Decibel Meter Online: Why It Reads Negative, Plus a 60-Second Room Check

You open a decibel meter in Chrome, sit in a silent bedroom, and it reads -58. Nothing in the room changed.

That's because the meter was never measuring your room. It measures how far your mic's signal sits below the loudest level that mic can record. Know that, and the number turns useful.

Our take: a browser meter is an excellent pass or fail check for your own mic in your own room, and it tells you nothing about hearing safety. The trap is reading one scale as if it were the other. Official hearing limits are sound-pressure numbers in dBA, and a browser never learns how sensitive your mic is, so no uncalibrated laptop can produce them.

What does the number on a browser decibel meter actually mean, and why is it negative?

Sound is measured in decibels, but a decibel is a ratio. It always needs a reference point, and that reference decides what the number means.

0 dBFS is the loudest sample the converter can store. Anything louder clips: the tops of the waveform get sliced flat and you hear harsh distortion.

• dBA is dB SPL with A-weighting, which turns down the low and very high frequencies the ear hears less well.

Can a website measure real dB SPL? Not on its own. The browser receives a stream of numbers from the mic and nothing about how sensitive that mic is.

At the loud end, a shout close to a built-in phone or laptop mic, or a loud concert, can push the signal to 0 dBFS. The meter then stops rising while the real sound keeps getting louder. A browser meter under-reports exactly the sound your ears care about.

The meter on this page shows your mic's peak and average in dBFS and starts with the browser's audio processing switched off. The audio stays on your device.

Why do two online meters, or a phone app, show different readings in the same room, and does Chrome change the number?

When a website asks for your microphone, the browser quietly switches on three filters unless the page asks it not to.

Follow the sound from the room to the meter and you can see where each stage bends it:

1. The room sets the real level.
2. The mic turns pressure into signal. Its sensitivity decides where a sound lands on the dBFS scale.
3. Operating system input gain (Windows: Settings > System > Sound > Input; macOS: System Settings > Sound > Input) raises or lowers everything after it.
4. Browser auto gain pushes a quiet signal up, so a silent room reads louder than it is.
5. Browser noise suppression pulls steady noise down, like a fan or an air conditioner, and can hide your real noise floor.
6. Browser echo cancellation subtracts what your speakers play, which can dip the level in bursts.
7. The call app (Zoom, Microsoft Teams, Google Meet) runs its own suppression on top.

Two meters can disagree just because one turned processing off and the other left it on. Other culprits: which mic the browser picked, whether the meter shows peak or RMS (average power, which reads lower than peaks), and how long it averages.

Try the toggle on our meter: processing on, then off. The gap shows how much processing reshapes the room you think you're measuring.

How do I check my room's noise floor before a video call or recording?

Your noise floor is what the mic picks up when nobody talks: hum, hiss, the fridge, traffic, the fan. It sits under every second of your voice.

1. Close noisy apps and switch off fans and air conditioning you control. A laptop working hard spins its own fan, so close heavy tabs too.
2. Plug in and select the mic you'll actually use, in the browser and in the operating system.
3. Set the input gain where it will stay for the session.
4. Leave browser processing off on the meter.
5. Put the mic in its session spot, aimed at your mouth. With a directional mic, the window, the computer fan and the fridge hum from the kitchen belong behind it, never in front.

Then say one sentence at normal volume and note the peak. The gap between voice peak and noise floor matters more than either number alone. Peaks touching 0 dBFS mean the gain is too high.

The usual mistakes:

• measuring with auto gain on
• measuring with the laptop on your lap instead of on the desk where it will sit
• measuring at 2 a.m. and recording at noon with traffic outside
• changing the gain after you measure, which shifts every number

What dBFS noise floor reading is good enough for a podcast, a voiceover or a Zoom call?

No official dBFS standard exists for a Zoom call.

That applies to the finished file, and raising your voice to ACX loudness raises the noise with it. The wider your raw gap, the easier the target.

Your room reads -48 dBFS with processing off and your test sentence peaks at -10. A 38 dB gap: fine for a Teams call. For a voiceover, closing the window and switching off the desk fan might bring the floor to -57. Move the mic from arm's length to about a hand's width from your mouth, lower the gain until peaks land at -10 again, and the floor reading can fall into the mid -60s. The room got no quieter. The ratio did.

Fixes, cheapest first:

1. Kill the noise source.
2. Move the mic closer and lower the gain.
3. Add soft furnishings (a rug, curtains, a duvet over a clothes rack) to cut echo. They won't stop traffic.
4. Swap the built-in mic for a headset or dynamic mic, which picks up less of the room.

For a longer test, use the Mic or Room bench on our home page.

When is a browser reading useless, and how do I get a real decibel number?

If your question is "is my setup quiet enough?", dBFS answers it.

• a professional noise survey by an acoustic consultant or industrial hygienist for workplace or legal disputes

Teachers: Bouncy Balls and Classroomscreen's sound level widget work well as a visual cue, because you set a relative threshold on one device. They don't measure real loudness and shouldn't be quoted as a hearing-safety reading. For that, you need the numbers below.

What decibel level is safe for your hearing, and why can't a browser meter tell you?

According to the National Institute on Deafness and Other Communication Disorders (NIDCD), sounds at or below 70 dBA are unlikely to cause hearing loss, even after long exposure. Long or repeated exposure at or above 85 dBA can cause it, and the louder the sound, the sooner damage happens.

NIDCD's rule of thumb: avoid noises that are too loud, too close or last too long. A sound can harm when it is too loud even briefly, or when it is loud and long-lasting. Familiar sources include earbuds or headphones at high volume, playing in a band and loud concerts.

A CDC study of 2011-2012 data found at least 10 million US adults under 70 (6%), and perhaps up to 40 million (24%), have hearing test features suggesting noise-related loss. Data from 2005-2006 put it at up to 17% of teens aged 12 to 19.

Human hair cells in the inner ear don't grow back, so most noise-induced hearing loss is permanent. A temporary dulling can disappear within 16 to 48 hours while some damage remains. Loud noise can also cause tinnitus, a ringing, buzzing or roaring in the ears. NIDCD calls noise-induced hearing loss the only completely preventable type.

NIDCD's advice is to wear earplugs or other protective devices, and to move away from noise you can't reduce or protect yourself from. If ringing or muffled hearing doesn't clear, see an audiologist or your doctor for a hearing test. This page is information, not medical advice.