One of the most popular articles on sound system EQ is How to Tune a PA System for Live Sound by Brad Pack. In this article I will compare and contrast it against my own methods.
Things I do the same
Start by testing the system with a few high-quality tracks that you know really well.
Yes! Listen first. The first step on my sound system tuning checklist is always Console Check (aka output verification).
Walk around and see what it sounds like up front, way in the back and over at the bar.
Yes, but be careful here. Your echoic memory is only about 20 seconds, and that’s in absence of interference. In other words, don’t expect to accurately compare the orchestra to the balcony.
It’s important to periodically bypass the GEQ and see what it sounded like before you made any changes. It’s really easy to make a mix sound worse by over-correcting, and you’ll never know if you don’t A/B (compare) your settings against the original.
Yes! My biggest takeaway from my interview with Jamie Anderson is to always incorporate strategic stop and check points to listen and make sure you are going in the right direction.
Things I do differently
The key to a good sounding show is a good sounding room, but if you don’t have one of those, you can try to fix a not-so-good sounding room by using a graphic EQ to “tune” the PA system.
While it is true that a good sounding room starts with the architecture, there are two problems with this sentence. First, you can’t fix a room with EQ. That requires a sledge hammer. What you treat with EQ are speaker-to-speaker and speaker-to-room interactions.
Second, the graphic EQ is the wrong tool for the job. The logarithmically spaced fixed frequency and bandwidth filters of the graphic EQ will never be able to create the shapes necessary to treat the frequency response created by the interactions just mentioned.
When using a graphic equalizer, try to think of each band as having a “Q” value of 100 — they’re super-narrow notches.
I used to think the exact same thing, that a GEQ was this super precise instrument. Then, I measured one. The truth is that if your GEQ has a filter bandwidth of ⅓ octave, then its Q is 4.31, not 100. There will be varying degrees of overlap between filters depending on the design.
With all of the mics “live”, slowly turn up the gain on each vocal mic and identify the frequencies that feedback
Find the band for that frequency (or the closest one to it) on the systems graphic equalizer, and slowly lower the fader until the feedback stops.
First, this won’t work with a GEQ because you’ll never be able to select the correct frequency. Does your GEQ have a filter at 839.2Hz? I don’t think so. Instead, you’ll end up damaging the frequency response of your speaker and reducing the system’s gain-before-feedback.
Second, this method will only work if the vocal mics you are using along with the speakers remain in the exact same position through soundcheck and show. A 9kHz sound wave is only 1.5 inches long. Moving the mic 0.75 inches is enough to alter the path length by 180º from a nearby surface and cause a 6dB peak where you previously had a valley.
Instead, I like to insert filters to reduce feedback on microphone groups with a parametric EQ. I send all of my vocal microphones to a vocal group. I follow the steps that Pack recommends, but instead of leaving the filters inserted, I bypass them until soundcheck. Then, during soundcheck, if I can’t get enough gain before feedback, I enable the necessary filters.
To quickly and easily find the most offensive frequencies, use a stagehand or mic stand to point the mic directly at its respective stage monitor.
Take a look at a microphone’s polar pattern. It is not uniform from back to front. Don’t expect to get the same results measuring from the back as you do from the front.

For more on my methods for dealing with microphone feedback see 6 Smart, Proven Methods To Control Feedback Onstage (Without EQ).
Gotta say that I really enjoy this articule, I think your completely right about the thing you do differently and it creates a great combo with the video of measuring EQs, cause if you still think that GEQ have a really high Q you gotta to do that exercise of measuring it out.
Agreed! Seeing/hearing, is believing.