Audio
Audio issues are common but hard to diagnose. Test your microphone levels, verify speaker output, and check left/right stereo channel balance — all from your browser.
Microphone Test
microphone test, mic test, mic test online
[READY]
Open Tool →Speaker Test
speaker test, sound test, speaker checker
[READY]
Open Tool →Stereo Test
stereo test, left right audio test, channel test
[READY]
Open Tool →Audio
Audio issues are common but hard to diagnose. Test your microphone levels, verify speaker output, and check left/right stereo channel balance — all from your browser.
Privacy First audio Testing Tools on usbRecognized run entirely in your browser using secure Web APIs. No software download, no installation, no data upload — your privacy is 100% protected. Simply click on a tool above to start testing.
Some tools in this category are available now (marked with ✓). Others are coming soon — each coming soon page includes a detailed preview of what the tool will do and how you can test your hardware in the meantime.
What Browser Audio Tests Can and Cannot Measure
A browser can drive a known signal out of a chosen channel and capture a known signal from an input device, and that is enough to answer most practical questions about a sound card. It can tell you whether a channel is dead, whether the two channels are swapped, whether the output produces distortion at high volume, and whether a microphone is picking up anything at all.
What it cannot do is measure the analogue stage. Total harmonic distortion, signal-to-noise ratio and frequency response all depend on the converter, the amplifier and the transducer, none of which the browser can observe directly. A measurement taken through the browser describes the whole chain ending at the browser event loop, so it includes the operating system mixer and any enhancement the driver applies.
This boundary is worth keeping in mind because it changes how a bad number is interpreted. A microphone that reports a high noise floor through the browser may be an electrically noisy microphone, or it may be a laptop with aggressive automatic gain control that is amplifying the room. Only a second measurement with the enhancement disabled separates the two.
Channel, Balance, and Phase
Channel identification is the first audio check to run and the one most often skipped. Playing a tone out of the left channel and hearing it from the right speaker is a wiring or driver fault that no amount of equalisation will fix, and it is invisible if you only ever play content that mixes both channels equally.
Balance is distinct from channel identity. Both channels can be present and correctly assigned while one is consistently quieter, which usually points at a faulty amplifier stage, a partially inserted plug, or a driver that has applied a balance offset. A stereo test that plays a tone at equal level in each channel, toggled back and forth, makes a level difference obvious in a way that simultaneous playback does not.
Phase is the third property and the one that produces the strangest symptoms. If the two channels are wired with opposite polarity, the sound appears to come from behind the listener, bass frequencies cancel out, and the image collapses when the track is summed to mono. It is rare on consumer hardware but common after a cable repair or a custom adapter, and it is easy to confirm by summing the channels and listening for the low end to disappear.
Microphone Level, Noise Floor, and Clipping
A microphone test reports two numbers that matter far more than the peak level people usually watch: the noise floor with the room silent, and the level produced by normal speech. A microphone that peaks loudly when you speak but sits in audible hiss when you stop has a poor signal-to-noise ratio, and no amount of gain adjustment in a call application will repair it.
Clipping is the opposite failure and is usually self-inflicted. When the input gain is too high, loud passages are flattened at the maximum value and the waveform becomes a square wave, which sounds harsh and cannot be recovered afterwards. Because laptop microphones frequently ship with automatic gain control enabled, the gain can rise during a pause and clip the first syllable of the next sentence.
The distance from the microphone matters more than any setting. Doubling the distance roughly halves the captured level while leaving the room noise unchanged, so a microphone that sounds thin and echoey is often simply too far away. Moving it closer improves the ratio between voice and room without touching a single control, and should be tried before any software change.
When the Fault Is Software, Not Hardware
Audio faults are more likely to be software than any other category on this site, because the signal passes through more software layers. A browser processes audio through the operating system mixer, which may apply per-application volume, then through any driver enhancement, then through the application layer. Each of these can attenuate, equalise or spatialise the signal independently.
Exclusive mode is a common culprit. When an application takes exclusive control of an output device, other applications may be silenced entirely, which looks like a dead channel until the application is closed. Communication-mode ducking is another: many drivers automatically reduce all other audio by a fixed percentage when a call begins, and the setting that controls it sits in the driver control panel rather than in the operating system sound settings.
A simple test separates software from hardware. Play the same tone through the browser and through a system-level player, and compare. If both are wrong, the fault is below the application layer and worth investigating in hardware. If only the browser is affected, the browser or the driver enhancement applied to it is responsible.
Audio FAQ
Do I need to install anything to test my microphone or speakers?
No installation is required. The tests use the Web Audio API, which every current browser implements. The trade-off is that the browser can only observe the signal after the operating system mixer has processed it, so the measurements describe the whole chain rather than the hardware in isolation. That is usually the more useful number anyway.
Why is one speaker louder than the other?
First confirm the channels are assigned correctly, then check for a balance offset in the operating system sound settings. If both are correct, the difference is likely in the amplifier stage or the physical connection. A plug that is not fully inserted can leave one contact partially engaged, which produces exactly this symptom and costs nothing to check.
How do I test left and right channels separately?
Use a channel test that plays a tone in one channel at a time rather than a stereo track. Playing the channels in isolation tells you two things that simultaneous playback cannot: whether the channels are assigned to the correct side, and whether a tone intended for one channel is bleeding into the other. Bleed indicates a wiring or driver fault rather than a balance setting.
My microphone sounds quiet. Is it broken?
Probably not. Check the input level in the operating system first, then move closer to the microphone and test again. Distance has a larger effect on captured level than most people expect, and room noise does not fall off the same way. Only if the level stays low with the microphone close and the gain at maximum is the hardware itself suspect.
Can a browser measure audio latency?
It can measure round-trip latency through the browser, which includes the audio stack, the operating system mixer and the driver buffers. That figure is much larger than the hardware latency a specification sheet quotes, so it is useful for comparing configurations on one machine but not for validating a manufacturer claim. Bluetooth devices in particular add latency the browser can measure but cannot reduce.
Why does my headset only play in one ear?
Work through the chain in order: confirm the plug is fully seated, check that the channel test produces sound in both sides, then check the balance setting, then check whether a mono audio accessibility option is enabled. A mono setting downmixes both channels into one, which sounds like a dead ear but is a single checkbox and is a surprisingly common cause.
What is a normal microphone noise floor?
It depends heavily on the environment, which is why a comparison against a known-good setup on the same machine is more useful than an absolute figure. Record a few seconds of silence and look at the level meter: a meter that sits near zero and only moves when you speak is healthy. A meter that shows constant visible movement in a quiet room indicates either a noisy input stage or a very noisy environment.