Tuning a System with Graphic/Paragraphic EQ — Tools

Key facts

EQ is the LAST step in tuning a system — fix broken gear, polarity, and level issues first, or the correction curve is chasing the wrong problem.
Two basic tuning approaches: ring out the system to hunt individual feedback nodes, or measure the frequency response with an RTA and correct toward a target room curve. Real jobs often use both.
Measure from more than one listening position; if the curves broadly agree, average them into one representative measurement before you start correcting.
Expect several rounds — three or four measure/correct passes is typical before the response settles.
Always confirm on real program material, not just pink noise or test tones — and document the final EQ settings alongside the curve they produced.
Default subsonic high-pass filter: 20 Hz cutoff, 18 dB/octave is the sweet spot (12 dB/octave is acceptable, 24 dB/octave is usually more slope than you need and adds extra phase shift).

Pre-EQ Checklist — Don't Skip This

Check every piece of gear in the chain — and the system as a whole — for basic health: frequency response, noise, distortion. One misbehaving loudspeaker or a hum problem will sink your EQ work before it starts. Verify polarity is consistent across every input source (mics, DIs, playback) all the way through the console, processors, amps, and boxes — everything should be in phase with everything else. Bi-amp/tri-amp crossover points are the exception; follow the manufacturer's spec there, since some drivers are wired reversed on purpose. Listen and test at real operating levels with real program material. Test tones run quiet (or cranked to extremes) can hide problems that show up clearly once the system is actually working.

Measure the Room

Pick a measurement tool that matches your EQ's resolution — if you're correcting in third-octave bands, your RTA or analyzer needs at least third-octave resolution to be useful. Take a reading from a typical listening position, then repeat from three or four other spots around the room. Compare the readings. If one or two positions are wildly off from the rest, that's often a coverage or aim problem, not an EQ problem — consider re-aiming or repositioning boxes, or adjusting relative levels, before blaming the EQ. If the curves broadly agree, average them together into one representative response to work from.

Correct, Then Re-Measure — Repeat

Apply correction to lift obvious dips and knock down obvious peaks in the averaged curve. Measure again from the same positions and compare. Expect to go around this loop three or four times before the response settles down. Before calling it done, listen to actual program material through the system — not just noise or tones. If something sounds off despite a clean-looking curve, trust your ears and split the difference between the measurement and what actually sounds right. Write down both the final measured response and the exact EQ settings that produced it — future you (or whoever touches the rack next) will need the record if the settings ever get bumped.

Quick reference — high-pass/low-cut filters

Default: a subsonic high-pass around 20 Hz on everything, to strip out stage rumble, footfall, and handling noise before it eats amplifier headroom for no audible benefit.
18 dB/octave is the sweet spot for slope; 12 dB/octave is a fine compromise; 24 dB/octave is usually overkill and adds more phase shift than it's worth.
Go easy on (or skip) the high-pass on sources that actually live down there — kick and bass channels, or keys patches built around real sub content — since an aggressive cutoff will rob them of the low end you actually want.

Source: Davis, Gary D. & Jones, Ralph. The Sound Reinforcement Handbook, 2nd ed. (Hal Leonard/Yamaha). Section 14.1.8-14.1.9, pp.254-256. Rewritten in Ryan's own words from the source material.

Tuning a System with Graphic/Paragraphic EQ System Tuning

Pre-EQ Checklist — Don't Skip This

Measure the Room

Correct, Then Re-Measure — Repeat