Audio Latency Test: Measure Bluetooth & Headphone Delay
Overview
Audio Latency Test. The Audio Latency Test measures the delay between a sound firing and you actually hearing it, two ways: click Measure Now to read your browser's reported output and base latency, or run the Play Sound / I Heard It! tap test to get an estimated total based on your own reaction time. That delay is what causes lip-sync problems in gaming, live streaming, video calls, and music production. Choir members and karaoke fans use the vocal range test online for music basics, such as picking a song key that suits the voice.
Every earbud, speaker and sound card adds a little delay before a sound reaches your ears. This audio latency test is a latency tester you run in your browser: it measures that gap in milliseconds, so you can tell whether your device and headphones are quick enough for games, calls or music.
How the Audio Latency Test Works
The check runs in two phases. First you tap when the screen shows a flash, which records how fast you react to what you see. Then you tap when you hear a beep, which records your reaction plus the audio delay. Subtract one average from the other and the latency remains:
$$\text{latency (ms)} = \overline{t}_{\text{audio}} - \overline{t}_{\text{visual}}$$
Each phase runs five rounds, so one stray tap can't decide the result.
Calibrate Your Reaction Time First
The calibration phase sets your baseline. Your reaction time to a visual cue is often near a quarter of a second, but it varies, so the test measures your reaction speed on every run. Sit in a quiet room, set a comfortable volume, and keep the flash in your peripheral vision rather than staring at it.
Tap to the Beep and Read Your Bluetooth Audio Latency
A worked example from a bluetooth audio latency test on Bluetooth headphones:
| Round | Visual tap (ms) | Audio tap (ms) |
|---|---|---|
| 1 | 248 | 409 |
| 2 | 262 | 421 |
| 3 | 239 | 397 |
| 4 | 255 | 515 |
| 5 | 251 | 413 |
The visual taps average 251 ms. Round four's audio tap is one of the outliers, about 100 milliseconds slower than the rest, so it is dropped and the audio average becomes 410 ms. The result: 410 − 251 = 159 ms. Some tools use a slider instead: you delay the flash until it and the click feel simultaneous.
Accuracy and the Wired Headphones Baseline
Accuracy describes how consistent your taps were; the visual taps above stay within about 12 ms of their average. To isolate the wireless delay, run the check on wired headphones too. Here the wired run reads 17 ms, the end-to-end audio output latency of the device and cable, so the Bluetooth link adds 159 − 17 = 142 ms. Repeat the Bluetooth latency test a couple of times and keep the median.
Typical Wireless Delay by Bluetooth Codec
The codec your phone and headphones agree on sets most of the wireless delay: To memorise how 100 Hz, 1 kHz and 10 kHz sound, practise with repeated tones, streaks and replay in the free frequency identification training until the pitches feel familiar.
| Connection | Typical delay | Comfortable for |
|---|---|---|
| Wired headphones or speakers | 3-20 ms | Everything, including live instruments |
| LE Audio (LC3) | 30-60 ms | Gaming, video |
| aptX Adaptive | 50-80 ms | Gaming, video |
| aptX | 100-150 ms | Video, casual games |
| AAC (most earbuds) | 120-200 ms | Music, most video |
| SBC | 170-280 ms | Music |
Both ends must support a codec, or the link falls back to AAC or SBC.
Why Bluetooth Latency Runs Higher
A cable carries a continuous signal. Bluetooth audio delay comes from the radio link: audio is encoded, split into packets, buffered against radio interference, then decoded in the headphones. The buffer adds the most, as the transmitter trades delay for a stable stream.
Gaming Mode, aptX Low Latency and Other Fixes
- Turn on gaming mode or a low latency mode in the headphone app.
- Pick hardware with aptX Low Latency, aptX Adaptive or LE Audio on both ends.
- Stay within a few metres of the source device.
- For fast games, plug in: a cable gives the shortest headphone delay.
How Much Audio Delay Is Acceptable?
Read your result against what you use it for. Your hearing tolerates lag when nothing depends on timing: on an album, even 250 ms goes unnoticed. Once sound must match a picture or your actions, the limit drops fast.
Competitive Gaming and Video Calls
In competitive gaming you react to sound cues such as footsteps, so every millisecond of lag delays your response; most gamers aim for under 50 ms. Video calls are more forgiving: talk flows up to roughly 150 ms, beyond which people talk over each other.
Music Production, Monitoring and Lip-Sync
Music production is the strictest case. When musicians, or anyone singing, hear themselves through in-ear monitors during a live performance, more than about 10 ms sounds like an echo, so monitoring stays wired. For video editing, video editors need picture and sound in sync; the ITU-R BT.1359 broadcast recommendation puts the point where viewers detect sound lagging picture near 125 ms. It's the same lip-sync problem you notice when watching movies over Bluetooth on a TV, soundbar or projector, where an AV sync setting shifts the picture to match the sound source. So a reading above 10 ms means using a cable on stage, and one above about 125 ms means dialogue visibly trailing the picture.
