The first measurement showed a smooth response with a gentle roll-off above 10 kHz. Good. The second measurement, taken thirty seconds later at the same position, showed a 6 dB notch at 4 kHz. Concerning. The third measurement showed neither — instead, a broad bump at 2 kHz that wasn't in the first two. I began to question my equipment, my training, and my fundamental understanding of acoustics.

Five measurements, five different results. The coherence was adequate — above 0.75 — so the measurements weren't noise. They were real responses. But how could the same room, same speakers, same microphone position produce five different frequency responses in two minutes?

The answer walked past the microphone during measurement number six. Literally walked past. A person crossed between the speaker and the measurement mic, creating a momentary shadow and a reflection from their body. I looked around the room. It was a rehearsal space. The band was setting up. People were crossing the floor continuously, carrying cables, adjusting stands, having conversations. The acoustic path between the speaker and my microphone was being constantly modified by human-shaped obstacles.

A human body absorbs and diffracts mid and high frequencies significantly. A person standing near the measurement path can create 3-10 dB of variation depending on frequency. I asked everyone to leave the room. The next five measurements agreed within ±0.5 dB.

The Moral: Clear the measurement path. SonaVyx's averaging modes (linear, exponential, peak hold) help smooth out transient variations — but the best practice is to measure in a still room, because physics includes the people in it.