Start with a live microphone check, then use practical tests for your camera, display, calls, and peripherals. Device permission is requested only after you start the relevant test.
Run the test for 5–10 seconds while speaking at normal call volume. MeterSee will retain the peak, noise floor, and clip count when you stop.
Browser readings are diagnostic estimates, not calibrated laboratory measurements. dBFS describes digital signal level, not acoustic loudness in dB SPL. Audio is analyzed in this tab and is not uploaded or recorded.
Choose a device or start with a symptom.
Browser tests help narrow down a problem. They cannot certify hardware, scan installed software or replace an in-app call test.
Read our measurement methods →Actual browser readings, with unavailable values clearly marked.
Know what the result can—and cannot—tell you.
Change one thing, repeat the test and compare before replacing hardware.
MeterSee does not ask you to judge a microphone from a single moving bar. The browser records a short local sample only after you press Start, then reports several browser-observed measurements together. Comparing those readings helps separate a permission or routing problem from low input level, clipping, and steady background noise.
Close meeting apps that may already be using the microphone, connect the device you intend to use, and choose a quiet room. Press Start microphone test and approve access for this site. If the expected device is not selected, open Test details and choose it before sampling again. Speak at the distance and volume you would use in a call. Headphones help prevent speaker sound from reaching the microphone, but they are not required.
Level is reported in dBFS, where 0 dBFS is the digital ceiling rather than a target. Repeated peaks at the ceiling indicate clipping; very low speech may indicate the wrong input or insufficient gain. The noise-floor estimate describes the quieter parts of the captured sample, so fans, typing, traffic, and automatic processing can affect it. Sample rate, channel count, echo cancellation, noise suppression, and automatic gain control come from the browser or media track when those values are exposed. They describe the captured stream, not the microphone’s advertised specification.
Use the result as a screening check, not a laboratory certificate. If permission is granted but no useful signal appears, confirm the active input in the operating system and meeting app. If speech clips, lower input gain or move slightly farther away. If the background reading is high, mute nearby sources and repeat under the same conditions. Change one variable at a time and compare a second device or browser before replacing hardware. Stop ends the current track; audio and results stay in the browser and are not uploaded to MeterSee.
Step-by-step instructions to resolve device, audio, video, and performance issues.
Separate permission, input selection, audio gain, and app-routing problems step by step.
Open guide →Video & CameraCheck privacy permissions, device conflicts, camera selection, and desktop client state.
Open guide →System PerformanceInterpret Windows, macOS, and Chromebook memory pressure signals without installing cleaner tools.
Open guide →Wireless AudioTest speech arrival, battery status, Bluetooth routing, and left/right microphone selection.
Open guide →Display & RefreshCheck Windows Advanced Display, cable version, port bandwidth, and GPU driver capabilities.
Open guide →Gaming HardwareMeasure neutral axis values, dead zones, and identify hardware drift versus game sensitivity.
Open guide →A moving meter only proves that some signal reached the page. Use the pattern below to isolate the layer before changing equipment.
Confirm the active device name, then check the operating-system input meter.
Lower hardware or OS input gain until speech peaks remain below −6 dBFS.
Mute nearby sources and compare with a second input before replacing the microphone.
The test reads browser-delivered audio samples and calculates RMS level in dBFS: 20 × log10(RMS). The displayed floor is −96 dBFS. A low percentile of valid samples estimates the background during this sample; it is not a separate calibrated noise measurement.
These levels include microphone gain and browser processing. They are not room loudness in dB SPL. Speak normally, pause briefly, and repeat after changing one input setting.
Change one variable at a time, repeat under the same conditions, and compare with the operating system or the affected application. An unavailable reading is not a pass.
