Watch Timegrapher

A browser-based mechanical watch accuracy analyzer. Hold your watch up to your microphone and get professional timegrapher metrics — rate, beat error, and amplitude — with no app to install and no data leaving your device.

---

What It Does

Watch Timegrapher listens to the tick-tock of a mechanical watch through your microphone and runs a full signal processing pipeline entirely in the browser to produce three key metrics:

• Rate (s/day) — how many seconds per day the watch gains or loses. ±10 s/d is COSC certified; ±20 s/d is typical for a regulated movement.
• Beat Error (ms) — the asymmetry between the tick and tock intervals. Under 0.5 ms is excellent; over 1 ms may need regulation.
• Amplitude (°) — the arc of the balance wheel's swing. 250–310° is ideal; below 200° usually indicates low power reserve or worn parts.

It also renders two diagnostic charts after each capture:

• Averaged Waveform — a noise-reduced view of a single beat cycle
• Beat Timeline — a scatter plot of interval stability across the full sample

---

How It's Built

The app is a pure static web app — HTML, CSS, and vanilla JavaScript — hosted on Firebase. There is no backend server. All audio capture and signal processing happens client-side using the browser's Web Audio API.

Signal Processing Pipeline

1. Capture — the browser records a 6-second clip via MediaRecorder
2. Bandpass filter — a 4th-order biquad filter (2000–8000 Hz) strips low-frequency hum and high-frequency hiss using OfflineAudioContext
3. Envelope — the absolute value of the filtered signal is smoothed with a moving average to produce clean, unambiguous peaks
4. Autocorrelation — a custom Cooley-Tukey radix-2 FFT (no library) implements the Wiener-Khinchin theorem to find the beat period efficiently
5. Sub-sample refinement — quadratic interpolation on the autocorrelation peak gives floating-point precision on the period, essential for accurate rate
6. Metrics — rate from period ratio, beat error from alternating interval groups, amplitude via an arcsin formula using the lift angle

Tech Stack

	Technology
UI	HTML5, CSS3, Vanilla JS
Charts	Chart.js
Audio	Web Audio API (OfflineAudioContext, MediaRecorder)
Hosting	Firebase Hosting (static)

---

Using the App

Prerequisites

• A modern browser (Chrome or Edge recommended for best Web Audio API support)
• A microphone — a phone mic works, but a USB mic or lapel mic near the watch gives cleaner results
• A mechanical watch

Steps

1. Open the app in your browser
2. Set your movement parameters in the Parameters panel:
   • Search for your caliber (e.g. NH35, ETA 2824-2) — BPH and lift angle will auto-fill
   • Or set BPH and lift angle manually if your caliber isn't listed
3. Click "Enable Microphone" and grant mic access when prompted
4. Position your watch close to the microphone in a quiet environment
5. Watch the live waveform — you should see a clear repeating pattern. Adjust the Gain slider if the signal is too quiet or clipping
6. Click "Capture 6s Sample" — a 6-second recording is taken and analyzed automatically
7. Results appear in the Results panel. Color coding indicates quality: green (good), yellow (marginal), red (needs attention)
8. Click "Copy Results" to copy a text summary to your clipboard
9. Enable Continuous Mode to keep capturing automatically for ongoing monitoring

Tips for Clean Results

• Hold the watch case directly against the mic or use a folded cloth to dampen ambient noise
• A quiet room makes a significant difference
• If you see a "Weak signal" warning, move the watch closer or increase gain
• Dial position (crown up, crown left, etc.) affects amplitude — note your position for consistent comparisons

---

File Structure

/
├── static/
│   ├── index.html       # App shell
│   ├── script.js        # All signal processing and UI logic
│   ├── style.css        # Styling
│   └── movements.js     # Movement database (BPH + lift angles)
├── firebase.json        # Firebase hosting config
└── README.md

---

License

MIT