Sound Meter

Validation Protocol & Data

“Accuracy” should be reviewable. This page defines a repeatable validation protocol, publishes data templates, and explains how to reproduce checks and submit verification runs.

What this validation is (and isn’t)

The goal is to measure repeatability and tracking under controlled setup, and to document how results vary by device and browser. It is not a claim of certification to IEC 61672 or legal-grade compliance.

For definitions, limitations, and calibration guidance, see Accuracy & Calibration.

Protocol (SOP)

Setup

  • Use a steady source: tone (1 kHz), pink noise, or stable fan noise.
  • Mark distance (e.g., 1 m) and keep microphone height consistent.
  • Avoid corners and reflective surfaces for room-average checks.
  • Disable audio enhancements if possible (OS-level noise suppression / AGC).

Runs

  1. Baseline run: 60 seconds avg/min/max.
  2. Repeat baseline 3 times to measure repeatability.
  3. Change one variable (distance or door open) and repeat 60 seconds.
  4. If a reference meter is available: record its average in parallel.
  5. If you use Align reading: record the offset value and the reference used.

Download templates

Device & browser matrix

List what you tested so results are comparable.

device-matrix-template.csv

Validation runs template

Record runs (avg/min/max + context + optional reference).

validation-runs-template.csv

How to reproduce using our tools

  1. Open Online Sound Meter.
  2. Optionally open Speaker Test (tones) in a second tab to generate a stable 1 kHz tone.
  3. Measure 60 seconds per run and save a report after each run for an artifact.
  4. Fill in the templates with run averages and context notes.
  5. Keep the same distance and placement across runs and record any changes explicitly.

Practical measurement workflow: Measure decibels more reliably.

Submission checklist (for verifiable evidence)

  • At least 3 baseline runs (60 seconds each) with the same setup
  • At least 2 verification runs with one variable changed (distance or context)
  • Saved report images (one per run) or screenshots including system clock
  • Device + browser + OS listed in the device matrix template

We will only summarize results that include enough context to be reproduced or audited.