Updated October 2026
Of the three numbers this tool reports, amplitude is the one most worth reading the fine print on. Rate and beat error are genuinely useful estimates derived from real timing math. Amplitude is explicitly labeled "Experimental" in the results for a specific, honest reason explained here.
In watchmaking, amplitude means the number of degrees the balance wheel swings from its rest position to the extreme of each oscillation — a genuine angular measurement, typically somewhere in the range of 200-320 degrees for a healthy watch depending on position and power reserve. A low amplitude (well under roughly 200 degrees) is a real, clinically meaningful signal of a problem: weak mainspring power, excessive friction, or a lubrication issue.
Dedicated timegrapher machines use a calibrated microphone or piezo sensor physically clamped to the watch case in a controlled fixture, with known acoustic coupling and a fixed, repeatable distance and contact pressure. The relationship between the sound's timing characteristics (not just loudness — the shape and timing of the tick's acoustic signature) and actual angular amplitude has been established and calibrated against that specific, controlled setup. That calibration is what turns an acoustic signal into a trustworthy degree measurement.
This tool uses whatever microphone happens to be built into your device, picking up sound through open air at a distance and angle that varies every single time you use it — closer one test, farther the next, a slightly different angle, a different phone or laptop model with different microphone sensitivity entirely. There is no fixed, known relationship between "how loud this particular tick sounded to this particular microphone, held this particular way" and "how many degrees the balance wheel actually swung." Any number claiming to be a precise angular-degree reading under these conditions would be fabricated precision, not a real measurement — which is exactly what this tool deliberately avoids doing.
Rather than inventing a fake degree number, the amplitude metric here is a relative loudness percentage: every detected tick's amplitude in a capture session is compared only against the single loudest tick detected in that same session, and shown as a percentage of that session's own maximum. This makes it genuinely useful for one specific comparison — louder versus quieter within the same recording, for example noticing that ticks faded somewhat partway through a capture, or comparing one position against another back-to-back in the same sitting with the same mic placement. It is not useful, and not intended, as an absolute number you record today and expect to compare directly against a different recording next week.
Because the percentage is normalized against each session's own maximum, two separate captures of the exact same watch, in the exact same position, could show different amplitude percentages for reasons that have nothing to do with the watch — slightly different mic distance, a slightly louder or quieter individual tick happening to set this session's ceiling. This is why testing before and after a service specifically warns against comparing amplitude percentages across two different sessions, even when rate and beat error from those same two sessions remain meaningfully comparable.
If amplitude specifically — not rate, not beat error — is the number you need to trust (for example, diagnosing a suspected mainspring or lubrication issue, or documenting a professional service), that calls for an actual calibrated timegrapher machine or a watchmaker's bench equipment, not this or any other open-air microphone tool. See how to tell if your watch needs service for how to weigh this tool's numbers against that limitation.
Why show amplitude at all if it can't be trusted? The relative, within-session comparison is genuinely informative for the specific uses described above — the goal here is honest labeling of what it can and can't tell you, not omitting a genuinely useful (if limited) signal.
Does a higher percentage always mean a healthier watch? Not necessarily, and not across sessions — within one session, a louder relative reading in one position versus another can be a meaningful comparison; across different sessions it mostly isn't, for the reasons above.
Will this tool ever add a calibrated amplitude mode? Not with a standard phone or laptop microphone — true calibration would require known, fixed hardware, which is exactly the constraint a browser-based, any-device tool can't satisfy.