Mic Level Meter: Free Online Sound Level Meter for Your Mic
Overview
Mic Level Meter. The Mic Level Meter gives you a focused live input level meter with peak hold and a target band, so you can set your microphone gain correctly. Watch the meter while you speak and adjust your input until it sits inside the target band — clipping and too-quiet input are the two most common call-audio problems, and this catches both. Try the free midi test for a quick, no-install way to check this yourself.
Before you hit record, join a call or go live, it pays to know how hot your signal really is. This mic level meter listens to your microphone in real time and turns the incoming audio into a moving level in dB, so you can see the average, the loudest moments and any overload at a glance. It works in your browser like a pocket acoustic sound level meter, but its real job is to show how close your mic is to its digital ceiling, and, once calibrated, roughly how loud the room is.
How the Mic Level Meter Measures Your Microphone
Every sound level meter does the same three things: it captures pressure changes with a mic, converts them into numbers, and puts those numbers on a logarithmic scale measured in decibels. This online sound level meter uses your device microphone through the web audio interface, and all processing happens locally on your computer or phone. Nothing is recorded or uploaded; the page only keeps a rolling history of levels to draw the graph. Use the airpods mic test tool whenever you want a fast, repeatable way to confirm this.
Because the dB is a ratio rather than a fixed unit, every reading needs a reference: for a digital signal it is the loudest value the converter can store, for sound in air it is a tiny standard pressure.
From Air Pressure to a Digital Signal
Sound waves are small, fast changes in air pressure. Inside a mic, a thin diaphragm moves with them and produces an electrical signal. A condenser microphone is the usual choice for a measurement microphone because its diaphragm is light, stable and responds evenly across the audible range. Laptop and phone mics are much smaller, so their output for the same sound pressure varies from model to model.
That output per unit of pressure is called microphone sensitivity. A professional sound level meter knows the sensitivity of its capsule exactly, which is how it converts voltage back to pressure. A web-based meter does not, so its raw numbers describe the digital signal your sound card delivers, not the room itself.
Peak, RMS and LUFS Readings
Each block of samples gives the meter two numbers: the highest absolute sample value, and the root-mean-square amplitude, which reflects the energy of the signal and tracks what you hear far better than single spikes. Both use the same formula, where \(A\) is the amplitude as a fraction of the maximum:
$$L = 20 \log_{10}(A)$$Here is a worked example. You speak into a USB mic and the meter finds a root-mean-square amplitude of 0.0821 and a highest sample of 0.412. The average level is \(20 \log_{10}(0.0821) = -21.7\) dB and the top of the waveform sits at \(20 \log_{10}(0.412) = -7.7\) dB. The 14.0 dB gap between them is the crest factor of your voice, and the 7.7 dB left above the loudest sample is your headroom.
Alongside the live value, the meter keeps running statistics:
- Min and Max: the quietest and loudest averaged sound level since you pressed start, which show the noise floor and the loudest phrase.
- Avg: an energy average over the session, the same idea as the Leq or LAeq figure on a professional sound level meter.
- Peak level with peak hold: the highest sample, frozen on screen for a moment so you can read it.
This meter reports the loudest sample and an energy average, not LUFS, the broadband program-level measure broadcast tools use across roughly 20 Hz to 20 kHz (Momentary 400 ms and Short-Term 3 s windows). For steady speech the two usually land within a few dB, so the −21.7 dB average above is a fair first guide when a platform quotes a program-level target.
Frequency Weighting: dBA, dBC and dBZ
Human hearing is less sensitive to deep bass and very high treble at moderate levels. Frequency weighting filters the signal so the number follows that response:
- A-weighting (dBA) follows the equal-loudness contours for quiet sounds and is the legal standard for workplace noise.
- C-weighting (dBC) is nearly flat and keeps more low-frequency noise, so it suits loud music and peak noise checks.
- Z-weighting (dBZ) applies no filter, which is what you want when judging your raw mic signal.
By default this meter reads unweighted (dBZ); its A and C weighting options are approximate filters, not a calibrated sound level meter. Unweighted, traffic rumble can read several dB above an A-weighted meter, because that noise sits in the bass. A-weighting is closer to perceived loudness, but not a model of it.
Time Weighting: Fast, Slow and Impulse
Time weighting controls how quickly the reading rises and falls. A jumpy display is hard to read, so sound level meters smooth it:
- Fast: a 125 ms time constant that follows syllables and short noise bursts.
- Slow: a 1 s time constant that steadies a fluctuating sound level such as traffic or an air conditioner.
- Impulse: a quick rise and a very slow fall, kept on some meters for impulse noise like claps but dropped from the current standard.
The display update rate is a separate setting: it only changes how often the screen redraws, not how the level is averaged.
Reading Your Audio Level: Digital vs Acoustic Scale
The most common mistake with any online sound level meter is treating its number as the level of the room. Out of the box it shows a digital audio level, ideal for relative comparisons (louder than a minute ago, mic A against mic B) but silent on the absolute noise around you. Use the free hear yourself live monitoring whenever you want a fast, repeatable way to confirm this.
