# Headphone Test > Just bought headphones? Detect reversed channels and learn what to check, compare loudness ear to ear, then run a sweep with a plain-language listening guide. - URL: https://bytescope.dev/en/headphone-test/ - Locale: en · Also: [ja](https://bytescope.dev/ja/headphone-test.md) · [zh-Hant](https://bytescope.dev/zh-Hant/headphone-test.md) - Runs: Your files never leave your browser — all processing is local. - Requires: Any modern browser (Chrome, Edge, Firefox, Safari), desktop and mobile. - Updated: 2026-09-08 ## What it does Put the headphones on, press play, and answer what you hear. The tones are generated in the browser with Web Audio — no audio files are downloaded and nothing is uploaded — and the instruments are aimed at the questions that come with a new pair. The left / right / both channel test ends in a verdict, and if your answers show the sides reversed, the page says so outright and lists what to check. The 20 Hz to 20 kHz sweep has a live readout and a button to mark the frequency where the sound vanished. The stereo test pans a tone from ear to ear, and separate low- and high-frequency checks probe the two ends of the range on their own. Start at low volume: the drivers sit millimetres from your eardrums, and a sweep at high volume is genuinely unpleasant — at the top of the range, it is not good for anyone. ### When left and right come out reversed A reversed pair is more common than it sounds, and it is almost never the drivers themselves. When the verdict says swapped, check in this order: the cable connector, if the cable detaches — some seat in either direction but are wired for one; the L and R markings, misprinted or simply misread on housings that look identical; the operating system's balance and channel settings, where a virtual audio device or a leftover surround profile can mirror the output; and any adapter or extension in the path, because a mis-wired TRS adapter swaps the pair silently. Rerun the test after each change — it takes ten seconds and tells you immediately whether the fix worked. Only when all of that is eliminated does the fault point inside the headphones, and a pair hard-wired backwards is a warranty case, not something to live with. ### Listening for frequency response — honestly The listening guide walks the sweep in plain language: what a mid-bass hump sounds like, what treble that is present rather than boosted means, where voices sit and why. It is explicitly subjective — you are listening, not measuring, and this page will not dress an impression up as data. Results depend on the system volume, any OS or application equaliser, and over Bluetooth on the codec and its resampling. The sweep's loudness is evened out by an approximate A-weighting compensation clamped to ±12 dB, which makes it listenable, not flat. And the frequency where you stop hearing the sweep is bounded by the whole chain at least as much as by your ears — do not read it as a hearing test. What this page does settle reliably: whether the sides are swapped, whether one driver is quiet or dead, and whether the pair you are deciding to keep has a fault worth returning while you still can. ## Limits Does not measure frequency response with instruments; it plays left/right, phase and sweep tests you judge by ear. Start with the volume low. The drivers sit millimetres from your eardrums, and a sweep at a level that felt fine for music is genuinely painful around 1–4 kHz. ## Machine output Copy JSON copies an envelope `{ tool, url, generatedAt, userAgentHint, result }`. `tool` is `"speaker-test"` or `"headphone-test"`; `result` is the same listener-test shape on both pages. - `testsPlayed` (string[]) — Instruments the listener actually started, in on-screen order: `"channels"`, `"sweep"`, `"pan"`, `"band-low"`, `"band-high"`. Empty until something has been started. - `channelAnswers` (array of objects) — Left, right, both — always that order. Empty until the channel run finishes. Each item is `{ played, heard }`; `heard` is `"left"` | `"right"` | `"both"` | `"nothing"` or `null` when that prompt was not confirmed. - `channelVerdict` (string | null) — `"correct"` | `"swapped"` | `"left-silent"` | `"right-silent"` | `"both-silent"` | `"mono"` | `"inconclusive"`. `null` until the listener has answered enough for a verdict. - `sweepPlayed` (boolean) — Whether the frequency sweep was started. - `sweepMarkedHz` (number | null) — Frequency the listener marked. `null` = sweep not marked. - `panPlayed` (boolean) — Whether the pan sweep was started. - `bandLowHeard` (string | null) — `"yes"` | `"no"`. `null` = low-band prompt not answered. - `bandHighHeard` (string | null) — `"yes"` | `"no"`. `null` = high-band prompt not answered. - `sampleRate` (number | null) — AudioContext sample rate in Hz. `null` = context not opened. ## FAQ ### One side sounds quieter — is it real or am I imagining it? Three steps separate the suspects. First check the OS balance slider — it is the most common cause and the only free fix. Then run the channel test, which plays the same tone at the same level into each ear, so a difference you hear there is a real difference in the chain. Finally, wear the headphones reversed — left driver on right ear — and repeat: if the quiet side follows the headphones, the fault is in the pair; if it stays on the same ear, it is the source or your own hearing, which is rarely symmetric. A clear level difference that follows the headphones on a new pair is exactly what warranty windows are for. ### The page says my channels are swapped — are the headphones defective? Usually not. Work through the cheap explanations first: a detachable cable seated the wrong way round, an L/R marking you misread, the OS balance and channel settings, and any adapter in the path — a mis-wired TRS adapter reverses the pair without any other symptom. Each takes seconds to test with a rerun. If every one of those is ruled out and the verdict still says swapped, the pair is wired backwards internally — that is a manufacturing defect and a legitimate warranty claim, and no amount of settings will fix it. ### Can I use the sweep as a hearing test? No — and headphones make the temptation stronger, because they reach far higher than laptop speakers. The number where the sweep vanishes still describes the whole chain: the headphones' own response, the system volume you happened to choose, any equaliser, and over Bluetooth the codec, which trims the top of the range first. The A-weighting compensation is approximate and clamped to ±12 dB — it keeps the sweep bearable, it does not calibrate anything. An audiologist's booth controls all of these variables; a web page controls none of them. Treat the readout as a fact about this pair on this system, nothing more. ### Do Bluetooth earbuds ruin the results? It depends on the question. Left/right identification and swap detection survive any codec — compression does not relocate a channel. Loudness comparison mostly survives too. The sweep and the listening guide are where Bluetooth intrudes: the codec resamples and discards signal, high frequencies go first, and many earbuds run their own processing and ANC on top, so what you hear is the earbuds' opinion of the tone rather than the tone. For anything about frequency response, test wired if the option exists at all. ### Can the sweep damage my headphones — or my ears? Your ears are the honest concern. Headphone drivers handle sine tones at reasonable levels without complaint, but they sit millimetres from your eardrums, and a sweep concentrates everything at one frequency — around 1–4 kHz it gets painful at volumes that felt fine for music. Start the volume low, raise it only until the test is clearly audible, and never launch the sweep at a level set for something else. If a tone distorts or buzzes at high volume, turn it down before wondering about the drivers: the component at risk in this test is you. ## Related tools - [Speaker Test](https://bytescope.dev/en/speaker-test.md) — Test speakers and a soundbar for crossed wiring, a dead channel and bass that never arrives. - [Mic Test](https://bytescope.dev/en/mic-test.md) — Test your microphone with a live VU meter, waveform and spectrum, plus record-and-playback.