

    New music markets need new digital infrastructure to help indie labels manage releases and coordinate their relationships, not just with the music services they supply, but also the other services that they need to buy in. Platforms can transform industries, becoming almost like a computer operating system in the way they connect resources and processes. A platform is what connects the rights, the assets, the businesses and the people together to deliver all the other benefits of the digital world to independent music.

    …And What Won’t Work…

    And it is worth setting out what clearly won’t work, as aggressive major labels and VC funded businesses continue to take bites out of indie music infrastructure and catalogue.

    The first and foremost of our impediments is the attitude that scale doesn’t matter in technology – that somehow a two person startup with a rented database and some code is an adequate answer to the massive platforms we now interact with daily. Our dependence on technology I have heard described, depressingly, as ‘the tail wagging the dog’. This particular tail has been wagging the dog for at least the last 15 years.

    And a new respect for the strategic importance of technology should be accompanied by a scale up mentality in how we nurture service providers, rather than a start-up culture with the inevitable failure rates and, even worse, success through acquisition. All the best new talent is being snapped up by the big music services, leaving us data and capability poor. Better prospects over the longer term for the smartest people will mean that the other great benefits of working with music – the culture, and the community – really do help us stop the brain drain out of the supply side of our industry.

    I remain optimistic that another two years could see us starting to gain traction with a global community of independent music companies and service providers, leading the conversation with the big tech platforms, and setting the terms for our businesses and for music. We just need to straighten up and fly right.

  • My Platform, Right or Wrong!

    It was an early hero of the United States navy, Stephen Decatur, Jr., who gave the world the dubious chant of nationalist zealots, reportedly in an after dinner toast to,

    Our country – in her intercourse with foreign nations, may she always be in the right, and always successful, right or wrong.

    a notion which has horrified educated and sensible people ever since.

    Along with much of the rest of our cultural and communication industries, music finds itself ever more beholden to a group of platforms which combine massive technical capability and the ability to operate at global scale. What happens to our world when more of our industry moves on to those platforms? GK Chesterton’s response to the jingoism of Decatur – “my wife, drunk or sober” – does not quite capture the indifference of the technology to the impact it has, or the awed helplessness with which we watch it reshaping our lives.

    Mappa Mundi
    The Royal Higden map – England is the red bit, bottom left.

    These platforms are now our central strategic challenge; how to service them and their users with our music in such a way as to shape a sustainable and profitable future for professional recorded music. We can see and measure the value that is draining from the surrounding ocean into these relatively closed ecosystems; with truly remarkable precision an EU study found during 2013-2015 that ‘137 Spotify streams appear to reduce track sales by 1 unit’, with an overall neutral effect on gross revenue. (Industry watchers will note the name Waldfogel on the cover – he’s got form in pushing an academic opposition to music industry lobbying.)

    But they are competitors for the value of music as much as these platforms are customers. To understand this from music’s perspective with harsh clarity it’s only necessary to look at how playlists change the marketing game. In the last few years we have moved from offering people a choice of platforms to buy our music from, to competing for mindshare within each platform with no sense that we either can or should ask a listener to switch from, for instance, Apple Music to Amazon. We know they won’t not matter how much windowing we do. The wholesale price captures only part of this value as platforms find more scalable and profitable revenue in advertising and consumer data.

    With a degree of complexity far beyond the old service or retail model these platforms are multi-sided business, which makes the challenge of servicing them and optimising our performance within them many times harder. At a technical level it’s no longer enough to set up some simple pricing and descriptive metadata, then drop the music and hope. More and more we are being offered continuous feedback on performance, and the tools to try to improve it. At a business level, competing within each platform while coordinating activity between platforms is becoming like three dimensional chess.

    Serpiente alquimica
    Serpiente alquimia – an ouroborus.

    For an analogy we can look at the adtech industry, itself a direct response to the impact of big platforms on media and marketing. Within the platform economy adtech creates no new value from the attention of consumers, but rather chases and coordinates value within an already closed system. Funding for adtech businesses reached $3.2b a year in 2015, after 5 years of growth.
    That is a lot of pricing power being hypothesised in what is essentially a tertiary business sector.