What dBFS Measures and Where Clipping Starts
On the digital scale, readings are decibels relative to full scale, written dBFS. The largest value the converter can store is 0 dBFS, so every real reading is negative: −21.7 dB means the signal sits 21.7 dB below that ceiling. Push past it and the waveform is flattened into a harsh, crackling distortion you cannot remove later. The clipping indicator lights when samples hit the ceiling; if it flashes, lower the gain on your interface or move back from the mic.
Sound Pressure Level and dB SPL
Sound pressure level measures the actual pressure wave in the air against a reference of 20 micropascals, roughly the quietest sound a young listener can hear:
$$L_p = 20 \log_{10}\left(\frac{p}{p_0}\right), \quad p_0 = 20\ \mu\text{Pa}$$A dedicated sound pressure level meter reports it as a positive number: about 30 for a quiet bedroom, 60 for conversation, 85 for heavy traffic nearby. A phone or laptop sound level meter can only estimate it after calibration.
Target Readings for Voice, Music and Streaming
Set the input volume so normal speech lands in the middle of the band, and check that background noise stays well below it when you are silent.
| Use | Average sound level (digital) | Loudest moments | Silence / noise floor |
|---|---|---|---|
| Podcast or voice-over | −24 to −18 dB | −10 to −6 dB | below −60 dB |
| Video call or meeting | −26 to −18 dB | −12 to −6 dB | below −55 dB |
| Singing or instruments | −20 to −14 dB | −8 to −4 dB | below −60 dB |
| Live streaming with music | −22 to −16 dB | −6 to −3 dB | below −55 dB |
The worked example, −21.7 dB average with 7.7 dB of headroom, sits inside the podcast row.
Passing an Audiobook Spec with the Decibel Meter
A narrator in a converted spare room needs to know one thing before recording a 9-chapter audiobook: will the raw takes pass the ACX submission check, which asks for an average level between −23 and −18 dB, loudest samples no higher than −3 dB, and a room-tone floor of −60 dB or lower? The short answer from the first test is no, and turning the mic up would not fix it.
The setup is a cardioid USB mic on a desk arm, 30 cm from the mouth, interface knob at about two o'clock. The narrator records 10 seconds of silence, then reads a 40-second test paragraph, watching the level reading this mic meter gives in its unweighted mode. The figures:
- Room tone: −57.2 dB, 2.8 dB too loud for ACX.
- Average while reading: −25.6 dB, 2.6 dB too quiet.
- Loudest sample: −11.4 dB, safely under −3 dB.
The tempting move is to raise the input level by 3 dB. Here is why that fails: the knob lifts voice and room tone together, so the gap between them stays at −25.6 − (−57.2) = 31.6 dB. ACX needs at least −23 − (−60) = 37 dB between the quietest allowed average and the highest allowed floor. No knob position closes a 5.4 dB shortfall.
So the narrator changes the two things that move voice and room separately. Halving the mic distance to 15 cm adds about 20 × log10(2) = 6.0 dB to the voice and nothing to the room. Switching off the ceiling vent drops the floor to −63.8 dB on a fresh 10-second recording.
| Setup | Average | Loudest | Room tone | ACX result |
|---|---|---|---|---|
| 30 cm, vent on | −25.6 dB | −11.4 dB | −57.2 dB | fails average and floor |
| 15 cm, vent on | −19.6 dB | −5.4 dB | −57.2 dB | fails floor |
| 15 cm, vent off | −19.6 dB | −5.4 dB | −63.8 dB | passes, 44.2 dB gap |
I would watch the loudest-sample column most closely from here. At 15 cm a hard "p" can jump 3 to 4 dB above the test paragraph's −5.4 dB and cross the −3 dB line, so a pop filter goes on before chapter one. Also, a 40-second reading is not a chapter: ACX measures the whole file, so the narrator re-checks the first finished chapter in an editor before recording the rest.
Calibrating a Noise Meter to Real Decibels
An uncalibrated sound level meter on a laptop can tell you the dishwasher is 6 dB louder than the fridge, but not whether either is 45 or 65 dB. To turn it into a usable noise meter you add a calibration offset, a fixed number of dB that maps the digital reading onto the acoustic scale. Calibration improves accuracy for your setup; it does not make a phone a certified instrument.
Using a Reference Meter or Acoustic Calibrator
Place a trusted reference meter next to your device, play a steady sound such as pink noise or a 1kHz tone, and note both readings. The offset is the difference:
$$k = L_{\text{reference}} - L_{\text{digital}}$$Say the handheld reads 68.4 dBA while this meter shows −41.7 dB. The offset is \(68.4 - (-41.7) = 110.1\) dB. Enter it, and every later reading converts as \(L \approx L_{\text{digital}} + 110.1\). A blender reading −27.3 dB on the same laptop now displays an estimated 82.8 dB(A).
An acoustic calibrator does the same job more precisely: it seals over a measurement capsule and produces a known tone, usually 94 dB. Keep the source and mic at the same distance and angle every time, and note the offset for each device.
Why the Offset Only Holds for One Setup
Built-in mics have an uneven frequency response, so an offset found with pink noise can be several dB off for a bass-heavy sound. Changing the OS input level, the mic position or even the web app can shift it. Professional instruments are verified to a published tolerance; yours is calibrated to one room, one sound and one day, so recheck it whenever the setup changes.
