Device test · Camera & mic
Video Call Test
Checks your camera, mic and playback before a meeting.
Press “Start the check” to open the microphone and then the camera. Each asks for its own permission, and refusing one does not stop the other.
Input level is RMS in dBFS, measured against digital full scale. The browser has no calibrated reference, so it cannot give room loudness in dB SPL. There is no correct input level, so this row prints numbers and never a pass or fail. RMS is the latest sampled window; the peak and the clipped-window count hold since you started or last reset. “Reset peak & clip” clears those two without releasing either device or changing your answers. For a moving meter, waveform and signal-to-noise reading, use the microphone test.
Press “Record 3s and play it back”, speak for three seconds and listen for playback through your current output. Turn the volume down before recording. The clip stays in memory; this page does not save it to disk or upload it, and discards it when you stop or leave.
Did you hear the playback? The browser knows the clip played; only you can say whether you heard it. Answer after listening to the clip.
Does the picture look right? Check the preview for framing, lighting and focus; the page sees only that frames are arriving.
Speaker-to-microphone bleed
Take headphones off first, and turn the volume down. Then select your speakers as the output in your system or browser sound settings: with headphones selected there may be no acoustic path to the microphone, and the result would say nothing about your speakers. The button plays a 1.2-second sine tone at −20 dBFS through the current output, ramped in and out over 30 ms to avoid clicks. How loud that is at your ears depends on your volume, and through headphones even a short tone can be an uncomfortable squeal. Stop if the sound is uncomfortable.
No feedback loop can form through this page: the microphone analyser is not connected to the audio output, and the probe oscillator is not fed by the microphone. The tone plays only after you press the button, never on page load.
Keep this tab in front and stay quiet. The probe measures a narrow band near 2 kHz for 1 second with nothing playing, then plays the tone for 1.2 seconds and reports how far the level rose, in dB. Below 6 dB the tone did not rise above the floor; 20 dB or more is a large rise, meaning the microphone is clearly picking up the tone.
How to read the results
Each row says how it is known. measured covers access checks and calculations made by this page, reported repeats settings supplied by the browser, and confirmed by you records your answer about the playback or preview. The summary is built from the pass/fail rows; the level, processing, format, frame-rate and bleed rows report values with no single correct answer.
| Check | Evidence | Result |
|---|---|---|
| Microphone access | measured | not run |
| Microphone signal | measured | not run |
| Input level | measured | not run |
| Audio processing | reported | not run |
| Track format | reported | not run |
| Camera access | measured | not run |
| Camera frames | measured | not run |
| Camera format | reported | not run |
| Presented frame rate | measured | not run |
| Speaker-to-mic bleed | measured | not run |
| Playback heard | confirmed by you | not run |
| Picture looks right | confirmed by you | not run |
How this test works
Pressing “Start the check” makes two getUserMedia requests, microphone first and camera second, so a problem with one device is named on its own row while the other is still checked. The microphone feeds a Web Audio AnalyserNode for RMS, sample peak and a count of inspected windows containing a clipped sample. The camera request prefers 1280×720, and the browser can return a different size. Camera format shows the size and nominal frame rate the track reports, using the video element for size if needed, while Presented frame rate is counted by the page from video-frame callbacks. Without those callbacks, Camera frames follows the video element's clock and Presented frame rate reports that it cannot be measured in this browser. Audio processing and channel count are settings the browser reports, and this page does not ask the browser to turn its processing off. The sample rate shown is the audio graph's rate, which can differ from the hardware's. Missing settings read “not reported”.
For the optional bleed check, the page reads a narrow band for 1 second with nothing playing, then plays a 1.2-second sine at −20 dBFS through the current output. It skips the first 150 milliseconds of the tone readings for the ramp and acoustic delay, then subtracts the baseline median from the tone median to give a difference in dB. Medians keep a single loud moment, such as a door closing, from setting the result, though a sound that changes across the two windows still shows up in it. The probe sits near 2 kHz because small laptop speakers reproduce it and it is clear of mains hum; the page snaps it to an analysis bin using the audio graph's sample rate and prints the frequency it used. Each window needs at least 12 finite readings or the result is withheld, so keep the tab in front while it runs.
There is no far end here: no other participant, call or server is involved, so the page cannot measure network performance, sync between sound and picture, what a remote participant receives or whether a meeting app works with your devices. The bleed figure describes this device at the current output volume, with any browser processing already applied. If the browser reports that echo cancellation is on in this stream, a small figure can mean the browser is suppressing the tone, the output is low or little sound reaches the microphone. Nothing here scores the microphone or camera; the preview and the playback are there for you to judge. If the preview looks grainy, read why webcams look grainy and how to fix it.