Where Apple sees the future of streaming

Right now Apple Music is laughably bad at providing metadata beyond the most basic track information. Where does the company want to go?

By Mihir Sharma

The small but diverse group of classical-music lovers is in deep mourning after one of the pillars of its community died. Primephonic, a Dutch-American app that streamed a wide catalogue of classical music went dark last month, after being acquired by Apple Inc., which aims to fold the service into Apple Music. But why would the world’s largest company be interested in a closely held startup with a relatively small user base, a few dozen employees and no startling technological innovation to boast of?

The answer: Primephonic understood the future. Apple has realised that streaming services will succeed or fail depending on whether they master the four things the tiny company, along with its classical-music peers such as Idagio, have figured out: metadata, discovery, curation and quality. And that will hold true for video streaming, too, not just music. With the buyout, Apple is hoping to absorb Primephonic’s DNA.

Right now Apple Music, like most music-streaming services, is laughably bad at providing metadata beyond the most basic track information. The software developer and composer Gene de Lisa regularly tweets out howlers such as Apple Music listing the composer of the Pathétique Sonata as “unknown.” (Poor Beethoven.)

Without the right metadata, the user interface breaks down; you don’t know which movement is playing from which symphony. Search breaks down, too: Without metadata for conductor, orchestra, composer or movement, you can’t find the recording you want. And, worst of all, effective recommendations and music discovery become virtually impossible.

This isn’t only a problem for classical music ‑ it matters for anything more complex than the latest Top 40 single. In India, where we listen to Hindi film music, you might remember a song based on its singers, its composer, its lyricist, which film it was from, which year it was released or even which actors lip-synced the song in the movie.

Classic rock is the same: when a Beatles track pops up, you should know immediately if it’s the stereo or mono mix, the original single or the album version, the 1995 Anthology alternate take or the 2000 remaster. Search for an early R&B number and you might get some bad knock-off “re-recording” made in the 1980s when the original group was down on its luck. For anyone with more than a casual interest in music, streaming without proper metadata is a disaster.

Primephonic, because it worked on getting metadata right – and searchable – learned about its users faster as well. Especially compared to, say, Tidal, which knows I have spent two years listening to the Vienna Philharmonic but still recommends hip-hop or (grudgingly) a playlist of movie soundtracks.

Knowing your user matters. We don’t always open a streaming service already knowing what we want to hear or watch; the discovery process is half the fun. Primephonic didn’t just get algorithmic recommendations right, it had actual experts curating playlists – famous violinists picking their favorite violin-focused recordings, for example.

As you listen to an album you’ve discovered on Primephonic or Idagio, you can read the CD liner notes ‑ and even, through Primephonic’s “maestro” service, get real-time notes on what to listen for in the orchestration. In the video world, the Criterion Channel regularly gets major filmmakers to pick and introduce their favorite movies.

Then there’s sound quality. Tidal – and its French competitor Qobuz – sell themselves as having high-resolution “better than CD” tracks on offer. Primephonic did the same without much fuss and at a fraction of the price. Apple recently announced “lossless” streaming, and Spotify promised high-fidelity music months ago. But high-resolution tracks aren’t really a priority for Apple compared to the far weirder world of “spatial audio.” Apple has worked hard to install Dolby Atmos on hardware such as Airpods, for instance. But the company’s implementation of high-resolution streaming through their proprietary Airplay streaming protocol is buggy and confusing.

This is a mistake. Scientists disagree about whether high-resolution streaming makes a perceptible difference to the human ear. But it’s hard to find a classical fan who believes it doesn’t. And it is niche, high-end classical music buyers who drive new standards in music reproduction. RCA Victor’s “Living Stereo” orchestral recordings began to push stereo into the mainstream in the 1950s. Pop took a long time to catch up: the Rolling Stones were still recording mainly for mono as late as the “Beggar’s Banquet” album in 1968. CDs were designed for classical music – Sony and Philips agreed that they should by 74, not 60 minutes long so that they could accommodate the entire length of Beethoven’s Ninth Symphony.

Perhaps Apple is scared off by the fact that integrating music streaming with audio hardware is a mess of contradictory protocols. The implementation of Airplay’s main competitor, Google Cast audio, is also very buggy. Speaker companies such as Sonos Inc. exist to overcome these contradictions and investors are already rewarding them: Sonos stock has jumped almost 170 per cent over the past year.

Classical music is demanding on both reproduction and search. But that’s why classical music fans are who you should seek out if you want to figure out the future, as Apple has just done. Streaming, both audio and video, is going to come down to competition over metadata, quality (are all your movies available in 4K?) and discovery, not breadth of catalog or cost.