openmixerGitHub →

You are reading the manual for build c1b18be.

Room

Governed by: docs/design/specs/2026-09-08-room-analysis.md

Part of the PA Setup tab, alongside Speakers, on any output that feeds a box — MAIN, a mix, a matrix. It reads the room the way Speakers reads a loudspeaker's arrival, against the SAME measurement microphone, but it keeps the whole curve instead of reducing it to one number: where the room is loud or thin, what bounces back and when, where it rings. Align is a different tab — a channel-level instrument for two microphones on one source, not a PA-setup step.

The PA Setup tab draws a numbered strip above Speakers and Room — Aligned → Positions → Modes confirmed → Proposals → Applied → Re-swept — reading each step straight off the same rows the two panels below it already carry. It is a map of where you are in the procedure, not a control of its own.

This page names what it finds — a curve, a set of reflections, a set of modes, a decay time per band — and, for a confirmed mode, a narrow EQ cut you may apply. A reflection is never corrected here; see below for why. A mode is only ever corrected on your own APPLY gesture, and only while this output's own FBS row is set to live — see "Room modes and mic positions" below.

Room — the transfer curve and the reflections

Pick the measurement microphone the same way you do for Speakers (the picker lives in the same PA Setup tab; Room reads the same pick). Play pink noise through the output under test — one output at a time, exactly Speakers' own rule, because a microphone hears the sum of whatever plays.

The panel draws:

  • The transfer curve — magnitude across the band, with coherence shaded underneath it. A flat line means the room is not colouring that range; a shaded band that drops out means the reading there is not to be trusted yet.
  • Reflections — every distinct echo the desk can pick out of the first 80 milliseconds after the direct sound, each as time after the direct arrival (milliseconds), how far that is past the direct path (metres), and how loud it is relative to the direct sound (dB). A reflection is never "fixed" here — see below.
  • Reverberation time (RT60), one small bar per octave band (63 Hz to 8 kHz), read from the last take you recorded of this microphone. A room usually rings longer at the low end than the top, which is why one broadband number would hide the thing worth knowing.

The measurement runs because the panel is open. Opening it arms the desk to measure this output against the microphone and no other; closing it stops. Nothing is measured that nobody is looking at.

Why nothing here is "fixed" automatically

A reflection is a second arrival in time, not a level problem at one frequency. An EQ cut that flattens the notch it leaves in the curve does nothing for a seat two metres away — the notch moves with where the microphone stood, the EQ band does not. The honest fix for a reflection is acoustic: absorption, diffusion, or re-aiming the surface or the box. This instrument names the reflection and stops there.

Room modes and mic positions

A room mode is a standing wave — a property of the room's shape, not of where the microphone happened to be. The tell is that its frequency does not move when the microphone does, while a reflection's comb pattern moves with every position.

So confirming a mode takes more than one reading: measure, press Measure this position, move the microphone, measure again. The desk compares the low end of each position's curve and keeps only the peaks that show up, within a fraction of a bin, at every position measured. Two positions is the floor, never the ceiling — keep adding positions until the confirmed set stops changing, which the panel reports as enough.

Each confirmed mode is shown with its frequency, how far above the surrounding curve it reads, and how many positions confirmed it.

Clear resets the position ledger for this output. It is explicit only — the ledger never ages out on its own, because a sweep you stepped away from mid-room is still the sweep you will come back to.

Proposed cuts and APPLY

Each confirmed mode with real excess energy over the room's own floor lists as a proposed narrow cut: the frequency, the depth, the confirming position count, and its own Apply button — never a bulk "apply all". Applying spends the SAME depth ceiling and gate rails FBS's own feedback notches use; a proposal that would have gone deeper says so ("bounded") rather than quietly offering less than the room asked for.

Apply only lands while this output's own FBS row reads Live. Room Analysis carries no switch of its own for this — it reads the FBS panel's own mode. In Off or Ring-out the row refuses the write and the row explains why; the proposal stays on screen either way, so you can see what would be spent before arming FBS live.

Applying the same proposal twice does not plant a second band — it retunes the one already there. Clear cuts retires every band this instrument has planted on this output, and nothing else: an operator's own EQ bands and any FBS or HRP bands are untouched.

Once applied, keep the tab open (or reopen it) and the same reading keeps running — that is the re-sweep, no separate button for it. A residual figure appears beside the applied proposal once the fresh curve has a reading at that frequency: how much of the mode is left after the cut, so you can see what it bought.

Related

  • Alignment — Speakers, the PA Setup tab's other half, this reuses the same measurement microphone and safety rule from, and the loudspeaker delay this desk offers to fix arrival time; also where the tab's numbered procedure strip is described.
  • EQ and dynamics — where an applied room-mode cut lands, tagged the same way an FBS or HRP band is.
  • Recording — the take RT60 is read from.