Bluetooth Audio Latency Test
Find out how many milliseconds your wireless headphones actually add.
Volume warning
This test plays tones that can be loud and may damage hearing or speakers. Turn your volume down before starting, then raise it gradually.
How this test works
Every Bluetooth connection adds delay; the question is how much, and whether it matters for what you are doing. This measures it two ways. First it reads the output latency your browser reports for the current device, which costs nothing and moves when you switch to Bluetooth on Chromium — though Firefox and Safari report zero regardless. Second, and more usefully, it runs a tap round: eight beeps at random intervals, and you press the moment you hear each one. That round on its own is meaningless, because your reaction time is far larger than the delay being looked for. So you save it as a reference on a wired output, switch to Bluetooth, and run it again. Reaction time is the same in both rounds and cancels in the subtraction, leaving the delay the wireless path added — measured on your ears rather than reported by an API.
What the results mean
- Added latency
- The difference between this round and your saved reference. This is the real measurement, and it appears only once a reference exists. Under 40 milliseconds is imperceptible, up to 100 is fine for music, up to 200 shows on video, and beyond that is unusable for anything that has to line up with a picture.
- Output latency the browser reports
- The audio stack's estimate for the current output device. On Chromium it tracks reality closely enough to be useful on its own. A zero here means the browser does not report the figure, not that there is no delay.
- Your reaction this round
- The median of your usable presses. It is mostly your own reaction time, which is why it is never presented as a latency. It exists so you can see that the two rounds are comparable.
- Tap consistency
- How much your presses varied. A wide spread means the difference between two rounds is noise rather than a codec, and the result says so instead of quoting a number.
Common problems and fixes
- The measurement comes out around 150 to 250 milliseconds
- That is standard SBC or AAC behaviour and not a fault. Both ends have to support a low-latency codec such as aptX Low Latency or aptX Adaptive for anything better, and where one end does not, the connection silently falls back without telling you.
- Latency is worse than it used to be with the same headphones
- Multipoint is the usual cause — holding a connection to a phone and a laptop at once adds buffering. Disconnect the second device and run the test again. A congested 2.4 GHz band does the same thing, so a room full of wireless devices is worth ruling out.
- Audio and video are out of step but only when watching
- That is what this delay does, and the delay is constant. Most players and televisions have an audio delay setting designed exactly for it — set it to the figure measured here rather than trying to make the headphones faster.
- The test says it needs a reference run
- Run it once with wired headphones or the built-in speakers, press the button to save that round as your reference, then connect the Bluetooth device and run it again. Without a reference there is nothing to subtract your reaction time against.
Frequently asked questions
Why can it not just tell me the latency in one run?
Because a single round measures your reaction time plus the delay, and human reaction time is 200 to 300 milliseconds — several times larger than the number being looked for, and different for everyone. There is no constant that can be subtracted. Comparing two rounds removes it exactly, which is why the reference step exists rather than being a convenience.
Is Bluetooth latency a defect I should return the headphones for?
Almost never. It is what the codec costs, and the common codecs cost 150 to 250 milliseconds. If yours land there, they are behaving normally. Low latency is a purchasing decision — a codec both ends support — rather than something a working pair can be fixed into.
Does the delay affect music?
No. A constant delay is inaudible when nothing has to synchronise with it: the whole track arrives late by the same amount and sounds identical. It only matters when a picture, a game or an instrument has to line up with the sound.
Why does my phone seem better than my laptop with the same headphones?
Different devices negotiate different codecs and run different audio stacks. The same headphones commonly get AAC on a phone and SBC on a laptop, which is a visible difference in this measurement. Running the test on each is the only way to know rather than guess.