How to Test Browser Audio Latency and Read the Result

A click-after-tone browser test can compare output setups quickly, but it does not isolate hardware latency. The result combines your reaction time with browser scheduling and the selected audio path.

Reviewed and updated July 31, 2026

What the browser test measures

The timer starts when the page schedules and plays a short tone, then stops when you click that you heard it. The number therefore includes auditory reaction time, input-device response, browser work, operating-system routing, and output buffering.

Treat the value as a response-delay estimate. It is most useful when the same person compares two setups under controlled conditions.

Why five trials and a median help

A single click can be early or late because attention varies. Running five trials exposes that natural spread. The median uses the middle result after sorting and is less affected by one unusually slow click than an average.

Keep the device, browser, tab visibility, volume, and hand position unchanged. If the five values are very scattered, run another set before drawing a conclusion.

Comparing Bluetooth, wired, and speakers

Bluetooth audio often adds codec and buffering delay. Wired headphones can remove that wireless stage, while speakers add an acoustic path through the room. Audio interfaces can have their own driver and buffer settings.

Run a complete set for one output, record the median, switch the output, and repeat. Do not compare one Bluetooth trial with one wired trial.

What browser timing fields mean

The Web Audio API may report baseLatency or outputLatency, depending on the browser and device. These are browser-level estimates and may be unavailable. They do not include your reaction time, and they may not describe every hardware or wireless buffer.

Use the reported sample rate as context. It identifies the audio context rate but does not by itself determine the final end-to-end delay.

When to use a loopback measurement

Music production, live monitoring, rhythm games, and driver tuning often need physical round-trip latency. Use a cable loopback, microphone recording, or dedicated measurement system that compares an emitted reference with a recorded return.

A physical test can include digital-to-analog conversion, acoustic travel, microphone or input conversion, driver buffers, and application processing that a click test cannot isolate.

Frequently asked questions

Is browser audio latency the same as hardware latency?

No. A click test includes human reaction and browser behavior, while hardware or round-trip latency requires a physical measurement path.

Does Bluetooth always have more latency than wired audio?

Bluetooth commonly adds codec and buffering delay, but the exact difference depends on the devices, operating system, codec, and browser path.

Why is outputLatency not reported?

Browser and platform support varies. An unavailable field does not mean there is zero output latency.

Should I use the average or median?

For a small set of human-response trials, the median is usually more robust against one unusually slow click.