James was a believer in the art of EQ. He spent three hours during soundcheck adjusting the system, guided entirely by his ears. A touch more warmth here. A bit more air there. Tighten the low-mids. Open up the presence region. He tweaked forty-two parameters across the system processor and six outboard units. He pronounced it 'transformed.'

The problem was that I had taken a transfer function measurement before he started. And I took one after. And when I overlaid the two traces on SonaVyx's display, they were — within measurement tolerance — identical. The before and after differed by less than 0.8 dB at any frequency. The two curves lay on top of each other like they were spooning.

James had made forty-two adjustments across six devices, and the sum total effect was indistinguishable from doing nothing. Some of his boosts were cancelled by cuts elsewhere. Some were below audible threshold. Two of his devices were patched into the signal chain but had their processing bypassed. One outboard unit wasn't receiving signal at all — the send cable was disconnected. He had been adjusting its knobs for forty minutes while listening to the system's unchanged output and hearing improvements.

I showed him the overlay. He stared at it for a long time. 'The measurement must be wrong,' he said. The measurement was not wrong.

The Moral: Trust measurements, not impressions. Confirmation bias is powerful — if you expect to hear a change, you will. SonaVyx's before/after comparison provides objective proof of whether your adjustments actually did anything. Sometimes the most valuable thing a measurement tells you is: you're done.