Live Session: Barometric Turn a Storm Into a Set
Barometric booked this session off a forecast. For fifty-one minutes a collapsing low pressure system chose the tempo, opened the filters, and decided when the set was over.
The date was picked by a forecast. Barometric provisionally booked this session eleven days out, moved it twice, and then called on a Sunday evening to say Tuesday was the one: a deep low crossing the coast between seven and nine, with the front arriving more or less over the roof. Which is how you come to record a live set on the least glamorous night of the week, in a building where the fire door has to be held shut with a chair.
The instrument is on the roof
Ines Halloran and Tomas Ek run their weather rig from the top of Northlight Studios, and it is scruffier than the concept deserves. An anemometer zip-tied to a scaffold pole. A pressure sensor in a plastic lunch box gaffered to the handrail. A humidity probe tucked into the lee of the stairwell housing, where Ek admits the reading is "a lie, but a consistent lie." One cable down four flights into a laptop, polled every second and averaged over thirty so the set does not twitch.
The mapping is the same one they described when we put five questions to them in the autumn, with one addition:
- Barometric pressure to filter cutoff - falling pressure opens the machine up
- Relative humidity to delay feedback
- Wind speed to tempo
- New: any gust above 18 metres per second fires a single snare hit, which neither of them can trigger by hand
That last rule is the reason to watch rather than just listen. Twice in the video someone flinches at a snare before it happens, because they heard the gust hit the stairwell door first.
What the storm actually did
At 19:40 the set opens at 1004 hectopascals and 118 beats per minute, and it sounds like a soundcheck, because it is basically a nice evening. By 20:30 the pressure is at 993, gusts are touching 24 metres per second, and the tempo has climbed to 141 without either of them touching anything. The filters are wide open. The delay is drowning.
Then the front goes through, the wind drops to six, and over about four minutes the whole thing decelerates to 96 and stops. Fifty-one minutes. Nobody chose the length, nobody chose the ending, and the ending is better than one they would have written.
It is not quite pure. At 20:12 the feed dies for ninety seconds - water in a connector, entirely predictable and entirely their fault - and they cut to a canned dataset from a 2019 storm they keep on the laptop for exactly this. They volunteered the detail before I asked. "You can hear the seam," Ek says. "Real weather is untidy. The 2019 file has been cleaned up, so for ninety seconds it is too smooth. Once you know, it is the only thing you can hear."
Two questions in the stairwell
You could fake all of this with an LFO and nobody would know.
Halloran: "Nobody would know and we would know, and more to the point we would start choosing. The whole point is that the system is indifferent. It is not building to a drop because you are having a nice time."
What happens if you book a night and the weather is boring?
Ek: "Then it is boring, and about forty minutes in people stop waiting for something to happen and start listening to what is actually there. We have had two of those. One was the best set we have played."
Discussion