    In music we can afford a slightly different perspective, in that we can demonstrate the ability songs and recordings have to create new value. Apple’s colossal growth was at the very least seeded by the combination of great music and the iPod platform to unlock digital mobility, a transition that was partly funded by piracy as people filled up devices with illegal downloads. We need nevertheless to understand that we are essentially competing with our biggest customers to grab our share of the value that we jointly create with them, and we also need to build strategies that acknowledge the asymmetry of this relationship.

    Put simply, they have more data about our music and the people who listen to it, and more brains and more computers to understand that data with than we ever will. This is not however an argument for ceding the platform advantage and simply becoming tied tenants. The same reach and scale that enables the platforms we work with is available on the supply side, and we have the distinct advantage of being able to create competition and diversity against the inherent conservatism of big tech.

    To achieve this we need to take some simple but brave steps.

    First we need to break down our own data silos, to level the data playing field. Big data is truly our friend here as data needs to be sufficient rather than complete for patterns to emerge that really can guide and inspire business decisions. But no single label’s data is remotely near sufficient, so we need an environment that incentivises and rewards sharing.

    Next, we need to understand our competitive dynamics better, so we know when by fighting each other we are harming our whole industry. Many industries face similar strategic problems, and find ways of overcoming them, for example by forming standards bodies, or pooling intellectual property. Perversely music’s most collectivised parts – the collecting societies – are foot draggers in the world of open standards.

    Finally, we can look at what really creates new value and expands the market, and reward and celebrate it. This means a degree of unbundling of the unholy trinity of rights, supply chain, and marketing, along with much more transparency, at least for interested parties. Observable and measurable excellence should be what we honour with awards; we can and should invest in and recognise real expertise in relevant fields.

    None of these are easy, and all rely on music growing its own supply side platform ecosystem, as well as some parties reconsidering where they think their competitive advantage comes from. But we already do similar things in many areas of music, through collecting societies, through indie rights licensing body Merlin, through the standards consortium DDEX, and sometimes via trade associations. Strength, coordination and dynamism in the middle layer of the market will make a much better world for musicians and music lovers; it deserves our best efforts.

  • Record Labels! Beware of Fake Independence!

    Giant corporate groups are a fairly recent development in the music industry. The most extraordinary and wonderful music has always been produced outside the city walls by innovative and committed independent record labels.

    Around 40% of recorded music industry value is created by independent record labels, but 65% of that independent music is captured by major label owned distributors. Only 14% is truly independent all the way through the supply chain.

    This is not really so surprising. Scale is very important, and bigger platforms can provide global coordination between digital and physical products that is not available to the flock of indie distributors. That is an active choice and is fair enough. Much music migrates into major label systems by default though, and much is there despite the preferences of the indie labels.

    How does it get there? Some labels see themselves as incubators of talent, intending to pass successful artists up to the layers above for a price, and for them the relationship with a major label is valuable. Many labels get bought, or do a deal for financial support which locks them in.

    Probably the biggest factor is that many distributors start independent, but either run out of cash or have investors who want to get a return on their money. The Orchard, for example, started out as a beacon for independence, took investors’ money, and sold out to Sony Music. On its way it took other potential icons with it into the corporate belly, a migration taken by IODA, IRIS, DRA, Essential, and others. Feeding the beast seems to be a Sisyphean task. It’s not just Sony; the other global music behemoths are just as hungry.

    Here’s what this looks like from a drone:

    If music distribution was as simple and transparent as it should be these major label owned, pretend independent distributors would be a great adornment to the commonwealth. For many reasons they are a blight. Here’s why

    One: Stealing by stealth from the famous black box

    Fake indie distributors provide a very easy way to funnel unattributable revenue – the famous ‘black box’ – into central corporate bank accounts. Other benefits too, such as delivery fees, deal fees, marketing and promotion opportunities – all these are essentially stolen and allocated to global priorities and corporate profits.

    Two: Gatekeeping the label’s own metadata and music

    Labels and artists have many reasons for wanting access to their metadata and files, from public performance to promotion and synch opportunities. Understandably the effort of preparing digital releases and data is not something they want to go through twice. Instead of sticking to their service ethos, some distributors (and not just the fake indies) forget where the value comes from and end up acting as a gatekeeper to a label’s own catalogue and assets, or worse, trying to herd them into rights admin deals instead of letting them collect directly where they can.

