James Hill – everblog

The Paint Shop and the Art Gallery

Imagine a paint shop. A huge, sprawling warehouse filled with cans of paint; every colour you can imagine and then some. Imagine you could have any amount of any colour you want for just $9 a month. What a deal! If you need to paint your house or bedroom, you can’t beat it. Nowhere else can you get more paint for less money. This is Spotify.

Imagine a small art gallery next door to the paint shop. They sell original 1-of-1 paintings and limited-edition prints. The paintings are surprisingly expensive and the prints are more affordable. Nobody comes into the art gallery asking to buy “the canvas with the most paint on it” because they’re trying to get a good deal. Here, the value isn’t in the paint itself but in the artwork. This is ever.fm.

The point is: the paint shop and the art gallery can happily co-exist. They do not cannibalize one another. If a customer wants to paint their wall, they go to the paint shop. If a customer wants a painting for their wall, they go to the art gallery.

The paint shop and the art gallery can happily co-exist. They do not cannibalize one another.

The paint shop and the art gallery both sell the same product – paint – but they do so in entirely different ways. You’ve seen photos of a bird sitting inside a crocodile’s mouth, cleaning its teeth? This is a bit like that. There’s a symbiosis between the two because they don’t compete directly.

This is how I think the near-future of music consumption will look. Spotify isn’t going anywhere. There are too many people who need to paint their walls with music and the allure of high-volume, low-cost music-by-the-gallon is undeniable. But Spotify doesn’t sell 1-of-1 original art pieces by my favourite artist. It’s simply not something they offer. To get that I need to go next-door to the art dealer.

The Fear of Change

Over the years, I’ve spoken with many artists, fans and collectors about the idea of “recordings that change.” In the early days of ever.fm (before it was even called ever.fm), I sent out a questionnaire to 100 friends, fans and colleagues. Here’s what I found.

One of the questions on the survey was, “What if a recording could re-write itself every time it was played? For example, there might be new lyrics, a different guitar solo or extra harmony parts. How would you feel about that?” The responses were really interesting. 63% said they think they would like it, while 11% said they probably wouldn’t. The rest were undecided.

I was especially interested in the open-ended comments. “Sometimes I like a song because of the predictability of it,” wrote one person. “It’s the familiar tunes that feed my soul and give comfort, bringing back important memories,” wrote another. “I like the connection I feel through familiarity,” A theme begins to emerge as you read the comments: anxiety. Specifically, anxiety about change.

Of course, “fear of change and the unknown” tops the list of the most common anxieties we feel. But what struck me in the survey responses was that it seemed to affect artists and fans equally. Artists feared that if any part of their musical creation were to change, the work might lose its meaning or essence. For fans, the reaction was even more visceral. They seemed to be saying, “when I fall in love with a song and then something about it changes, I fear that I’ll fall out of love with the song.”

A theme begins to emerge as you read the comments: anxiety. Specifically, anxiety about change.

Isn’t this an echo of the way we feel about everything, including our personal relationships? When we fall in love, we make ourselves vulnerable. When we let someone (or something) into our hearts, we do so in spite of the risks. The “emotional Trojan horse” scenario is terrifying: the thing you fell in love with turns out not to be the thing you fell in love with.

What’s a heart to do? Not love? Never let down its guard? Protect itself at all costs? We know where that leads. Total protection is self-destructive, like a plant that “protects” itself from water and sunlight. Water drowns. It also feeds. Sunlight burns. It also heals. Change itself isn’t the culprit; it can go either way. It may be true that to change a single hair on the head of your favourite song would ruin it forever. It might also be true that you’ll fall more deeply in love with it as you discover more about it.

That’s the promise of ever.fm: discovery. Here, for the first time in history, is an audio medium that allows artists to create worlds inside a single song. Layers to dive through, secrets to reveal, hidden gems that reward the intrepid listener, and sounds that might even surprise the artist. I’m not saying this is not a time to be fearless; fearlessness breeds recklessness. No, this is a call to embrace fear. To stand on the banks of a cold, clear swimming hole, to know it will take your breath away, and to will your legs to jump.

Fans, Speculators, or Both?

