Sound Localization Test — L/C/R

A sound localization test scores how accurately you locate short stereo cues. Eight shuffled trials play a 440 Hz tone with one hidden cue, a sub-millisecond timing difference, a 9 dB level difference, or a hard pan, and you pick left, center, or right after each one. Everything runs in your browser; nothing is uploaded.

How to use it

1. Use headphones in a quiet room. Set one low volume and leave it fixed for all eight trials, because the test compares your answers, not your playback level.
2. Press Start test. The page shuffles eight trials and lays out eight chips in the scoreboard.
3. Press Play trial. One 440 Hz tone plays; the readout shows which trial you are on and the answer buttons become active.
4. Answer with your first impression. Press Left, Center, or Right. The buttons stay disabled until a trial has played, so every answer follows a sound.
5. Read the feedback. Correct answers become green chips, misses become red chips, and the score tracks every answer as you go.
6. Repeat or reset. Press Restart test for a fresh shuffle. The page stores your last score locally with a timestamp, and Clear saved score removes it.

The three cue types

The eight trials are built from three cue families, each rendered with a different Web Audio node.

Timing (ITD)

Three trials use interaural time difference: left leads by 0.5 ms, right leads by 0.5 ms, and one trial has no delay at all, which should sit in the centre. The graph holds the leading ear at 0 ms and runs the other ear through a DelayNode, so the only difference between the ears is arrival time.

Level (ILD)

Two trials use interaural level difference: a 9 dB gap with either the left or the right ear louder. The quieter ear runs at 0.355 gain against 1.0 on the louder side, which is a 9 dB drop. At 440 Hz the head shadow a real source would produce is only a few dB, so the test uses a deliberately larger cue to make the direction unambiguous.

Pan

Three trials use a StereoPannerNode at hard left (-1.00), centre (0.00), and hard right (+1.00). This is the strongest cue of the three and the easiest to answer, which is why the pan trials act as a control for the timing and level trials.

What the physics says

The largest interaural time difference a human head produces is about 0.65 ms, so the 0.5 ms cue is roughly 77% of the physiological maximum. That delay equals a path difference of 343 × 0.0005 ≈ 0.17 m between the ears. Below about 1.5 kHz the brain reads direction mainly from timing; above that, the head shadows the far ear and level differences take over, reaching up to about 20 dB at high frequencies. A 440 Hz tone sits firmly in the timing-dominated region, which is why the test uses it for every trial type.

Direction cues are ambiguous on their own. Every ITD and ILD pair fits a whole circle of possible directions, the classic cone of confusion, so front and back can trade places. Head movement normally resolves that ambiguity, but the test asks you to keep your head still so each trial stays consistent.

Scoring is simple enough to interpret directly. With three possible answers, guessing scores about 2.7 out of 8 on average. Scoring 6 of 8 happens by chance roughly 1.7% of the time, and 8 of 8 has a chance probability of about 0.015%. A single miss is not meaningful, because attention and fatigue move the number by one or two trials.

Trial	Cue	Parameter	Expected answer
Timing, left leads	Delay on the right ear	0.5 ms	Left
Timing, right leads	Delay on the left ear	0.5 ms	Right
Timing, no delay	None	0 ms	Center
Level, left louder	Gain difference	9 dB	Left
Level, right louder	Gain difference	9 dB	Right
Pan hard left	Stereo panner	-1.00	Left
Pan center	Stereo panner	0.00	Center
Pan hard right	Stereo panner	+1.00	Right

Accuracy, limits and troubleshooting

• Headphones are required. Built-in laptop and phone speakers blend the two channels in the air before the sound reaches your ears, and a 0.5 ms cue cannot survive that crosstalk. It arrives at both ears as the same signal.
• The timing resolution is not a limitation. A 0.5 ms delay is 24 samples at 48 kHz and about 22 samples at 44.1 kHz, far above the one-sample resolution of a DelayNode. The cue is rendered exactly.
• Bluetooth usually keeps relative timing intact. A2DP carries both channels in one codec stream, so the inter-ear relationship survives the transport delay. Head-tracked spatial audio and some true wireless sets with independent per-ear processing can still disturb it, so disable spatial features.
• Mono settings destroy the test. An operating-system mono audio setting folds both channels into both ears and turns every trial into a centre answer. Check that setting before you start.
• Keep the volume fixed. The level trials are built into the signal, so changing your system volume between trials only adds noise to your results.
• Swap the cups if one side keeps missing. If errors follow the headphone cup, the hardware is at fault; if they follow your ear, that is a screening signal about you rather than the page. If you are concerned about your hearing, see an audiologist.
• One run is a screen, not a diagnosis. Eight trials cannot measure a hearing threshold or replace an audiogram. Repeat later, and compare runs rather than reading a single score.
• Levels are reported in dBFS, not dB SPL. The readout is digital level relative to full scale, and the page has no microphone or calibration.
• Keep sessions short. Eight trials is a few seconds of sound. Take a break rather than grinding through repeated runs, because fatigue flattens the result.

Related tools

• Stereo Test — The system-side check: channel identity, a pan sweep, and mono compatibility.
• Headphone Test — Confirm that both drivers work and compare them before you trust a localization score.
• Frequency Check — A quick per-side check of which frequencies you can hear.

Aliases covered on this page: sound localization test, stereo localization test, and left center right test.