    Three: Sitting on sales and analytics data

    All distributors act as a conduit for sales and other analytical data. But beyond feeding some selected parts into a pretty dashboard, many don’t pass it through to the label, and so end up with more insight into the music and artists than the label is able to get for themselves.

    A fair tradeoff?

    These might by themselves be a fair tradeoff for the services and scale that bigger companies can offer in the marketplace, and the advantages of access to some major label deals might well balance out some leakage. But more recent developments are highlighting just how pernicious fake indy distribution can be for genuine indie labels.

    So What’s New?

    A major label buy out was always a reasonable exit for an indie label owner wanting a quieter life. That is how the majors ended up so big in the first place. But digital distribution has almost certainly accelerated this as new dependencies make that sell out happen sooner. And given the data asymmetry, the buyer now has much more information than the seller about what a catalogue is worth. So from being that indie beacon, The Orchard is now one of the biggest acquirers of catalogues and labels in the market, on behalf of its corporate master, Sony.

    What this means is that with each acquisition there is more internal competition between owned catalogue and distributed catalogue, and you don’t need to be a genius to guess which will win.

    And back to that data asymmetry, which sets up a very bad incentive for the distributor and its owner by giving insights into artist development way ahead of when a label might be thinking it’s time for an artist to move up to a bigger home. Anecdotally some distributed labels have already seen artists getting targeted by the corporate owner of their fake indie distributor. The fact that major labels are buying sales report processing companies, such as RoyaltyShare and Korrect, shows you how strategically valuable this data is.

    Again this might be a fair and reasonable tradeoff for the scale, service, and access provided. But it might also be a toxic mixture of foreclosure and unfair competition. And if the latter what can be done about it?

    Prevention is better than cure…

    Starting with the basics, it helps to know who owns a distributor, and what their long term intentions might be. As a rule of thumb, the major labels are gatekeeping the more profitable revenue lines and trying to commoditise the rest, as well as ensuring as far as they can that there is less competition to sign the most promising new artists. Any distributor with venture capital or private equity funding – no matter how much they protest their indie credentials – will be looking to sell itself at some point, and will have given up to the financiers the right to decide who the buyer is.

    Next, read the contract! Carefully!! It’s normal in many contracts where there are potential future conflicts of interest to include ‘change of control’ clauses in case you don’t like the new owners after a sale. But at the very least it helps to know what rights and privileges are changing hands before it becomes an issue, and what you can expect not to be provided as well as what is being offered. And any gotchas, like the right to use your sales data without your knowledge or permission.

    And the practical steps. Keep your metadata and assets where you can use them without interference from your distributor. Those web based product uploaders might be convenient, but you can end up with no organised metadata at all about your releases and recordings. Store your sales and trends data too where you can use them yourself if you need to.

    Of course the best defence against being packaged and sold is to remain independent, and only work with partners who value independence as much as you do.

  • Streaming versus Downloads? The Fight is Far From Won

    The recent decline in the music download market, coupled with an acceleration in the streaming revenue numbers, has led many to predict the kind of transition that saw the end of the cassette tape, or indeed the cart and horse. There are a few reasons, some a bit less obvious, why this might not be such a certainty.

    For a start, building the higher bandwidth mobile networks required for pervasive access is expensive and difficult. Ofcom, the UK telecoms regulator, says most of the country has 50% or less 3G coverage. There are plenty of ‘not spots’ due to topography and building materials, so there are likely to be connectivity constraints and disappointments for the foreseeable future. On one of Britain’s main train lines there’s only a 10% chance of maintaining a 15 minute mobile phone conversation. Even if we wanted mobile music on demand we’re not likely to get it cheaply and reliably for years.

    The second aspect to look at is the nature of demand for music. Studies show that most people are happy with a library of 1000 tracks or fewer, a minute fraction of the 50 million tracks now available on demand. And by far the majority of music buyers, even in music mad countries such as the UK, traditionally listened to the radio and bought one or two albums per year.

    In a recent study consumers cited the ability to play music offline as the biggest benefit of a paid for music subscription. They get that with downloads, and with CDs, which perhaps explains the resilience of compact discs, along with the resale opportunity of course. What’s more, mobile data storage is going to get bigger and cheaper much faster than any improvements in mobile broadband. A small personal library, with a radio, and the occasionally impulse track download, is perhaps the ideal package for the casual music buyer.

