Where the Red Light Was

Every recording happened somewhere. The ledger, honest to a fault, will only say where for a sliver of them: of 36 million dated recordings, the share with a documented recording place peaks at one in ten (the 1980s) and has fallen to one in thirty (the 2020s). So this case is really two investigations wearing one coat — a geography of recording, and a geography of documentation, and the court will keep saying which one is testifying.

Within the documented world, the concentration collapse is the finding. In the placed 1940s, New York alone holds 24 percent of the map — the industry was a city. Decade by decade the summit erodes: London's 1980s lead is 7.9 percent; by the 2020s the leading city on Earth — still London — holds 3.1 percent, with Paris and Berlin behind it, and the rest is a fine mist of ten thousand small rooms. The red light did not move from capital to capital, in the end. It shattered. The forty-minute album is now assembled in places the ledger has no name for, which is to say: everywhere, which is to say — as a matter of filing — nowhere. The placeless share is not proof of bedrooms (absence of paperwork never proves anything, a lesson this collection keeps relearning), but the documented mist and the missing addresses point one way, and it is the way of the laptop.

And then there is the anomaly the census kicks up, the kind of detail an investigator circles twice. In the placed 1950s, the single busiest recording location on the planet is not a New York tower or a Hollywood soundstage. It is Hackensack, New Jersey — 14.3 percent of the documented decade — and in the 1960s the crown moves nine miles up the Palisades to Englewood Cliffs (13.3 percent). Both addresses are one man: Van Gelder Studio, the room of engineer Rudy Van Gelder, first in his family's house, then in the purpose-built studio, 8,197 recordings at the main node and a second node holding two thousand more. One engineer's two rooms outweigh, in the documented record, every city on Earth for two consecutive decades. Part of that is the man — half the jazz canon sat in those rooms — and part is the flashlight: jazz discography is loved with an archivist's love, session by session, address by address, while equally busy rooms in equally busy cities went unrecorded by the record. The ledger's map of music is a map of where the librarians shone the light. Hackensack earned the beam; the beam also made Hackensack.

Which is the honest close for a geography case in this house. The red light was in New York, then briefly and gloriously in New Jersey, then in London and everywhere, then nowhere the file can name. The crews had cities; the addresses had leaderboards; the present tense has a mist. If some future archivist loves the bedroom era the way the jazz people loved Van Gelder's log books, the 2020s map will someday light up street by street, and this article will need a retraction of the happiest kind. The graph will be here. It keeps the light on.

Method: recordings with dated years 1930–2026; "placed" = carries a recorded-at edge (254 placed recordings in the 1930s, half a million in the 2010s — shares are of each decade's dated total). Metro = first named area up the place's containment chain, so adjacent nodes (Manhattan / Midtown / New York) split totals and city figures are conservative floors; the exhibit's New York and London segments sum each decade's listed fragments (boroughs, neighborhoods, St John's Wood, Hampstead) and remain floors. "Van Gelder's New Jersey" = Hackensack + Englewood Cliffs. Documentation bias is the subject, not a caveat: placed shares measure archival attention as much as geography, and are read as such throughout. Computed August 2026 against the 2026-07-25 MusicBrainz build.