Extra! Extra! The NFT universe is expanding at light speed! The era of digital collectibility is upon us! NFTs will disrupt the music industry! It’s all very exciting (and I mean that sincerely) but something is sticking in my craw. Something is keeping me from fully believing the message of music NFT evangelists. What is it? What is it?

Maybe it’s this: NFT platforms often assume that fans are speculators and that speculators are fans. I think it’s a mistake to conflate fandom with speculation. A fan is a person who falls in love with an artist and their work. Love is blind and irrational. Love cares nothing for markets, secondary sales, or the inner workings of cryptography. Fandom is as fandom does. A speculator, on the other hand, loves the thrill of the chase. Understanding markets, secondary sales and cryptography is critical in generating ROI. The fan and the speculator are two very different animals. While some people embody a bit of both, generally speaking they’re two distinct species.

[We] often assume that fans are speculators and that speculators are fans. I think it’s a mistake to conflate fandom with speculation.

Were you the kid who collected comic books to read them, follow the story lines and fall in love with the characters? Or were you the kid who kept your comics in plastic covers, hoping to sell them for a profit down the road? Maybe you did a bit of both but one approach or the other probably resonates more strongly with you. My point is: fans and speculators are not one and the same. What does that mean for the way we design products, platforms and services around collectible digital art?

It means what it has always meant: know your customers and speak to them in their language. Early-adopters, especially in the crypto world, are likely to be more tech-savvy and have a stronger speculator streak. It makes sense to put technology and ROI at the forefront of the conversation. For speculators, the fun is in the hustle. But these same talking points can be alienating to fans who have no interest in ROI. The bigger the NFT world becomes, the less important ROI will be to the average collector.

Were you the kid who collected comic books to read them, follow the story lines and fall in love with the characters? Or were you the kid who kept your comics in plastic covers, hoping to sell them for a profit down the road?

The question for artists, developers and platforms at this still-early stage is: who are you speaking to? Fans, speculators or both? Do you indulge the speculator’s appetite for tech and ROI or speak in simplified terms that appeal to the fan’s love for the artist and their work?

Anyone who has ever built a campfire knows that the fuel changes as the fire grows: first it’s paper and twigs, then thicker branches and, finally, logs. If it’s true that the NFT revolution is upon us, it means that we’re no longer burning paper and twigs. It means we’re not speaking to exclusively early-adopters. It means we can’t assume that fans want to be hustlers. It means we’re entering an era defined more by irrational fandom than by calculated speculation. It’s a transition that I, for one, welcome with open arms.

What Does Generative Music Look Like?

The Evolution of ever.fm’s “Mandala” Visualizer, from stick-figures to beautiful art pieces.

Seeing is Hearing

On ever.fm, what you see is what you hear; every rendition is accompanied by a beautiful visualization. These “sound mandalas” serve two important functions: 1) they beautify the ever.fm experience, and 2) they help you make sense of what you’re hearing.

At first, the colourful, circular visuals might seem random, but they’re far from it. They’re generated from the audio data in a song and show a circular “timeline” of the piece, each concentric ring representing a different track in the master recording. Furthermore, the artist is able to set a number of parameters that affect how the visualizer will behave. In this way, the visuals are truly an extension of the artist’s vision.

*Artists can create a wide range of generative visuals to accompany their works.*

Dolls, Blobs and Stick Figures

When I first approached Andreas, my co-founder, about creating a visualizer for ever.fm, my ideas were, to put it mildly, a hot mess. I knew we needed visuals so that listeners could see the change from one rendition to another (even if they might not always hear the change), but that was pretty much it.

My initial ideas look pretty hokey in retrospect: I imagined a kind of side-scrolling video-game animation in which a figure would walk through a landscape, the background of which would represent the waveforms on each track. Seriously. That was my initial idea. We still roll our eyes about it.

*One of my initial sketches for the visualizer.* 🙄

I also brainstormed a few other possibilities: a doll-like figure whose outfit changes with each rendition of a song, a zodiac-looking arch with “constellations” generated from the audio data, and a “blob” that would float on-screen and change shape based on a rendition’s audio properties. We had a long way to go but we knew two important things: 1) we needed a visualizer, and 2) the visuals should be generated from the waveform data in a song. It was a start.

