The system processor's EQ page was so densely populated with filter nodes that it looked like a connect-the-dots puzzle. I counted them. Forty-seven active parametric EQ bands across the left output channel alone. The right channel had fifty-one. No two were the same.

Each band had a story. Some were clearly feedback notches — narrow cuts at specific frequencies, the scar tissue of past shows. Others were broad tonal adjustments applied at some unknown date for some unknown reason. A few appeared to be corrective filters that were correcting other corrective filters. One was a +0.1 dB boost at 847 Hz that couldn't possibly have any audible effect but had survived at least three firmware updates.

The cumulative effect of forty-seven filters on phase response was catastrophic. Each filter introduces phase shift at and around its frequency. Forty-seven of them, at various frequencies, created a phase response that looked like modern abstract art. The group delay varied by 15 milliseconds across the audible band.

I bypassed everything. All forty-seven bands. The raw PA response was ±4 dB from 55 Hz to 18 kHz. Not perfect, but substantially better than the 'corrected' version. The group delay became smooth. The system sounded clearer, punchier, and more coherent than it had in years. Forty-seven corrections, and the system was better without any of them.

The Moral: More EQ is not better EQ. Every filter has a phase cost. SonaVyx's diagnostic engine recommends the minimum corrections needed — because the best-sounding system is usually the one with the fewest filters between the music and the audience.