Chapter 01The room matters more than the phone.
A mechanical watch's tick reaches the phone microphone roughly 50 dB above the room's noise floor. That's a comfortable margin, but fragile: any continuous sound that lives in the same frequency band as the tick (typically 2–8 kHz) raises the floor and partially masks the beat. WatchScope's peak-tracker loses precision because every peak becomes a flatter hill, and rate estimation gets noisier.
You don't need an anechoic chamber: a closed room, a few placement tricks and an awareness of which noises are truly enemies are enough.
Noises to silence before you start
- Fans — PC, console, air conditioning. They are the worst offenders because their noise is broadband and steadily covers the tick's band. Switch them off or leave the room.
- Fridge / heat-pump compressors. They cycle on and off: even if the room feels quiet when you begin, the compressor may kick in mid-test and ruin 20 seconds of measurement.
- Street traffic and children. They are impulsive (non-stationary): they spoil single samples but WatchScope rejects them as outliers. Still, avoid them on COSC tests where every s/d counts.
- Music, TV, radio. Even at low volume, speech and instruments sit in the tick's band. Total silence during the session.
Placement and resting surface
Rest the watch flat on the phone, with the caseback about 1 cm from the microphone (usually on the lower edge of the device). Avoid glass or metal tables: they amplify the vibration of a fridge compressor 3–4 metres away and feed it into the microphone as structural noise. A felt mat, a folded tie or a mouse pad damp the mechanical transmission and noticeably improve the signal-to-noise ratio.
Before launching the real test, run a 5-second quick test and watch the oscillogram on the phone. If you see regular patterns other than the tick (e.g. a continuous 50 or 60 Hz buzz, or periodic thumps from a compressor), find the source before launching the actual test. You'll save 30 seconds and land a clean session.
Chapter 02Hardware calibration: the correction factor.
Every smartphone microphone uses a MEMS crystal as the clock reference for audio sampling. That crystal has a small temporal drift — typically a few parts per million (ppm), corresponding to fractions of a second per day — relative to the system's main clock. For a music app it doesn't matter; for a timegrapher measuring rates a few s/d wide it matters a lot.
WatchScope solves this with a calibration procedure: from Profile › Calibrate timegrapher, the app records 3 minutes of silence, compares the actual duration of the audio samples against the kernel clock, and produces a correction factor (typically a number between 0.9995 and 1.0005). The factor is then applied to every subsequent reading: where you used to see a constant 2–4 s/d offset, accuracy now drops below 0.5 s/d.
When to recalibrate
- Phone change. The factor is specific to the MEMS crystal — a new device must be calibrated from scratch, even if it's the same model as the previous one.
- External microphone change. If you switch to the piezo (see chapter 03) or change USB-C / jack adapter, recalibrate: the factor is tied to the whole acquisition chain, not just the phone.
- After a major Android firmware update. Rare, but some vendors touch the audio drivers in major releases and the factor can drift slightly.
Recalibration takes 3 minutes of total silence — less time than a 5-position COSC test. Do it when the room is really still; a cough halfway through and you have to start over.
On a 60-second quick test of an in-spec watch (±5 s/d), skipping calibration introduces a systematic error that can double the measured value. That's the same gap that separates a hobbyist timegrapher from a watchmaker's bench instrument.
Chapter 03The external piezoelectric contact mic.
The phone microphone captures the tick through the air. It works, but with two structural limits: the signal arrives attenuated (the caseback is steel, air is an inefficient medium) and it mixes with any ambient noise on the same band. A piezoelectric contact microphone — the classic contact mic luthiers use to amplify acoustic violins and guitars — solves both problems: it sits directly on the case and picks up the mechanical vibration of the escapement bypassing the air.
Result: signal-to-noise ratio up to 10× higher than the phone microphone. Stable readings even in noisy environments (workshops, fairs, sales floors). And a unique use case: measuring watches under a closed glass cloche, where the air is isolated but the piezo transmits the tick from the case to the phone through the glass itself.
How to pick one
The market of contact mics for musical instruments offers suitable products starting at 15–30 euros. Criteria:
- Piezoelectric pickup in a circular platform 10–20 mm wide. Classical-guitar models work great: a watch caseback is roughly the same size as an acoustic-guitar pickup hole.
- 3.5 mm TRS jack connection for phones with a headphone jack, or USB-C / Lightning ↔ jack adapter for newer devices. Verify that the adapter supports mic input (not just headphone output).
- Non-permanent adhesive or coupling gel to fix the piezo to the case without leaving residue. Tac putty ("poster putty") works perfectly and is fully reversible.
After connecting the piezo, always run a fresh Timegrapher calibration — the correction factor changes because you're using a different acquisition device. Then start the quick test as usual: you'll immediately notice the oscillogram is cleaner, with taller peaks and an almost flat noise floor.
Chapter 04Checklist for a bench-grade measurement.
Before launching any important test, scan this list. Four items, sixty seconds.
Still room
Closed room, fans / AC off, no music or TV. Fridge in eco mode if possible.
Damping surface
Watch on felt or fabric, not on glass / metal. Microphone distance ≈ 1 cm.
Recent calibration
Correction factor applied. If you changed phone or microphone, recalibrate before starting.
Adequate wind
The watch should have been wound at least 8 hours: end-of-power-reserve measurements underestimate amplitude by 20–30°.
Three minutes of calibration,
a lifetime of better readings.
Download WatchScope, run the timegrapher calibration once and every subsequent test will inherit the correction factor. Free on Android.