Stereo Test — Pan & Phase Check

A stereo test checks that your system images correctly: left is left, right is right, motion between them is smooth, and the mix survives a mono sum. The page plays a 440 Hz tone, loops an 8 second pan sweep, and adds polarity inversion, a mono sum, and a level meter. Everything runs in your browser; nothing is uploaded.

How to use it

1. Check identity. Choose Left, Both, or Right and press Play channel. A 440 Hz sine is panned hard to the chosen side; Both sits at centre.
2. Follow the pan sweep. Press Play pan sweep. The tone travels left, centre, right, centre on an equal-power curve, looping every 8 seconds, with the pan value shown every 100 ms.
3. Set a manual pan. In the phase panel, drag Manual pan from -1.00 through 0.00 to +1.00 in 0.05 steps.
4. Invert polarity. Press “Polarity: normal (0°)” to switch it to “Polarity: inverted (180°)”, which inverts the right channel only.
5. Switch the mono sum on. Press “Mono sum: off” to make it “Mono sum: on”. The mono path adds the two channels with a 0.5 gain and switches the stereo path off.
6. Read the verdict. With the mono sum on and polarity inverted, the meter falls below -24 dBFS for more than 250 ms and the verdict reports cancellation; with normal polarity it reports reinforcement.

What the three panels test

Each panel checks a different property of the stereo field, and they are best run in order.

Identity: left, centre, right

The identity panel routes a 440 Hz sine to hard left (pan -1.00), centre (0.00), or hard right (+1.00). A correct system puts Left in the left speaker, Right in the right, and Both in the phantom centre. Swapped labels mean a reversed cable, adapter, or operating-system setting.

Pan sweep: smooth travel

The sweep follows pan = -cos(2πt / 8), so it starts hard left, crosses the centre twice per loop, and holds its loudness as it moves. Listen for continuous travel with no clicks, no sticking at the centre, and no sudden jump; a dip in the middle points at a level or phase problem, and a jump at a routing fault.

Mono and phase: compatibility

With the mono sum on and normal polarity, the channels reinforce and the verdict says so. Invert the right channel and the sum collapses toward silence, reported below -24 dBFS once it holds for 250 ms. That collapse is the mono-compatibility check: content that cancels this way sounds thin on a mono device or a single speaker.

How stereo perception works

Human directional hearing uses two cues. Below about 1.5 kHz the brain relies mainly on interaural time difference, the arrival gap between the ears, which reaches roughly 0.65 ms at the sides of the head, a path difference of about 0.22 m. Above about 1.5 kHz the head shadows the far ear and interaural level difference takes over, reaching up to about 20 dB at high frequencies. At 440 Hz the period is 2.27 ms, so a 0.5 ms offset is a 79° inter-ear phase difference.

The sweep’s even loudness comes from the pan law. An equal-power panner sets the left and right gains to cos and sin of an angle, so a centred image sits at 0.707 in each channel (-3.01 dB) and a hard pan sits at 1.0 and 0. The Web Audio StereoPannerNode uses this law.

On speakers, a centred image is a phantom: identical correlated signals from two drivers sum in the air, and the image holds only if the channels are matched in level to within a decibel or so and sit symmetrically in the room. The reference layout follows the ITU-R BS.775 arrangement, with front left and right at ±30° and the listener on the apex of a 60° triangle.

Rooms break the illusion predictably. If one speaker is farther away by Δd, identical mono content arrives twice with a path difference, and the first cancellation lands at 343 / (2 × Δd) Hz: 171.5 Hz for 1 m, 343 Hz for 0.5 m. The result is a comb filter that scoops narrow bands out of the response.

Cue or property	Dominant range	Typical magnitude	What it exposes
Interaural time difference	Below about 1.5 kHz	Up to 0.65 ms	Channel delay and time alignment
Interaural level difference	Above about 1.5 kHz	Up to 20 dB	Channel balance and head shadow
Equal-power pan law	Full band	-3.01 dB per channel at centre	Smooth travel without a level dip
Relative polarity	Full band	180° inversion	Wiring and mono compatibility
Mono sum	Full band	Doubles in phase, cancels inverted	Correlation and comb filtering

Accuracy, limits and troubleshooting

• Levels are dBFS, not dB SPL. The meter reports digital level on the stereo or summed path; it has no microphone or calibration.
• Cancellation is not digital silence. The two branches are never perfectly identical, so the mono sum settles below the -24 dBFS threshold instead of at zero, and the verdict waits 250 ms to avoid reacting to a brief dip.
• The mono sum uses a 0.5 gain. Do not read a mono measurement as a doubled level; compare it with the stereo reading to judge reinforcement or cancellation.
• Speakers include the room; headphones do not. Smooth headphone travel plus a rough speaker sweep points at the room rather than the source.
• Spatial audio renderers break the test. Windows Sonic, Atmos for headphones, and head-tracked spatial audio re-render the stereo field, so disable them before measuring pan and phase.
• Bluetooth and true wireless can add their own processing. A2DP carries both channels in one stream, but some true wireless sets process each side independently. Test wired if a result looks wrong.
• The pan readout updates every 100 ms while the audio pans smoothly, so the display lags the sound slightly. Judge travel by ear.
• This page cannot judge absolute polarity, because it only compares the two channels with each other.
• Keep levels modest: 440 Hz gets uncomfortable quickly.

Related tools

• Sound Localization Test — A scored eight-trial test of your own ITD and ILD placement accuracy.
• Speaker Test — Per-channel routing, balance, and polarity checks for loudspeakers.
• Headphone Test — Hardware checks and a slow ITD sweep for headphones.

Aliases covered on this page: stereo test, stereo perception test, left right stereo test, pan sweep test, and mono compatibility test.