Headphone positional capture & visualisation

Capture the same headphone at a centred seating and at controlled forward, backward, up, and down offsets. The graph shows how much of the measured response depends on the listener–headphone geometry, rather than treating one fixture position as universal.

Why seating position matters

Worked examples: Aune SR7000 and Grell Audio OAE2

A single measurement records one seating of the headphone on one fixture. With an angled driver, a small shift of the cup moves the driver relative to the ear, and the upper mids and treble can change by several decibels while the bass barely moves. A centred capture plus forward, backward, up and down offsets shows how much of the curve belongs to the headphone and how much to the fit.

Aune SR7000: the traces stay within about 3 dB below 1 kHz, then spread by more than 10 dB around 5 kHz.
Grell Audio OAE2: bass and mids hold together up to 1.5 kHz, while 3–6 kHz moves with position. The same review compares the Meze Strada, a conventional closed-back whose variation is in the bass seal instead.

1. Audio input / output

Measurement input — GRAS/B&K interface, Cosmos ADC, or USB-C audio input Playback output — headphone DAC/interface

Audio devices have not been scanned.

The automatic suggestion prefers names containing Cosmos, ADC, GRAS, B&K, measurement, or USB-C for the input, and an external DAC/headphone output for playback. You can override it.

2. Capture and rendering options

Headphone name Start frequency (Hz) End frequency (Hz) Sweep duration (s) Wait before sweep (ms) Wait after sweep (s) Noise-floor capture (s) Output level (dBFS) Optional center repeats Input channel Output channels Graph smoothing

Use a timing lead-in/referenceThe built-in matched sweep still deconvolves without it; the lead-in makes the capture timing more robust.

Subtract the measured noise floor from the responseOff by default, as in REW. A steady tone in the room (a fan, a transformer) can be over-subtracted into a narrow notch; the noise floor is still measured and reported either way.

Input calibration file — REW microphone/interface calibration, frequency + dB columns. It is tied to the selected input device and channel and comes back whenever they are selected again, even though this box reads “No file chosen”.

Remember this calibration on this browserNo calibration loaded

3. Target and preference bounds

Display mode Target curve

Your own rig target (frequency + dB columns) — uploading selects and remembers it. A remembered file stays in use even though this box reads “No file chosen”.

Normalisation Custom alignment (Hz) — or shift-click the graph Target tilt (dB/octave) Custom tilt (dB/octave)

Preference bounds Flat tolerance (± dB)

No target loaded

4. Positional workflow

Keep the fixture, signal level, and headphone seal unchanged. For each offset, move the headphone only in the indicated direction, then let it settle before capturing. Center is averaged when the selected repeat count is complete.

Measurement pinna mounted at the centre of a circular fixture with concentric offset rings — The outlined earcup marks the centred seating over the pinna. The fixture's concentric rings give repeatable offset distances — pick one ring and use the same one for all four offsets.

Positional response

Plot Seatings All-seatings average Spread band Difference from centre

Frequency rangedB spanDrag to pan · double-click resets · shift-click aligns the target there

5. Export and import

Load a project (.json)

REW export writes one text file per captured position, zipped together — REW imports one measurement per file (File › Import › Import frequency response). A single capture downloads as a plain .txt.
The project JSON keeps every captured response, calibration name, target, and capture settings. Load it again here to continue: positions in the file are added to whatever this session already has, and the file's centre counts as complete until you re-measure it.