Convolution IRs from a concrete stairwell

The building I work in has a fire stairwell that runs seven flights, end to end, bare concrete walls, painted steel railings, a steel door at top and bottom. I have wanted to capture impulse responses of it for two years. Last Tuesday at 03:00 the building was empty and I finally did.

The mic position is the recording

A stairwell is not a room with one impulse response. It is at least three:

• Source and mic at the same landing — the close field. Mostly direct sound, a single flutter echo between the walls, ~17ms tail.
• Source on the landing, mic two flights up — a ringing tube. The flutter becomes a pitched resonance at ~84Hz, the standing wave between the two end-doors.
• Source on the landing, mic at the top — a full reverb tail. RT60 of 4.8 seconds at 1kHz, 6.2 at 200Hz, 2.1 at 8kHz. Schroeder frequency around 240Hz.

The third position is the one that sounds like a reverb. The first two are more useful, because they have early reflections that are not generic.

Sweep, not pop

The mic was a pair of Earthworks QTC-50s in ORTF. The QTC-50s are omni, which means ORTF is doing nothing useful for the directional pickup — but the 17cm spacing still produces a stereo IR with a real ITD, which is what I want when I convolve a mono source through it later.

What I will use it for

Not as a reverb. The tail is not interesting on its own — it is generic. The early reflections in position one and two are the interesting part, and they are short enough that you can use them as a kind of pre-delay-shaped ambience: take the first 80ms of the IR, splice in a different tail (a hall, a plate, a different stairwell), and you get something that sounds like a real space without the seven seconds of low-mid mud that the actual room has.

The 84Hz resonance from position two is also unexpectedly useful as a sub bass texture. EQ it down 12dB and feed a kick drum through it and you get the body of a much larger drum than the one you recorded.