Video Synchronisation: Genlock, Timecode and Drift.

In a multicamera production, each camera acts as an independent clock. Although they all record the same scene, each sensor exposes frames at the moment determined by its own internal oscillator, which does not exactly coincide with the neighbouring camera. On a switcher, that offset shows immediately: jumping cuts, trembling mixes and shots that never quite fit together.

The problem worsens when video and audio arrive on separate paths. The sound desk works with its own sample clock and video with another; without a shared reference, drift appears within minutes and nobody knows for certain which audio track matches which shot. This guide orders the concepts resolving that disorder: genlock, timecode, LTC, VITC, black burst, tri-level sync, frame sync, sync converters, unsynchronised cameras, audio clock versus video sync, drift and post-production synchronisation.

What Synchronising Video Means

Synchronisation is more than two images looking alike. It means all equipment in the chain shares a time reference and each frame is generated at the same logical instant. Distinguish three layers: scan reference or genlock, numbering reference or timecode, and the system’s audio reference.

The first prevents flicker and tearing when cameras connect to a switcher. The second locates each frame later in editing. The third keeps audio aligned with the image. A failure in any layer causes different symptoms; confusing them sends you looking in the wrong place.

Consider the entire chain: camera, transport, switcher, recorder and encoder. Every step can break the reference if equipment is not locked or a converter does not respect the signal. Design synchronisation before setup rather than improvising on event day.

Genlock: The Synchronisation Reference

Genlock locks a camera or device to an external reference signal. Once a camera receives genlock, it abandons its free-running oscillator and generates frames following the common reference. All cameras locked to the same source scan simultaneously, letting the switcher cut between them without an image jump at the transition.

A small production with few cameras and a simple switcher sometimes operates without genlock, relying on each input’s resynchronisation. But every correction introduces a small latency and a risk of repeated or dropped frames. For regular live signal mixing, genlock becomes the installation’s foundation rather than a luxury.

Timecode, LTC and VITC

Timecode labels every frame with hours, minutes, seconds and a frame number. It allows two separate recordings to align afterwards without relying on hearing or a handclap. In live multicamera work, timecode is distributed from a master generator to all cameras and recorders, so every file starts with the same time reference.

LTC, or linear timecode, travels as a modulated audio signal, usually through an XLR connector or unused audio channel. It is robust and easy to distribute, making it useful for recorders and cameras accepting an analogue timecode input. Its main limitation is needing a dedicated channel and not travelling inside video.

VITC, by contrast, is embedded in the vertical lines of the video’s blanking interval. The concept explains why some formats allow reading timecode directly from the recorded image, even in slow motion or on a frozen frame: the label travels with the signal rather than relying on a separate track.

In practice, most modern equipment combines both: distribute LTC or network timecode for recorders and let some devices read the reference from the SDI signal itself. Choose one master source and avoid mixing generators within the same installation.

Black Burst and Tri-Level Sync

Before digital references, reference signals were built from analogue pulses. Black burst is a black composite video signal with its sync information, used as a scan reference. It remains the usual reference in SD installations and much legacy equipment.

HD popularised tri-level sync, a three-level reference signal replacing the classic sync pulse and providing a more stable reference for progressive formats. Both serve the same practical purpose: giving each device a common heartbeat to lock its scanning to. If the installation mixes older and newer equipment, check which reference each accepts and use a converter where they differ.

Frame Sync and Sync Converters

Frame sync is a block that resynchronises an input signal to the local reference. Placed at a switcher input or an upstream stage, it makes an unlocked camera appear locked. Its cost is one or several frames of delay; with many sources, accumulated delay can become noticeable in video returns and operator communication.

Sync converters generate, distribute or convert those references: black burst to tri-level sync, one genlock input to several outputs, or one timecode source to multiple formats. Correctly configured, they let a mixed installation share one timebase; incorrectly configured, they become a silent failure point only noticed when something loses alignment mid-broadcast.

Genlock, Timecode and Drift

Separate two frequently confused problems. First is scan offset: unsynchronised cameras do not share a reference, causing jumps on cuts or moving lines in the image. Second is drift: even with common timecode, two clocks are never identical, and their accumulated difference grows over time.

A typical drift case is one camera recording independently to a card while the others are genlocked to the switcher. When combining files, the first minutes match, but several frames separate them by the end of the set. No particular device has failed: two free-running clocks have simply run in parallel.

Next, decide which reference to use at every point in the chain and how to verify that it remains valid at the end of the show.

Audio Clock vs. Video Sync

Video and audio count different things. Video counts frames at a nominal rate, such as 25 or 50 per second. Audio counts samples, tens of thousands per second. Even with both clocks set to nominal values, a real oscillator deviates by a few parts per million; that tiny difference becomes milliseconds after an hour.

When the audio clock and video sync are not referenced to each other, the offset grows continuously. Over a ten-minute take it may be inaudible; over a two-hour session it becomes visible in a speaking close-up. The usual solution is to slave audio to the video reference through word clock or an audio network, or set the sound interface’s clock to the house reference.

Measurement is the only way to know whether the system is aligned. A clapperboard at the start or a shot of applause remains the quickest check of image/sound alignment before accepting a recording.

Post-Production Synchronisation

Recording with common timecode does not guarantee automatic editing, but reduces the work to checking and confirming. In post-production, import files, group them by camera and align them using embedded timecode. If some cameras lacked timecode, synchronise by audio or a common visual marker; that work multiplies with every additional angle.

Post-production synchronisation also corrects what live production could not: small drifts, one-frame offsets and audio differences. Document which camera had a reference, which files share timecode and which were aligned manually, because that saves hours when returning to the project weeks later. Details of recording and assembling those tracks are in Multicamera Recording and Post-Production.

Next Step

With cameras locked and timecode organised, the next problem is getting every signal to the switcher without breaking its reference en route. Video Transport: SDI, HDMI, NDI, SRT and Fibre reviews distances, converters and matrices for choosing transport on each installation segment. If you have not yet chosen your cameras, start with Cameras for Live Production and return here when it is time to match and lock them.