Timecode, genlock & word clock

Labels, pictures and samples.

A matching timecode display is useful evidence that devices agree on a label. It does not, by itself, establish that their sensors expose together or that their audio samples run at the same rate. A production sync plan assigns those jobs explicitly.

At 48 kHz, the nominal sample interval is 1/48000 second, about 20.833 microseconds. At 24 fps, a picture interval is about 41.667 milliseconds. They describe different timing grids. A shared reference system must establish their relationship where the workflow requires it.

Ambient’s sync explanation distinguishes matching a timecode value from synchronizing clock speed. Sound Devices’ 788T reference gives a concrete recorder example: its word-clock input controls sample rate, with a separate indication of lock.

Jamming sets a starting relationship.

With a conventional jam, a device takes a timecode value from a source, then continues on its own generator after disconnection. Its subsequent agreement depends on the clocks and the device’s implementation. A jam is therefore an event to record and verify; it is not a universal guarantee of continuous rate or phase lock.

A continuous external feed keeps a reference available, but the receiving device must be configured to use it for the intended purpose. Receiving LTC for timestamping and deriving sensor timing from LTC are different capabilities. Check the manual and status indications, including behaviour when reference disappears.

Free-run and record-run describe generator behaviour. Free-run continues between takes; record-run advances while recording. Neither term alone confirms genlock. For separately started cameras and sound, establish the intended timecode mode before the shoot rather than assuming every record-run generator shares the same timeline.

Ambient describes the limitations of timestamp-only camera operation: recording can begin at a different point within a frame interval, and independent recording clocks can drift. The practical question is what the particular camera does with its input.

Put a clock error into production units.

Suppose a hypothetical clock has a constant rate error of one part per million relative to a reference. Over a 3,600-second take:

That is 3.6 milliseconds. Two independent clocks at opposite errors of +1 ppm and −1 ppm can differ by 2 ppm: 7.2 milliseconds, or 0.1728 of a 24 fps frame over the same hour. At a nominal 48 kHz, 3.6 milliseconds corresponds to 172.8 sample intervals; this is a timing equivalent, not a claim that a file contains fractional samples.

Rounding 0.0864 to zero frames hides the measured interval; it does not correct the clock. Changing a clip’s start label can fix a constant editorial offset, but it does not change when a sensor exposed or prevent an increasing difference during a take. Test both the beginning and end of a representative long recording.

Check what the camera actually locks.

A blanket rule that LTC can never synchronize a camera would be wrong. ARRI’s Synchronization of ARRI Cameras with 3rd Party Devices describes configurations in which incoming LTC triggers sensor timing when the sensor rate matches the project rate. It also documents genlock options and model-specific conditions, including a service modification for genlock input on ALEXA Mini.

Use that as a reason to read the exact camera documentation, not as a setup recipe for another model. Confirm model, software version, input, project rate, sensor rate, supported reference format and lock mode. ARRI recommends tri-level synchronization for improved accuracy in its described setup. Integer 24 and fractional 24000/1001 remain distinct choices.

For an LED volume or VFX capture, common rate and phase reference still leave exposure timing, rolling shutter, display scan, processing delay and acceptable phase adjustment to be checked on the actual system. Matching labels cannot certify an artefact-free photographed wall.

Make the handoff testable.

1. Write the timing specification: exact picture rate, scan type, DF/NDF mode, sound sample rate and any off-speed capture.
2. Name the reference: identify the master and which signal controls each camera, recorder, display processor or capture device.
3. Verify the path: use each manufacturer’s requirements for supported signals, cables, distribution and termination; inspect lock status rather than only matching displays.
4. Record the operating plan: timecode mode, jam/check schedule, power changes and behaviour on lost reference.
5. Shoot a representative test: include the longest expected take and demanding LED or multicamera conditions. Check picture phase, sound alignment and end-of-take drift.
6. Carry evidence into post: preserve original metadata, reference recordings, known offsets and test results. Agree any correction with editorial, sound and VFX.

A clap or flash can help measure alignment in recorded media. It does not continuously discipline the clocks. Define the acceptable error for the job and test against it.

Calculate offsets, then verify the recording.

Timecode Pro can help calculate frame offsets and durations for production notes and post handoffs. Keep the rate and DF/NDF convention attached to every label; see drop-frame timecode and frame-rate conversion for those distinctions.

The app performs arithmetic. It is not a sync hardware generator, camera controller or NLE, and it does not measure genlock phase, discipline sample clocks or repair drifting media.