Welcome to Edition #49, and a happy April to all.
“Rustle twice if you’re in there.”
Upmixology in audio
I enjoyed this Jesse Jarnow article in Wired about how companies are using AI, specifically deep learning techniques, to isolate certain sonic aspects of old songs. Once a song can be broken into its component parts (“demixing”), it can be reassembled with a polished set of modern-quality instrumental stems (“upmixing”).
One of the core techniques at play is called spectral editing: treating sound as a visual object. Once a song becomes a spectrogram, individual elements can be identified and isolated using a range of approaches including neural nets. The result is that you could, say, programmatically recognize and extract just the signature representing the rhythm guitar track from a classic Beatles song. Or, as Spotify and others have done, eliminate the vocals from any number of pop classics to allow for a vast, easy-to-access karaoke archive.
My first exposure to a primitive version of this type of editing came in the innocuous form of backing track CDs that came in laminated sleeves at the back of the guitar workbooks I was given as a kid. On those CDs, you’d have jazz standbys performed by full bands, except for one instrument: yours. It was always a novel experience to hear those songs without one of their key ingredients; the flip side of the coin is to remove everything but that key ingredient.
The potential for sampling and remixed creations, using original stems, is massive. I’ve felt this firsthand, having thoroughly enjoyed experimenting with some (very bootleg) versions of this approach last year while trying to hip-hopify the bass and drums from a 1972 Tim Maia track for a 2020 remix. Thorny questions like copyright and permission do, of course, apply. At what point is it “ok” to produce polished, new-sounding versions of old classics? Do you need the artist’s consent, if you’re not actually changing any of the underlying music? And what is the level of differentiation at which a derivative composition becomes the work of the sampler, and not the original performer?
All these questions are nuanced and hard to answer at this stage. But, it’s exciting to see yet another frontier where technology can change the way we create (and consume) sound.
Spotif(ireside) chats
Live audio companies are leading their best lives. This week, Spotify announced its acquisition of Betty Labs: parent company of Locker Room, a sports-centric live audio platform. (Tech Crunch) I have, on occasion, used this product in a (futile) attempt to make myself feel better after watching my Celtics throw up on their collective shoes on national TV with apparent disregard for my fragile mental state.
It is expected that Spotify will leverage Locker Room’s functionality as the foundation for a more horizontal approach to live audio, going beyond sports to other core verticals. The acquisition fits nicely with Spotify’s prior acquisitions of podcasting companies and podcast enabling products (e.g., Anchor); it will also give creators another format with which to experiment and monetize.
More importantly, this deal gives Spotify a credible competitor to Clubhouse, Twitter Spaces, Mark Cuban’s Fireside, and the 6,703 other live audio apps that you have yet to hear about but that will probably become unicorns before the next issue of Trillium reaches your inbox. Spotify has long been explicit about its ambitions to “own the ear” and establish itself as the dominant player in audio. The first step was going beyond music to non-music inventory, and specifically podcasts. Now, the company can explore a more social—and live—audio format.
It will be interesting to see how quickly Locker Room’s product is integrated into the main Spotify app, and how quickly the company diversifies the audio rooms beyond sports. I’ll also be curious to watch how Spotify monetizes live audio, and whether other recent-ish acquisitions like Megaphone are utilized in order to help creators turn live conversations into commerce.
(Sub)stacking even more paper
Last week, I wrote about the Substack “backlash” over the company’s payment of writer advances to conservative writers (and the perceived lack of appropriate disclosure). This week, I am writing about how the company raised $65M on a $650M valuation. #Controversy!
The company’s round was led by A16Z. Funds will be used to “invest in writers”: expanding the company’s advance program, rounding out its community programming (fellowships, grants, educational resources, etc), and adding more wraparound benefits like healthcare, editing, design, legal, and so on. As Twitter, Facebook, and others have entered the paid newsletter space with a vengeance, Substack seemingly felt a need to add capital in order to accelerate development.
For all this talk of investing in writers, I’m still waiting for that first check. 🧐 The phone lines are open! All currencies accepted.
Cameo the Unicorn
Ah yes, the money machine continues to whir. Cameo—uncharitably described by Recode Media’s Peter Kafka as an app “that lets faded celebrities record birthday messages for a few seconds”—has raised a $100M Series C that values it at over $1bn. (Tech Crunch) The round was led by e.ventures and includes Google Ventures, SoftBank, Amazon Alexa Fund, and many others, including celebrities themselves. In related news: next week, I will announce that Trillium has raised $350M (using ETH as the funding instrument) on a $4.7bn post-money, and is debating a mega-SPAC situation.
I like Peter Kafka but, as someone who has used Cameo happily, I do see the merit in its product. Is it good? Yes. It is $1bn+ valuation good? This, I cannot answer for you. But, someone thinks it is, and that’s all that matters for right now! What I will say is that Cameo is well-positioned to capitalize on the “direct-to-fan” monetization trend that we’ve talked about in Trillium for months. And, it gives creators a way to commercialize the parasocial relationships—here’s looking at you, Reader CM—that the Web 2.0 era has strengthened.
Cameo itself is working on new ways to augment its original product, including Cameo for Business and other monetized services that creators can offer to their superfans. Cameo for Business strikes me as a no-brainer, and a great way to profit from corporate expense accounts and event budgets, particularly in The Zoom Zone. I would expect other experiments to focus on “gifting” behaviors, as that seems to be a natural entry point for many into the service.
Discordosoft? Microcord?
Recently, reports emerged that Microsoft is considering acquiring Discord for somewhere around $10bn. (NYT) Discord is best known for its popularity in the gaming community; its core product enables users to set up “servers” that can be customized to host a range of live text & audio conversations across themed channels. For those who haven’t used it, you might find it akin to an edgier Slack.
