Time / Phase Alignment / Smaart Setup & Alignment Guide

Smaart Setup & Alignment Guide System Calibration

My own field workflow for calibrating and tuning a PA system with Smaart, start to finish — software setup, mic placement, gain structure, delay alignment, and the EQ pass, repeated position by position across a rig. The gain-structure and delay rules below aren't optional steps to skip when you're in a hurry — skip them and every trace you capture after stops meaning anything relative to the ones before it.

0/ complete

Not saved anywhere — resets when you leave this page.

Key facts

Set input/mic gain once, during the very first measurement of the session — never touch it again once you start moving positions, or every later trace loses its shared reference point.
Mic height is always the ear level of the audience for that show, not floor height — a floor measurement's comb filtering isn't what the crowd actually hears.
Play pink noise (or program material) at the actual show SPL target, not quiet or cranked — the ear's frequency sensitivity shifts with level (Fletcher-Munson) even though the mic stays linear.
Re-find delay every single time the mic moves. A delay value from one position means nothing at the next.
If coherence has dropped to roughly 50% at a frequency, only apply about half the EQ correction the trace suggests — don't chase the full number.
Avoid Zoom mode when reading phase — it hides how many full cycles you're actually off by, which is how good techs end up several cycles out without noticing.

Software Setup

In I-O Configure, activate the input and output devices you need, then rename each individual input/output so it's identifiable at a glance later. In Measurement Config, set up a transfer function pair: Ref is always your source signal (pink noise or console feed), Mea is the mic actually listening to the system. One pair is enough to start. Leave FFT size on MTW (multi time window) — it blends multiple FFT windows automatically to give the best result, no manual tuning needed. Route pink noise (or a wav file) through the signal generator, and confirm both main and auxiliary output levels are set correctly before you start measuring. Learn the core hotkeys before you're on a clock: Spacebar captures the current trace, V resets/redoes the analysis, G toggles pink noise on and off.

Reading the Display

Magnitude window: 0 dB means no difference between reference and measured signal. Negative means missing energy, positive means excess — this is what you're correcting with EQ. Coherence line (100% down to 0%): shows how trustworthy the magnitude reading is at each frequency. It naturally drops in reverberant rooms or wherever comb filtering is heavy — scale back your EQ correction accordingly rather than trusting a low-coherence reading at face value. Phase window: the trace wraps every 360° instead of jumping cleanly from -180 to +180. Learn to read the wrap — a flat-looking line doesn't automatically mean two sources are aligned. Leave the ETC (envelope time curve) option off on the delay finder for day-to-day alignment work — it centers on the whole impulse envelope rather than the first arrival, and skipping it gives a cleaner, more horizontal phase trace to align by. Use the +/- phase-shift controls next to the delay readout to manually rotate the phase trace when it helps you compare it against another source.

Mic Placement

Baseline tops position: measure the distance between the left and right clusters, then place the mic that same distance out, facing away from one side — near the edge of the cabinet, not dead on-axis. For wide arrays, use the average-middle-of-array position instead, or the walk-back trick: step back and look for the air gap between roughly the 5th and 7th cabinet to locate the true middle of the array. For sub-to-top alignment, when the subs aren't on the same vertical line as the tops, place the mic on an imaginary line between the middle of the top array and the center of the sub line. Remember a digital delay sub arc shifts that effective center point. For front fills and outfills, position the mic where the horizontal AND vertical coverage between the two sources is -6dB — that's the point where summing them recovers the lost 6dB and keeps coverage even. Test your own mic's off-axis behavior before trusting it — even 5-10° off-axis can lose a surprising amount of high-frequency information.

Gain & First Measurement

Set the loop-back gain first, letting the signal just kiss the top of healthy headroom. Unmute one source at a time and bring up the mic/input gain to a healthy level against that same show-level pink noise. This is the only gain setting for the whole session — do it once and leave it alone. If the trace doesn't sit on the 0 dB reference line even though gain is set correctly, drag it up or down to apply a measurement offset rather than touching gain again. Run the system check starting from the highest frequencies down — a wiring or polarity mistake shows up immediately there and won't risk damaging the system the way starting low can.

Delay Alignment

Every time you unmute a different zone (subs, a fill, an outfill), hit Find, insert the delay, and store a fresh trace — never carry a previous position's delay value forward. Watch phase direction above the crossover to tell who's early: if the line is heading upward, that source (often the subs) is arriving early relative to the reference. Once you're in a session, reading delay manually — hovering the mouse over the impulse response peak — is the safer method at a given position, rather than re-running the automatic delay finder. If two phase traces still don't overlap after delay is corrected, try flipping polarity on one of them before adding more delay.

EQ Pass

Address the biggest peak first, one filter at a time: set frequency, gain, and Q, confirm the change against the measured trace, then move to the next problem area. Leave narrow suck-outs caused by comb filtering alone — they're a distance/geometry artifact, not something EQ can fix, and chasing them tends to make things worse. For broad correction at the frequency extremes (a hot top end, for example), reach for a shelf/FIR-style filter instead of piling up narrow filters. Once every zone (tops, subs, each fill) is individually aligned and EQ'd, only make small, conservative adjustments on the combined "ALL" trace. If the combined trace needs heavy correction, that's usually a sign to revisit an earlier zone instead.

Working the Whole Rig, Position by Position

Work one zone completely before moving on: unmute only the reference (usually tops) and store a trace, mute it and unmute the new zone to align and EQ it, then bring both back together and store the combined trace before moving to the next zone. Compare each new capture against your earlier saved reference traces (e.g. "tops post EQ") rather than judging a zone in isolation — the goal is a matching system, not a flat curve at one mic position. Move through the rig in the same order each time: tops, then subs (delay-align, then EQ, then touch up the combined bump), then each front fill, then each outfill, then a full-system check.

Working faster, and what to leave alone

Stop the pink noise once you have a stable stored trace — dialing in EQ or delay on a processor doesn't need live signal, and it saves everyone around you the noise.
It's fine to tune against program music instead of pure pink noise; burst pink noise briefly whenever you need a clean transient to nail a specific impulse response.
The averaging feature (blending two traces into one) exists but isn't something to lean on — the goal at each position is the best sound, not an averaged compromise.
Only build a custom mic calibration curve if you find yourself correcting the same frequencies by more than about 6 dB, show after show. Otherwise trust that a decent measurement mic is flat enough to work from directly.

Adapted from Ofer Dvir's Smaart system calibration & tuning course (Udemy). Rewritten in Ryan's own words from the course material.