Listening notes are useful, but they are not repeatable. So before a pair of headphones earns a listing we put it through a fixed sequence on the same rig, in the same room, in the same order. Nothing about the process changes between products — that is the whole point of it.

What the bench actually does

• Frequency response. Swept on a measurement fixture so the curve can be compared against another product on identical terms.
• Isolation. Measured passively with the electronics off, then again with active cancellation enabled, so the two effects are never conflated.
• Battery. Discharged at a fixed volume with a looped signal until cutoff, rather than repeating whatever the box claims.
• Bluetooth. Codec support is confirmed by connecting real devices, not by reading the spec sheet.
• Build. Hinges, headband adjusters and cable entry points are cycled, because that is where a pair of headphones usually dies first.

Why we publish the method

A number without a method behind it cannot be checked by anyone. Publishing the procedure means you can decide for yourself whether our test resembles how you actually listen — and whether our conclusion deserves your trust.

If a product does not survive the sequence, it does not get listed. That is the only editorial rule we have.