In some ways, Discord’s versatility reminds me of Roblox. It’s a flawed comparison, but both products started out with a focused gaming use-case, and have expanded to support a range of other experiences across verticals. I’ve seen plenty of creator Discord servers—some of which require paid membership to enter—and others corresponding to companies, projects, and a host of other topics. For many users, the primary community interaction around their favorite subjects or individuals is mediated through Discord.
I wrote about Discord and what I pseudo-ironically (ok, very ironically) referred to as “artisanal social networks” in July 2020. The general observation was around the profusion of these micro-networks, with creators as organizing principle, in which there existed not only creator-fan interaction but also fan-to-fan interaction. Allow me to quote myself:
If you visit these communities, you’ll find that they’re often relatively vanilla, with different threads for a range of standard topics like gaming, sports, movies, music, and so forth. As one hobbyist podcaster…with a small but rabid community of around 500 subscribers said to me: his fans talk about pretty much everything… often without mentioning the content that ostensibly brought them together…The only thing that distinguishes these mini-social networks is the implicit knowledge that everyone in them has been granted access because they like the same creator.
I agree with the conventional perspective that the Microsoft-Discord deal makes good sense because of Microsoft’s numerous investments in gaming—ZeniMax Media, Minecraft’s developer, and, of course, Xbox—and suspect that the majority of financial value and direct product integration would, at least initially, come from the company’s gaming business unit(s). But, I remain equally interested in the implications for creators, companies, and communities that are using Discord to stay connected. Given the suite of Microsoft products that could be leveraged by these individuals or organizations to build businesses, Discord could become a meaningful piece of Microsoft’s productivity/social/community “stack” beyond the world of games.
NFT(o The Moon)
Don’t worry: this is not going to be another self-congratulatory NFT thinkpiece about how Art, Commerce, and Probably Humanity Are Fundamentally Reshaped Forever. I just feel a particular responsibility to you, the global Trilliuminati, to keep your fingers on the collective pulse of this manic and still-rather-fascinating market.
Digital collectibles has always been a leading use-case for NFTs, starting with CryptoKitties and gaming before evolving into NBA Top Shots and sports. Now, Dapper Labs, the company behind both CryptoKitties and Top Shots, has raised $305M on a $2.6bn valuation. (ESPN) The company added 800k users since its public beta less than six months ago, and recognized $500M+ in sales.
Investors apparently included everyone with a bank account. Specifically: Coatue, A16Z, Chernin Group, the NBA, Warner Records, Ubisoft, and plenty of athletes and celebrities, naturally. Oh yeah, and MJ. I would love to see that cap table one day. Part of the thesis here is that Dapper’s underlying Flow blockchain will emerge as a credible alternative to Ethereum, becoming the basis for a range of other Web3 projects.
For more on NFTs & sports, I recommend Bill Shea’s piece for The Athletic: “Blockchain, crypto and NFTs are the risky ‘next generation of value creation’ in sports.” One area that I’d be curious to explore is NFTs that confer hybrid physical-digital value, meaning that ownership provides some sort of real-world benefits (special tickets to the Draft or playoff games, meet & greets with players, and so forth) beyond financial speculation.
Elsewhere in NFTerritory: OpenSea, self-reportedly “the largest NFT marketplace,” raised an additional $23M from investors including A16Z. (Coindesk) The company’s product also pulls in listings from other marketplaces, including Zora. [Ed.: this is how one can come to find my magnum opus for sale within its vast archives.] For the record: I came up with the “concept” for this masterwork as a way to amuse myself during a particularly excruciating meeting, and none of you successfully talked me out of it, so now we all have to suffer by having a Decentraland-themed “sheepish meme” etched for all time onto The Blockchain. [Ed.: shockingly, still no bids.]
In addition to OpenSea, SuperRare (an NFT marketplace that specializes in 1-of-1 editions) also secured a fresh $9M in funding. (Tech Crunch) Basically: creators need places to mint NFTs, and consumers need places to buy them. Companies that fill this role are raising money. Oh yeah, and Atari has debuted an NFT Capsule Collection that features Pong and Centipede-themed items, including “a 3D NFT of the original Pong cabinet featuring a digital autograph by Al Alcorn,” one of the game’s original engineers. What a world.
For your ears only
Every so often, an album gets released that completely captivates the attention of the Music Media Industrial Complex. That doesn’t necessarily mean the album is good, but it does mean that you can’t escape its existence.
So, it’s fair to say that I was somewhat reluctant to listen to Promises, the recent release from Pharaoh Sanders and electronic artist Floating Points (Sam Shepherd). I don’t listen to much jazz, and it’s been a few years since I spent any time with Floating Points. [Ed.: I liked Elaenia, from 2015, but couldn’t even make it through Crush, from 2019.] I have to admit, as happens all too frequently these days: I was wrong to be skeptical! But seriously: there were a lot of articles about this album.
Promises is one of my favorite albums of the Quarantine Era. I have listened to it constantly over the past week: at meals, while walking, when writing, prior to sleeping, and so forth. It is worthy of the hype. For context, Sanders is one of the all-time great jazz saxophonists; he has performed with legends like Sun Ra and John Coltrane. This album sees him collaborating with the London Symphony Orchestra and Shepherd (as Floating Points).
There’s an ambient aspect to the album, but it’s much more engaging because of the melodic themes that repeat throughout its nine movements. You’ll hear thematic elements carried by synthesizer and saxophone, string-oriented counterpoint (“Movement 6”), moments that sound like “Pyramid Song” mixed with 1950s film soundtracks, and many other sonic elements. It’s hard to describe, and very much worth your time.
See you all next week.
N