    It’s likely that streaming’s rapid growth represents the migration of heavy download buyers to a more attractive value proposition, rather than the rest of the market trading up. There might be other reasons for the casual buyers to be enticed into streaming services, for instance their value as profiled targets of advertising. But as things stand downloads serve all but the top few music buyers with a much better mix of price and functionality, so they are likely to be around for a long time yet.

  • Change How We Think First, Technology Second

    Many years ago I was asked to discuss with the board of a trade association how the music industry was changing. One prominent executive summed up by saying “complexity has served us well”.

    A few days ago Spotify’s James Duffett-Smith was moved to exclaim “music industry licensing and copyright structures are legendary in their complexity” in a blog post on how the company is responding to a mini-crisis in royalty payments. What bit Spotify was no conspiracy to defraud on a grand scale, just a simple expression of the inherent laziness of the music industry, and its abiding ability to move problems to where they are hardest to solve.

    Let’s clear up some fundamentals before trying to understand how we might help ourselves towards a better and cleaner approach to bad data. Duffett-Smith oversells complexity. The basics of copyright and licensing are simple enough. Somebody writes a piece of music, somebody performs it, someone records it, and someone plays the record. A play of the recording usually involves permission and payment to the people who contributed. There are only a few permutations of this basic requirement, and in fact the music industry and governments have over time found ways to simply even further through block licensing and some limited, mostly benign monopolies.

    Some complexity in how rights and contracts are managed however enables musicians to offer the right kind of permission at the best price, rather than having to accept a single flat rate for an impoverished range of uses. This finds its best expression in the use of music in films and advertising, where a great song can transform a movie or sell millions of pairs of jeans. Neither Spotify nor other music services, nor public radio and most TV producers, nor indeed bars, hairdressers, garages etc., have to worry about such matters. They get simple commercial or collective licences.

    But even after 15 years of the data driven 21st century the music industry has not yet managed to create a coherent supply chain. So music services get their information about the music they play from two sources, which do not cooperate to ensure either accuracy or completeness. And it’s not because they can’t, it is because they won’t.

    Those two sources of data are the owners of the recordings on one side, and the owners of the compositions on the other. For very good reasons the rights and royalties for these two groups have taken different paths; songs and records have a Venn diagram of interests, and they need some protection from each other, even though in many cases the individuals concerned are the same.

    But when it comes to supply chain and data management, the industry has built silos. There is no matching of data between the two sides before the tracks are sent into the market, and no collaboration before the official release date to make sure the songs and songwriters for each recording are properly identified. And the law excuses this, fully transferring the risk to the point where the music is played.

    Observers will note an irony here, as the music industry is notorious for its high profile plagiarism court cases. And added to this there is a fundamental right to attribution in most copyright codes (reinforced by an item in the UDHR) which is ignored on an industrial scale by the industry that most stands to benefit should it be universally respected.

    I have argued elsewhere for a simple tweak to either law or industry practice that would see us indemnifying parties which rely on the data we provide about our music and our rights. This seems to me to be simple good sense. But there is an ethical dimension to the question too, for it is surely everyone’s right whether they are a musician, or any other kind of creator, to be identified with their work wherever it is performed or exhibited. And that is the change in thinking we need in the music industry, to a cultural norm where to hide or omit the creators’ due credit is seen as the shameful and lazy act that it truly is.

  • The World Is Singing A New Song

    We live in an age of everyday activism, with every purchase an expression of values as well as needs. Ideas about fairness go far beyond the impact of certified fair trading schemes; they permeate consumer marketing. Social media drives a radical transparency in supply chains, with no exploitation tacitly excused.

    Alongside fairness we have discovered a new appreciation of the craft skills and tradition wrapped up in the idea of an artisan. We know that we can be fed and clothed, and that our homes can be furnished and decorated by machines. At some point however we lost the sense that mechanisation is progress, and what came to the surface is the powerful idea that all economic transactions are a medium for human communication. A loaf of bread is now, somehow, a laying on of hands as well as nutrition.

    A musician playing an instrument or singing with real skill and feeling is the very definition of an artisan. The century we’re in could have been made for musicians, whose work delivers all the meaning of a transaction with none of the burden of common wants and needs. If this seems too much of a stretch, consider a cup of fair trade coffee. How much does the ethical content really improve the world? Or how much is it just an excuse for a caffeine dose in a disposable cardboard beaker?