Sound Level Meters: Phone Apps vs Class 1 and Class 2 Instruments
Hand-held sound level meters are built to IEC 61672, which defines their weightings, time constants and accuracy. In the United States, OSHA and ANSI set related rules for workplace measurements. A web-based sound level meter borrows the vocabulary but none of the guarantees.
Classes, Tolerances and IEC Standards
IEC 61672 splits instruments into two accuracy grades:
- Class 1: tight tolerances over a wide frequency range, used for legal, environmental and noise pollution surveys.
- Class 2: slightly looser and cheaper, enough for most industrial noise checks.
Both classes of sound level meters must pass periodic verification against IEC 61672 test procedures. No phone or laptop meets either class.
Smartphone Applications and Browser Meters
Smartphone applications and web tools such as a sound decibel meter page are excellent for quick checks: setting up a mic or comparing two rooms. Treat their absolute numbers as estimates within a few dB at best and use them to spot changes. If a reading matters legally, hire a certified sound level meter conforming to IEC 61672.
Exposure, Dose and Hearing Safety
Loud sound damages hearing through both level and time. NIOSH recommends a limit of 85 dBA averaged over 8 hours, and halves the allowed time for every 3 dB increase:
$$T = \frac{8}{2^{(L - 85)/3}}\ \text{hours}$$At the blender's estimated 82.8 dB(A), that allows about 13.3 hours, so short bursts are harmless. Workplace dose meters integrate this exposure over a shift for health and safety records.
A calibrated sound level meter on your phone can warn you that a space is risky: set its warning thresholds near 80 and 85, and if the sound level stays above that, reduce exposure, since noise-induced hearing loss is permanent.
Spectrum View and the Real Time Analyzer
The spectrum graph splits the signal into frequency bands, the same idea as the RTA used for tuning rooms and PA systems. Professional analyzers work in octave or 1/3 octave bands. On a mic check, the graph shows hum at 50 or 60 Hz, hiss in the top bands, or a muffled mic missing everything above a few kilohertz.
Troubleshooting a Loudness Meter That Reads Wrong
When the meter shows nothing, sits stuck low or jumps around, the cause is almost always permissions, the wrong input device or system audio processing.
Microphone Permissions and Device Selection
- Allow microphone permissions when the page asks, then reload it.
- Pick the right input in your system sound settings; laptops often default to a webcam mic.
- Close other apps that hold the mic, such as meeting software, then restart the meter.
- If the sound level is too low, raise the OS input level or move closer, then re-check the Max value.
Turn Off AGC, Noise Suppression and Echo Cancellation
Operating systems and meeting apps add audio enhancements that rewrite the signal before the meter sees it:
- AGC (automatic gain control) lifts quiet passages and pulls loud ones down, so Min and Max drift toward each other.
- Noise suppression gates out steady background noise, making a noisy room look quieter than it is.
- Echo cancellation can mute your mic while speakers play, causing sudden drops in noise level.
Switch these off in your OS sound settings or mic driver for any reading you plan to rely on, keep the mic in a fixed position, and repeat each measurement twice.
Frequently Asked Questions
- Does the mic level meter send my audio to a server?
- No. The audio is analysed on your own device while the meter runs. Nothing is recorded or uploaded, and the page keeps only a rolling history of levels so it can draw the graph.
- What is the difference between dBFS and dB SPL?
- dBFS is the digital level of your mic signal relative to the loudest value the converter can store, so 0 dBFS is the clipping point and normal readings are negative. dB SPL is the pressure of the sound in the air, which only a calibrated instrument can measure. This meter shows dBFS until you add a calibration offset.
- Is the meter calibrated?
- Not by default, so its readings are best for comparing one sound or mic with another. To get an estimated dB SPL, hold a trusted reference meter next to your mic on a steady sound, or use an acoustic calibrator at 94 dB, and enter both readings. The offset only holds for that device, mic position and settings.
- What should my mic level be for recording or calls?
- For speech, aim for an average of roughly -24 to -18 dBFS with the loudest moments below -6 dBFS, and keep your silent room tone below about -55 to -60 dBFS. If the clipping indicator lights, turn the input down or move back from the mic.
- Which frequency weighting should I use?
- Use dBZ (no weighting) to judge your raw mic signal, dBA for noise and hearing-safety checks because it follows how people hear at moderate levels, and dBC for loud music or low-frequency noise. LUFS Momentary and Short-Term suit broadcast and streaming loudness targets.
- Why are my readings too low, too high or jumping around?
- Check that the page has microphone permission and that the right input is selected in your system settings, then close other apps using the mic. Turn off automatic gain control, noise suppression and echo cancellation, since they change the signal before the meter sees it, and keep the mic at the same distance each time.
- Can a phone or laptop replace a Class 1 or Class 2 sound level meter?
- No. Certified meters are verified to IEC 61672 tolerances as a complete system. A phone or laptop meter is fine for quick checks and comparisons, but use a certified instrument whenever a reading has legal or workplace-safety consequences.