*More early ideas for the ever.fm visual style.*

At some stage, our approach became (mercifully) more abstract. This was largely Andreas’ influence, as I recall. After a few of weeks of meetings, the dolls and stick-figures disappeared from our sketches and were replaced by concentric rings and line segments. We started a mood board and filled it with colourful, geometric, mid-century abstract art.

*Early sketches for the mandala, exploring how each ring could represent part of the master recording.*

On our weekly calls, we debated: how abstract is too abstract? How literal is too literal? How do we make each rendition unique and yet make them like they belong to a “family” of renditions? How do we ensure that every rendition is beautiful? Or do we? We were inspired by Tyler Hobbs’ work and devoured his essay on “long-form” generative art.

*More sketches. We’ve come a long way from dolls and stick figures…*

We took stock. Here’s what we had so far:

Circular design
Each ring represents an audio track
Allow artist to customize colours

At this point Andreas took over. He made quantum leap after quantum leap, putting together the visual editor in just a few weeks. Clarity is magic.

A example of waveforms in the mandala. This is an EDM track and you can see the regular beats.

The final touch: animating the visual so that it spins like a post-modern record as the song plays. Voilà! The story of the ever.fm visualizer. I’ll admit, I really love the results. I love that the visuals are flexible enough to evoke an artist’s unique vision and personality. I love that the visuals are engaging enough to pull the listener into a piece of music but abstract enough that they don’t distract from the listening experience. I love that ever.fm is an art gallery of sorts, an online audio-visual Guggenheim where music and art are beautifully entwined.

There are still many avenues to explore when it comes to visualizing generative music. Beyond spinning like a record, what other animations could help to bring the visuals to life? Could we add animated transitions between renditions to make the UX more seamless? How do ever.fm visuals interface with AR and VR? There’s always room to grow, dream and be creative. But for now, the ever.fm visualizer gives artists a ton of room to play, explore and express new facets of their songs.

There Will Never Be Another Beeple

Will music ever have a “Beeple moment”? Probably not. And here’s why I’m OK with that.

Where were you when Neil Armstrong set foot on the moon? Where were you on September 11, 2001? Where were you when Mike “Beeple” Winkleman’s Everydays: The First 5000 Days sold for $69 million at Christie’s?

The Beeple sale was huge. It was a huge pile of money and a huge moment for NFT art. On March 12, 2021, NFTs turned a corner and entered the public discourse. In an industry where “firsts” are prized, this was a “first” for the ages.

As a musician, I was cheering from the sidelines wondering if music might be next in line to reap the rewards of mainstream NFT adoption. If so, who would emerge as the Beeple of the music industry? Would it be Jay-Z? Beyoncé? Justin Bieber? Or a previously-unknown breakout artist? My guess: none of the above.

As a musician, I was cheering from the sidelines wondering if music might be next in line to reap the rewards of mainstream NFT adoption. If so, who would emerge as the Beeple of the music industry?

It sounds pessimistic but let me explain. I don’t think there will ever be another Beeple. It was a moment in time, the product of a perfect cultural and financial storm. What’s more, I don’t believe there will ever be an MP3 that sells at auction for $69 million. Even if it was Jay-Z singing opera with Beyoncé on jaw harp and Justin Bieber on hurdy-gurdy I still don’t think it would garner anywhere near that kind of money. Why? I wish I knew. We simply value visual art differently than we do audio art. A JPEG can sell for millions, but not an MP3. (The most anyone has every paid for an audio recording was the $4 million PleasrDAO acquisition of the 1-of-1 Wu Tang record Once Upon A Time in Shaolin.)

What does this mean for music? Well, if you can’t sell a single song for a million dollars, maybe you can sell a million songs for $1 each. But who has a million songs? Songs take time to percolate and to produce. And there’s the catch-22. Right there. In order to succeed in the music-NFT landscape, songs need to scale. For most artists, selling a song – or even 100 songs – to the highest bidder won’t be enough to generate a sustainable level of income. But what if one song could be assembled in a million ways? And what if each unique rendition of the song could be bought, sold and re-sold by fans?

If you can’t sell a single song for a million dollars, maybe you can sell a million songs for $1 each. But who has a million songs?

