Quick answer
What this guide helps you determine
Use a live level meter, waveform and a short local playback to check that sound reaches the browser.
Begin with the observable browser result, repeat it under the same conditions, and change one connection, setting or application at a time. A browser result narrows likely causes but does not certify hardware or decide a repair.
Scope of this check
A basic microphone test should answer three questions: did the intended input reach the browser, is its level usable without obvious clipping, and can a short local playback be understood? It is not a calibrated loudness, frequency-response or professional quality measurement.
Step-by-step method
- Lower speaker volume or use headphones to prevent feedback, then click Start and grant microphone permission. Select the intended input after labels are exposed and speak at the distance you normally use.
- Watch the live meter and waveform for movement. Speak quietly and at normal level; peaks should respond without staying pinned at the maximum. If available, record a short neutral sentence and play it locally once.
- Delete the temporary playback and press Stop. Verify the system microphone indicator turns off. Repeat inside the target meeting or recording app because it may select another input or apply noise processing.
What to check first
Use these checkpoints to keep the comparison focused. Record the visible behavior rather than assigning a fault label too early.
- Start only after reading the privacy note
- Speak at a normal distance
- Stop the media tracks when finished
How to read what you see
| Observation | What it can suggest | Useful next check |
|---|---|---|
| Meter and waveform move with speech | The browser received changing audio samples from the selected input. | Check playback and level consistency. |
| Meter stays at zero | Mute, wrong input, permission or device availability may be responsible. | Check hardware mute and device selection. |
| Peaks remain at maximum and playback distorts | Input gain or source level may be clipping. | Move back or lower input gain before retesting. |
Common mistakes to avoid
- Do not speak confidential information into a diagnostic recording.
- Do not use the meter as a sound-pressure or hearing measurement.
- Do not leave the microphone capture indicator active.
Build evidence another person can use
Write down the exact action, visible result and whether it repeats. Include the browser family, operating-system family, connection type and the single comparison that changed the result. Quote the observation directly—for example, “Meter and waveform move with speech”—then add where it occurred and which comparison changed it.
Change one condition per comparison and return settings that have no effect. Follow model-specific manufacturer instructions before firmware changes, cleaning, disassembly or warranty action.
When to seek qualified support
If the same problem remains in multiple applications and on another compatible connection or computer, collect the repeatable observations and contact the manufacturer or a qualified technician. Stop using equipment that is physically damaged, unusually hot, wet, sparking or unsafe. Device Test Bench cannot inspect electrical condition or make a repair decision.
Privacy during testing
Our public tests process key events, pointer values, media streams and controller input locally. Do not send private recordings, complete typed text or device identifiers when asking for help. A short description of the behavior, browser and page URL is normally enough.
Questions readers ask
Is the recording uploaded?
No. The short recording remains in browser memory for local playback until you delete it or leave.
Why does my meeting app sound different?
It may use another microphone, automatic gain, noise suppression, echo cancellation or network compression.
Primary technical references
The following specifications define the browser APIs discussed in this guide. They are primary web-platform references, not product repair manuals.