You're viewing an apartment and the agent swears the walls are quiet. You're choosing between sealing a window or replacing it. You built a music corner and want to know if the neighbors will hate you. In every case the question is the same — how much sound does this barrier actually block? — and the professional answer involves a calibrated speaker, a calibrated microphone, and a consultant's invoice.
Here's the thing though: the measurement doesn't fundamentally require calibrated anything. It requires two positions at once — and most households have a second phone.
The setup
One phone is the emitter: it sits on the far side of the wall and plays a stepped tone sweep — a sequence of pure tones marching through the frequency range, each announcing itself by simply being the only tone playing. The other phone is the measurer, and it captures the sweep twice:
- Reference capture — with a clear path to the emitter (same room, door open), record the sweep and note the level received in each frequency band.
- Through capture — with the barrier between you (door closed, or emitter in the next room), record the same sweep again.
The answer is the ratio between the two, per frequency band: how many decibels each band lost crossing the barrier. Low frequencies leaking through while highs are blocked is the classic signature of a lightweight wall — and exactly the pattern a single-number "soundproofing" rating hides.
The trick: only the same-receiver ratio is ever the answer
Why does this work with uncalibrated consumer hardware? Because everything that isn't the wall appears in both captures and divides out:
- The emitter's speaker is louder at some frequencies than others — identically in both captures. Cancels.
- The measuring phone's microphone has its own frequency response — the same one in both captures. Cancels.
- Absolute volume, distance conventions, phone model differences — all shared between the two captures. Cancels.
What's left is the one thing that differed between reference and through: the barrier. This is why the two phones need no clock synchronization, no cable, and no calibration against each other — the stimulus is self-identifying, so the measurer just listens for the tones, and the receiver compares only against itself.
The one rule that matters: both captures must happen on the same measuring phone. Swap phones between reference and through, and you've measured the difference between two microphones plus the wall, inseparably. The activity is structured so you never do this by accident.
The same idea, with light
Replace the speaker with a blinking screen and the microphone with a camera, and the identical logic measures light transmission: the emitting phone blinks its screen at a fixed 2 Hz, the measuring phone's camera watches for that blink directly (reference) and through the window, blind, or film (through). The blink rate is the self-identification — the camera isn't fooled by ambient light because ambient light doesn't pulse at exactly that rate. Same ratio, same cancellation, same absence of calibration.
Practical notes for a trustworthy reading
- Keep positions consistent. The ratio cancels the devices, not the geometry — measure reference and through from comparable distances.
- Mind the flanking paths. If sound reaches you around the wall (under a door, through a duct), you're measuring the room's weakest path, not the wall. That's often the more useful answer, but know which one you asked for.
- Repeat it. Two phones and three minutes means you can measure twice and check agreement — a luxury a consultant's single visit doesn't offer.
- Let one phone drive the other. With the optional device link, the phone in your hand tells the placed phone when to emit — so you're not running between rooms hitting play. The link is a convenience, never a requirement: the measurement itself stays link-free.
Why this shape of measurement matters
The two-phone ratio is the purest example of the design rule running through every tool on the platform: never build an answer on an absolute level. Absolute levels belong to the hardware; ratios belong to the world. The same principle powers the tap test's resonance features and the reflectance tools' torch-on/torch-off frame pairs — the tools overview maps them all.