A note from the bench

Iver Sandholm, on why Orrery exists

On a winter evening in 2011, a reverb I loved stopped working, and I found there was no one left to call.

It was a Halvyn R-8000 — a British studio reverb from the early nineties, the kind of box that cost as much as a car and earned it. I had mixed through it for years. That night it powered up the way it always had: the amber display lit, the fans turned, the meters twitched. But when I sent it a signal, nothing came back. The room it made — the deep, grained, faintly metallic space that was the entire reason I owned it — simply wasn't there. The machine was on. The instrument was gone.

I did what anyone would. I called the distributor, and the number rang out to a company that no longer existed. I looked for a service manual, and there wasn't one, because Halvyn had been bought, shut, and quietly dismantled three years earlier, and everything it had known — the schematics, the calibration notes, the reasons behind a hundred small decisions — had gone into the ground with it. The forums offered sympathy and no answers. I was holding a beautiful object that had become, in the space of an evening, unknowable.

So I opened it. I spent the better part of a year with a probe and a notebook, reverse-engineering a circuit no one had documented, redrawing its schematic by hand, chasing an obsolete part across three continents to replace the one that had failed. It was slow, and often stupid, and I am aware how it sounds — a grown man who could not throw away a reverb. But I remember the evening it came back. I sent it a signal, and the room opened up again, exactly as it had been: deep, grained, a little metallic in the tail. I sat inside it for a long time.

And then I understood something that has organized my life since. What had happened to my reverb was going to happen to all of them. An entire generation of instruments — the machines that made the sound of the eighties and nineties — was dying, one unit at a time, in silence. The companies had folded. The engineers had scattered into other industries or into retirement. The knowledge that made these things sing was not being stolen or smashed; it was simply being forgotten, quietly, for lack of anyone bothering to write it down. That is a worse kind of loss than destruction. Destruction, at least, gets noticed.

I am a digital-audio engineer by trade. I have spent my career writing the software that, if I am honest, is the very thing that killed these machines. When a reverb like my Halvyn became a plugin for two hundred dollars, the boutique that built the original could not survive it. The same story, with different details, ended all four of the makers I now keep. A Finnish company made the voice of the early mobile phone, and software codecs made its hardware obsolete. An American company built the instruments that measured digital audio, and measurement moved onto a PC and the instruments went dark. A Japanese company sold the dream of digital media, and the MP3 erased its formats in the space of two or three years. And a British company built the best boxes in the room and was undone, as I was reminded that winter, by a plugin.

Here is the thought I could not put down. Software ended them. Software is what I do. I could bring them back with the exact thing that erased them — not to undo the irony, but to honour it. There is something almost redemptive in that: the medium that made these instruments obsolete is the one that can keep them running.

That is what Orrery does, and I want to be precise about the word restore, because it is the whole discipline. We restore. We do not improve.

The temptation, always, is to fix. To add the bits the converter never had. To clean up the noise, extend the bandwidth, smooth the artifact, make the old thing behave like a new one. I refuse this completely, because that temptation is identical to the temptation to erase what made the instrument itself. The grit, the warble, the metallic ring, the way a codec half-swallows a voice — these are not defects to be corrected. They are fingerprints. A person who restores a painting does not repaint it in brighter colours, and a person who restores an instrument does not redesign it. Our task is to remove nothing and add nothing, and to be honest about how hard that restraint actually is.

And every restoration begins with a real object. We do not model a memory, or an average, or an idea of how a machine ought to have sounded. We find a specific surviving unit — often the hard part, sometimes the only part — measure it, and model that, down to the flaws that particular unit happened to carry. Then we print its serial on the plugin, so the claim is checkable. You are never holding a guess. You are holding the thing itself, made to run again.

I did not choose these four makers at random. I used all of them. I measured with a Tekura analyzer and tracked to a Yumeo deck; the telephone in my pocket spoke, for years, through a Salo codec; and I mixed, of course, through the Halvyn. As each company folded I started buying its gear at liquidation — then its manuals, and then, as estates settled and dormant patents changed hands, its actual archives: the design files, the calibration records, the last reference units. Orrery is the formalization of an obsession I did not plan to have.

The R-8000 that started all of this is the first thing in our catalogue. It is numbered OR-0001, it is not for sale, and it sits on a shelf here in Copenhagen, still running. Everything else we make follows from it — from one winter evening, and a room I was not willing to lose.

These instruments were made by people who cared a great deal, and then they were left behind. We keep them running. That is the entire idea, and it is enough.

— Iver Sandholm

Copenhagen