Audio latency is the gap between asking for a sound and the sound actually leaving your speaker. It matters for music apps, gaming, and video conferencing. Wired connections are usually 5–20 ms; Bluetooth headphones add 100–250 ms. The browser reports what it thinks the system latency is via baseLatency and outputLatency on the AudioContext — close to reality, but not always perfectly accurate.
The tap-to-beat test
This is a perceptual measurement. The page plays a click every second and asks you to tap along. The average offset between click and tap is the round-trip latency you actually feel — it includes the audio reaching your ear, your brain registering it, your hand moving, and the browser receiving the input event. Below 50 ms feels instantaneous; over 150 ms feels laggy.
Frequently asked questions
Why does my Bluetooth headset show 200 ms in the audio latency test?
Standard Bluetooth audio codecs (SBC, AAC) buffer significantly. Lower-latency codecs (aptX LL, LC3) drop it to 40–80 ms but require both transmitter and headphones to support them.
Can I use this audio latency test for music production?
Use it as a rough check. For tight monitoring you'd switch to ASIO or CoreAudio drivers in your DAW — those bypass the browser stack entirely and can hit single-digit milliseconds.
Why does the reported latency change between runs?
Browsers re-negotiate buffer sizes when other apps grab the audio device, when laptops switch power profiles, or when you change output. Re-run the probe after switching outputs.
Is the audio latency test accurate enough for gaming setup tuning?
Yes for relative comparisons (this headset vs. that one), less so for absolute targets. The tap-to-beat test catches obvious lag; if you need single-millisecond accuracy you'll want a hardware loopback meter.