The client arrived at the noise assessment meeting with absolute confidence and a free app on his phone. 'I've been measuring,' he announced, placing his phone on the conference table. 'Your numbers are wrong.'

His phone app showed the venue noise at 72 dBA. My Class 1 meter, calibrated that morning with a traceable calibrator, showed 94 dBA. A 22 dB discrepancy is not a rounding error — it's the difference between a quiet office and a rock concert.

The client's phone, a three-year-old model with a microphone designed to pick up human speech at arm's length, had a frequency response that rolled off a cliff below 200 Hz and above 8 kHz. It had no A-weighting filter — the app applied one in software using the phone's uncalibrated signal. The phone's AGC (automatic gain control) was compressing the signal before the app even saw it, and the microphone's maximum SPL was approximately 85 dBA before it clipped.

'We were next to a construction site,' the client continued, 'and the app said 75 dBA.' I looked at the photo he had taken. A diesel excavator was operating approximately eight meters away. That would produce roughly 95-100 dBA at that distance. His phone, physically unable to capture levels above 85 without clipping, had faithfully reported a number that bore no relationship to reality.

The Moral: Phone SPL apps without external calibrated microphones are entertainment, not measurement. SonaVyx's calibrated SPL meter with known microphone compensation provides measurement-grade accuracy — because science conducted with uncalibrated instruments isn't science, it's guessing.