Accuracy assessment

This page documents real test results comparing OnlineSoundMeter against calibrated reference equipment. We want you to know exactly what this tool can and cannot do before you rely on it for any decision.

What we tested

We compared readings from OnlineSoundMeter against a calibrated Class 2 sound level meter (Extech 407730) in controlled environments. The reference meter was calibrated within 6 months of testing using a 94 dB / 1 kHz acoustic calibrator.

Devices tested:

• MacBook Pro 2024 (M3 Pro) — Chrome 126, built-in microphone array
• Samsung Galaxy S24 — Chrome 126, bottom-firing mic
• iPhone 15 — Safari 17.5, bottom mic
• Lenovo ThinkPad X1 Carbon Gen 11 — Chrome 126, built-in mic array

Each measurement was taken three times with a 30-second sampling period. We recorded the minimum and maximum readings observed across all three trials. All devices were positioned within 15 cm of the reference meter, pointed at the same sound source.

Results by environment

How OnlineSoundMeter performed across different noise levels, compared to the Extech 407730 reference:

Key takeaway: In the 40–80 dB range (where most everyday decisions happen), deviation is typically ±3–5 dB. At extremes — very quiet or very loud — accuracy degrades significantly, especially on iOS.

Accuracy by range

Expected accuracy varies depending on the sound level being measured:

The 40–90 dB range is where consumer MEMS microphones operate within their designed sensitivity window. Outside this range, physical hardware limitations dominate regardless of software quality.

What affects accuracy

Multiple factors contribute to deviation between our readings and a calibrated meter:

• Microphone hardware — The single biggest factor. A $0.10 MEMS mic in a budget phone behaves differently from the array in a MacBook Pro. Sensitivity, noise floor, and frequency response vary widely across devices.
• AGC handling by browser — Even when we request autoGainControl disabled, some browsers and OS audio stacks apply gain normalization before the Web Audio API sees the data.
• Proximity to sound source — Sound pressure drops ~6 dB per doubling of distance. Small positioning differences between your device and a reference meter matter.
• Room reflections and acoustics — Hard walls amplify readings by 3–6 dB compared to open-air measurements at the same distance from the source.
• Device orientation and mic port obstruction — Covering the mic with a finger, case, or desk surface can attenuate readings by 5–15 dB.
• Background OS audio processing — Some systems (especially iOS) apply voice processing, compression, or noise gating at the driver level that cannot be disabled from a web page.
• Temperature — Affects microphone sensitivity slightly (typically <1 dB across normal indoor temperature ranges, but relevant for outdoor use).

Relative vs absolute accuracy

This distinction matters more than most people realize:

Absolute accuracy means "does the number match what a calibrated meter reads?" For our tool, absolute accuracy varies by 3–8 dB depending on device and environment. You might see 68 dB when the true level is 63 dB.

Relative accuracy means "if I compare two measurements taken with the same device, is the difference correct?" Here, our tool performs much better — typically consistent within 1–2 dB. If Room A reads 45 dB and Room B reads 62 dB on your device, the actual difference between those rooms is very close to 17 dB, even if the absolute numbers are both off by 4 dB.

For most practical decisions — "is my office quieter than the coffee shop?", "did the new window reduce street noise?", "which room is best for recording?" — relative accuracy is what matters. And relative accuracy is where browser-based measurement actually performs well.

Device ranking

Based on our testing, devices fall into rough tiers for measurement accuracy:

• Best: MacBook Pro (2022+), high-end Android flagships (Pixel 8+, Galaxy S24+). These have high-quality mic arrays and browsers that fully honor audio constraints.
• Good: Mid-range laptops with dedicated mic arrays, recent iPhones via Chrome on a separate device (use the phone mic with Chrome Android for measurement, not Safari iOS).
• Acceptable: Budget Android phones, older laptops (2019+), Chromebooks. Readings are directionally correct but expect ±5–8 dB deviation.
• Poor: Tablets with a single bottom-firing mic, devices older than 2017, any device where you cannot disable OS-level audio processing.

When this tool is sufficient

A browser-based sound meter with ±3–5 dB accuracy is adequate for:

• Comparing noise levels between rooms or locations (relative measurement)
• Checking if your environment is roughly safe for extended work (above/below 70 dB threshold)
• Monitoring background noise trends during the day
• Testing whether soundproofing improvements made a difference
• Identifying the loudest noise source in a room
• Getting a general sense of classroom or office noise levels
• Checking if your music/podcast listening level is in a safe range
• Verifying that a "quiet" appliance is actually quiet relative to alternatives

When you need professional equipment

Do not rely on this tool (or any phone/browser meter) when:

• Making legal noise complaints or documenting violations (courts require calibrated Class 1 or Class 2 meters)
• Workplace OSHA compliance measurements (requires Type 2 SLM minimum, often with dosimeter)
• Acoustic engineering or room treatment design (need calibrated measurement mic + analysis software)
• Determining exact exposure time limits at a specific dB level
• Measuring below 25 dB (studio/anechoic chamber work)
• Measuring above 100 dB (concerts, machinery) where hearing protection decisions depend on accuracy
• Any situation where being wrong by 5 dB has financial, legal, or health consequences
• Building code compliance for sound insulation ratings (STC/IIC testing)