Capturing clean field recordings: levels, monitoring and stereo

For a live coder, the sample library is the instrument — and nothing personalises a set faster than material nobody else has. This module builds toward one whole task: walk out with a recorder, capture a location in stereo on purpose (not by accident), prove the take survives the mono PA and phone speakers your audience actually hears, and then fold it back into your rig as sampler fodder for a sketch.

The arc starts with the two non-negotiable field disciplines: setting levels slightly low because digital clipping has no second take, and monitoring on closed headphones because your bare ears never hear the wind rumble and handling noise the mic does. Drill these until they are reflexes — they recur in every capture you will ever make. From there you move to a supported design exercise: compare the coincident X-Y approach (rock-solid mono summing from capsules at one point) against near-coincident arrays like ORTF (spaciousness bought with arrival-time cues), and weigh a dedicated stereo mic against a flexible two-mic bar. The 3:1 rule and the comb-filtering account of why phase damage shows up worst in mono give you the vocabulary to predict and diagnose problems before committing to the take.

Every required atom gates the capstone directly: you cannot choose an array deliberately — coincident, near-coincident or two-mic — verify mono compatibility, or keep the take usable without them; and the raw-material reframing is what turns the recording into a sketch instead of a document. The supporting atoms enrich rather than gate: mic-movement discipline saves you from nauseating image swings, M-S decoding previews post-capture width control, and schizophonia gives the whole practice its conceptual weight — every sample is a sound severed from its source.

Runnable examples

Generated from the context/ instrument corpus by concept (redistributable idioms only). Do not edit — regenerate with gen-module-examples.mjs.

gain-staging

(saw 110) * dbamp (-12) >> audio

_{punctual-0010 · CC0-1.0}

SinOsc s => Dyno d => dac; d.limit();

_{chuck-0028 · MIT}