This is the promise of ever.fm. ever.fm turns one song into many renditions, each like a spark struck from the flint of the song’s master. With no definitive version of a recording, a song becomes the sum of its renditions, each one helping to express a unique facet of the artist’s vision.

In my view, there will never be another Beeple. What’s more, audio art will always be valued differently (read: more modestly) than visual art. But this may be a feature, not a bug, especially during recessions and bear markets. If we can just solve the problem of song scaling – creating many meaningful 1-of-1 renditions from a single master – we may never need another Beeple.

Um… Why Does ever.fm Exist?

What’s the point of ever.fm? Why create recordings that change?

For thousands of years, songs were ephemeral and ever-changing. Songs were carried on the wind, growing like wildflowers wherever they landed. For thousands of years, songs were alive. Then, suddenly, they were turned to stone.

Edison may not have set out to create the record industry (after all, his machines were designed to replace stenographers, not musicians) but, nevertheless, the phonograph cast a powerful spell and there was no going back.

For thousands of years, songs were alive. Then, suddenly, they were turned to stone.

For the first century of its existence, the value of recorded sound followed a general upward trend, peaking in the mid-1990s with the convergence of cassettes and CDs. Then, in 2002, came Napster, the dawn of digital music sharing and two decades of economic tailspin for the music industry. Music-industry revenue declined sharply from a peak of $21.5 billion in 2000 to $6.9 billion in 2015. Since then, the industry has seen growth due to the mass adoption of streaming music services like Spotify, Apple Music and Tidal. However, the market value of a recording has changed forever: Spotify currently pays rights holders a fraction of a cent per stream. An artist must garner over 1,000 streams of a song just to buy a coffee.

An artist must garner over 1,000 streams [on Spotify] just to buy a coffee.

Success in streaming hinges on one thing: volume. Because of the way Spotify’s payouts are structured, artists with millions (or billions) of streams get a disproportionate cut of streaming revenues. That obscure band you binge-listened to all month long? They probably didn’t get penny from your subscription fee; it all went to Drake and Taylor Swift. While there is a movement among streaming services like Deezer and Tidal to adopt a user-centric payout model that might compensate artists more fairly, this approach has not been taken up by industry-leader Spotify.

“[Streaming] works if you’ve got thousands or millions of songs,” says Mark Mulligan of tracking firm MIDiA Research. “But if you’ve only got 20 or 30 or 100 songs then it doesn’t. You need scale of catalog to benefit.” Big artists get bigger, small artists get smaller. Without a drastic re-imagining of the recorded-music medium, the dream of the independent musician, fostered by 1990s-era optimism, is dead.

So let’s re-imagine it.

In the beginning was the wax cylinder. Then came a succession of physical formats: vinyl, 8-track, cassette, and CD. Finally, music was set free from its physical form with the advent of the mp3. Throughout this metamorphosis, one thing remains constant: recordings never change. Play, stop, repeat is all we ever get. It’s as true today as it was in 1877. One listen is like every other; it makes no difference if I listen now or later. There is little urgency or scarcity in this world of static music. More than ever, music has become the proverbial “wallpaper” in our lives: ubiquitous and forgettable.

But what if recordings could change? What if every listen was non-fungible? What if we realize the promise of disembodied music? Before the existence of streaming, blockchain and NFTs, this was unthinkable. Now, it may be a way to rebuild the value of music in the web3 era.

What if recordings could change? What if every listen was non-fungible?

ever.fm has created a new format for recorded music, one in which a song is reborn every time it plays. It’s streaming music with superpowers. As a song plays, the listener can shuffle the inner workings of the music with the touch of a button, cycling through myriad possible renditions until they find the one they love. These renditions are generated on-the-fly by ever.fm using two ingredients: audio samples and playback rules, both of which are created by the artist. With every rendition comes new discoveries: never-before-heard combinations of instruments, new lyrics, harmonies, alternate takes and more.

On ever.fm, a song is the sum of its renditions the way a flip-book is the sum of its pages. There is no definitive version of a recording. Instead, every song is a collection of possibilities, and each possibility is an NFT that can be collected, shared and re-sold.

On ever.fm, songs are alive.