Checking Headphone Delay Before a Lip-Sync Edit
I'm cutting a 30 fps interview on a laptop, wearing AAC earbuds, and the question is simple: can I trust what I hear when I line up dialogue with mouth movements, or do I need to plug in? The answer turns out to be no: the earbuds put sound 177.8 ms behind the picture, a little over five frames.
I run the visual phase first. My five flash taps land at 231, 244, 226, 238 and 229 ms, an average of 233.6 ms with a spread of about 6.5 ms, steady enough to use. Then the beep phase through the earbuds: 418, 402, 431, 409 and 397 ms, average 411.4 ms. The difference, 411.4 − 233.6, is 177.8 ms of audio latency. Dividing by one frame at 30 fps (33.3 ms) gives 5.3 frames. That is well past the roughly 125 ms where ITU-R BT.1359 says viewers start to notice sound trailing picture, so every sync point I set by ear would sit about five frames early.
Before deciding, I rerun with one thing changed each time:
| Setup | Audio tap average | Latency | Frames at 30 fps |
|---|---|---|---|
| Earbuds, default | 411.4 ms | 177.8 ms | 5.3 |
| Earbuds, gaming mode on | 336.0 ms | 102.4 ms | 3.1 |
| Wired, laptop headphone jack | 253.0 ms | 19.4 ms | 0.6 |
Gaming mode gets under the 125 ms detection point, which fools you into thinking it's fine, but three frames is still a visible slip on a hard consonant like a "p". Only the wired run stays under one frame. My call: earbuds for the rough assembly, the cable for every sync pass.
One caveat a single number hides: the flash phase also carries the screen's own display lag, so the test reports audio delay minus display delay. On this laptop's panel that's small, but on a TV with 40-plus ms of input lag the result would understate the earbuds by that much. Taking the wired run away from the Bluetooth one cancels it: 177.8 − 19.4 leaves 158.4 ms for the wireless link itself, and that's the figure to compare when shopping for lower-latency earbuds.
Measure Round-Trip Latency with a Loopback Cable
To measure audio latency without your reactions in the loop, connect the output of an audio interface or sound card to its input with a loopback cable, play a short sine sweep and record it back; cross-correlation finds how many samples later it arrived. Measure latency this way and you get round-trip latency (output plus input), a lower bound: a loudspeaker and microphone add the trip through the air, and any processing such as a filter or audio effect adds algorithmic latency. Convert samples to time:
$$\text{latency (ms)} = \frac{1000 \times \text{samples}}{\text{sample rate}}$$
So 384 samples at 48,000 Hz is 8 ms, fast enough for real-time audio processing.
Buffer Size and Sample Rate
Most of that comes from the buffer size your audio drivers use: a 256-sample buffer holds 5.3 ms at 48 kHz, a 128-sample buffer 2.7 ms. Smaller buffers cut delay but risk dropouts and underruns when the operating system can't refill them in time; a higher sampling rate shortens each buffer too. Match the driver settings to your software's sample rate and check for underruns after each stereo recording.
Browser Output Latency and the Web Audio API
Browsers report an estimate through the Web Audio API: baseLatency is the browser's processing buffer, outputLatency the time through the audio stack to the DAC. Neither sees inside a wireless device, so on Bluetooth a tap test reads higher than the reported audio output latency.
Frequently Asked Questions
- What is audio latency?
- Audio latency is the time between your device starting a sound and you hearing it. It builds up through the app, the operating system's audio stack, the DAC or Bluetooth link, and the headphones or speakers. Wired outputs usually add only a few milliseconds; Bluetooth can add well over 100 ms.
- How does this audio latency test work?
- In the default method you tap when the screen flashes (your reaction time) and then when you hear a beep (reaction time plus audio delay), five rounds each. The test subtracts the flash average from the beep average, after dropping stray taps, to leave the audio latency in milliseconds.
- How much audio latency is acceptable?
- It depends on the job. Under about 10 ms suits live monitoring, under 50 ms suits competitive gaming, and viewers start to notice sound trailing picture at roughly 125 ms. For plain music listening even 250 ms is fine, because nothing has to line up with it.
- Is 20 ms or 75 ms of audio delay noticeable?
- 20 ms is low enough for gaming and most monitoring, though some musicians will feel it when hearing their own voice. 75 ms is usually fine for video and casual games but feels slow in fast competitive play.
- Why is Bluetooth audio latency so high?
- Bluetooth audio is encoded, sent in packets, buffered against radio interference and decoded in the headphones. The codec matters most: SBC and AAC sit at the slow end, aptX Adaptive and LE Audio much lower, and both ends must support a codec to use it.
- How can I reduce Bluetooth latency?
- Turn on gaming or low latency mode in the headphone app, use a codec such as aptX Low Latency, aptX Adaptive or LE Audio if both devices support it, stay close to the source, and switch to wired headphones for competitive gaming or monitoring.
- How accurate is this test, and what does Accuracy mean?
- It is a perceptual estimate, not a lab measurement, and the result is only as steady as your taps. Accuracy shows how consistent your taps were; if it is low, retest in a quiet room and run it two or three times. For an exact figure use a loopback cable and an audio interface.