LTC Timecode Generator

Open the app — works on any phone or computer, installable as offline PWA.

A browser-based SMPTE LTC (Linear Timecode) generator. Designed to sync audio recorders (Tascam, Zoom, etc.) from a phone or computer — no dedicated hardware generator needed.

How it works

The app generates a standard SMPTE LTC audio signal in real-time using the Web Audio API. The timecode is encoded as a biphase mark modulated waveform — the same format used by professional timecode generators.

1. The current wall-clock time is captured and converted to a frame count
2. Each frame is encoded as 80 LTC bits using the SMPTE bit mapping (BCD-encoded hours, minutes, seconds, frames + 16-bit sync word)
3. Bits are converted to audio using biphase mark encoding (transition at every bit boundary, additional mid-bit transition for 1s)
4. Audio is streamed continuously at 48 kHz mono

The signal is generated sample-by-sample on the audio thread (AudioWorklet on desktop, ScriptProcessorNode fallback on iOS) ensuring frame-accurate output with no gaps.

Usage

Install as a PWA ("Add to Home Screen") for offline use. To self-host, serve the docs/ folder over HTTPS.

Setup

1. Connect your phone/computer audio output to your recorder's timecode input (TC IN / EXT IN)
   • Phone: use a USB-C DAC — the app outputs through whatever audio device the OS routes to
   • Computer: use any audio interface or headphone output
2. Open the app and configure:
   • Frame rate — must match your recorder's setting exactly (23.976, 24, 25, 29.97 DF, 30)
   • Time zone — defaults to your device's local timezone
   • Volume — start around 60%, adjust if the recorder won't lock (too loud = distortion, too quiet = no sync)
3. Tap Start
4. Your recorder should show "TC IN", "EXT LOCK", or similar within a few seconds

The timecode display runs continuously regardless of whether audio output is active.

Accuracy and limitations

Frame-to-frame timing is sample-perfect. Once running, the audio is clocked by your DAC's crystal oscillator. There are no dropped or repeated frames.

The initial timecode is seeded from your device's system clock, which is typically accurate to a few milliseconds via NTP. This is a one-time offset — it does not affect ongoing frame accuracy.

Long-term drift vs. wall clock: The DAC's oscillator and your system clock will drift relative to each other over time (~20-100 ppm, or roughly 0.1-0.5 seconds per hour). This does not affect sync between devices reading the same LTC stream — they all lock to the same signal. It only means the displayed timecode may gradually diverge from your wall clock over very long sessions.

This is not a precision time source. Do not rely on it for applications requiring sub-millisecond wall-clock accuracy. For syncing multiple recorders to each other during a shoot, it works well. For syncing to an external reference clock, use dedicated hardware.

iOS audio: iOS Safari has limited Web Audio support. The app uses a ScriptProcessorNode fallback and plays a silent audio element to bypass the hardware mute switch. If you hear no sound, check that your volume is up and try toggling the silent switch.

Python version

The original Python script (timecode_gen.py) generates LTC via sounddevice and numpy. It runs from the command line and requires a Python environment:

pip install sounddevice numpy
python timecode_gen.py

Configuration is hardcoded at the top of the file (sample rate, frame rate, volume, hour offset).

Files

docs/               PWA (served by GitHub Pages)
  index.html         UI + audio engine
  ltc-worklet.js     AudioWorklet processor
  sw.js              Service worker (offline caching)
  manifest.json      PWA manifest
  icon-*.png         App icons

timecode_gen.py      Original Python implementation

License

Do whatever you want with it.