    We in the music industry should not get carried away with enthusiasm for ideas that are easy to express but much harder to live up to. Some voices are now mooting a ‘fair trade’ approach to music. It is very hard indeed to see the parallels between the life of a musician in a G7 economy, and the fate of a worker on a coffee plantation. Nor would the Walk Free Foundation find many major label recording artists among the 35.8 million slaves it works so hard to free.

    Fair trade and authenticity in the music industry have to start from a very different place, and it will need the kind of grinding attention to detail that we have in abundance, but have not so far deigned to put on display. Today it is hard to find out who created a song, and who played the instruments. Such credits as there are don’t really tell the whole story. Was the performance recorded simply and directly, or did a studio wizard remanufacture it into something entirely wonderful but different? If we answer these simple questions we create trust, and trust is the foundation of both fairness and authenticity.

    But this comes with a warning. We won’t be correcting a great social injustice by going down this road, and we should be very careful not to wear the clothes of the genuinely oppressed and suffering. Fairness in music comes when everybody’s contribution is acknowledged and everybody’s rights respected; it’s not an argument over percentages of a payout.

    We have to admit in music that we are years away from the time when we can stand behind a full set of credits for a recording and swear that they are complete and truthful. Our audience, and consumers, have in many ways gone ahead of us, and it’s they who are singing a new song. We need to sing in harmony with them. It’s simple commercial good sense, as well as the right thing to do.

     

  • Listen to the Music, not the Audio Nihilists

    The trade offs that made digital music possible, when storage and bandwidth were scarce, are no longer quite so necessary through much of the developed world. Some brave services are trying to take advantage of this technological advance by offering higher quality music on demand. They face considerable obstacles, and oddly, a chorus of voices shrieking ‘snake oil’ and calling the whole endeavour a con. We should listen to the music, rather than the audio nihilists. Here’s why.

    A little compression is very hard to notice; unsurprisingly the more sound that gets removed the more likely the listener is to notice a degradation. This is only a small part of the story however. The best and most professional do everything they can to ensure the end product is as good as it can be, right the way through the recording and production chain. But they have a low target to hit, and nobody can easily justify overspending in a tight market. Many of course deliberately take short cuts in order to save time and money. Now that nobody can plead necessity, having a higher final target would be likely to raise standards generally, and this would be a good thing for music fans.

    Not all ears are equal, but recording and compression by their nature target a mid point of hearing acuity. And now the discussion becomes a bit speculative. Higher pitched sound is harder to compress, but is important both for the complexity of timbre as well as locating a sound in space. Location in a stereo signal is a bit of a fake anyway, as obviously a single audio source needs to be converted to two separate sounds to be perceived as either to the left or the right in a normal listening environment. In nature, as well as tiny time and level differences, it’s the wrinkles in your ears that subtly reflect and modify high pitched sound in ways that you have learned to associate with position. Because of this complexity the recording or sound creation process has already made many compromises in order to sound mostly OK to most people, and compression adds another layer of compromise.

    But still, as many experts say, most people can’t reliably tell the difference between the source and its compressed version, and express no preference either in listening tests. Surely depriving them of what they can’t hear does not hurt them!

    Sound, however, is not music, and hearing is only part of perception. Nor can simple tests really hope to discover the immensely complicated ways we respond to music, nor how that might be affected by the ways we experience the sound. We all know this of course. Our imaginations fill in the missing parts of familiar instruments, so that even a sketchy recording of a violin shares some of the wonder of a platonic Strad. We are continually exposed to newly imagined sounds, which we fit into our own personal collections of virtual instruments and performances. Musical training even alters the structure of the brain, it is that affective. Even if the audio testing equipment says that the sound you hear now is exactly the same as the one you heard five minutes ago, your perception of it has been irrevocably altered by both memory and imagination. We create what we hear as much as perceive it.

    So for these reasons we should not let the audio nihilists deprive us of the opportunity to create and experience a world of sound and music that could take us far beyond the standard fixed by the CD and then compromised by compression to fit the business needs of the late 20th century. We should and can encourage a music industry that can really stretch our ideas of what is possible in audio perception.