How we trace a lip-sync fault from microphone to playout

Starting with the complaint, not the equipment

Most audio synchronisation jobs begin with a sentence such as the presenter looks slightly late on the broadcast feed. Before touching any equipment, I ask for the details around that sentence: which output shows the fault, whether it appears on every programme or only on some, and whether it changes after a restart. A fault that moves around points to a different cause from one that stays fixed.

I also ask for a signal flow drawing, even a rough one on paper. Studios often have more devices in the audio path than anyone remembers: a wireless receiver, a digital mixer, an embedded audio stage, a frame synchroniser, a playout server. Each can add a small delay on its own, and the total only matters once it passes what viewers notice.

Only after this conversation do we agree what to measure and where.

• Which output shows the fault
• Whether it is constant or drifts
• The last change made before it began

The measurements we take on site

We measure rather than judge by ear. A test signal with a sharp flash and matching click goes through the same path as live material, and we compare picture and sound at the output with a dual-trace display. This gives the offset in milliseconds and tells us whether the sound leads or lags the picture.

We then repeat the measurement at several points along the chain, working backwards from the faulty output. The point where the offset appears is where the cause lives. It is slow, methodical work, and it is the main reason a written report can name one device instead of suggesting a general problem.

On a compact studio this usually takes a morning. Larger installations with several routes may need a second visit so we can check each route under normal load.

• Test signal with flash and click
• Offset logged at each stage
• Same path as live programme material

Where the delay usually hides

In our experience the cause is rarely a broken device. More often a processing stage was added or reconfigured without anyone compensating for the extra latency. A video frame store, an upconverter or a conferencing link can hold the picture back by one or two frames while the sound continues at normal speed.

Digital audio has its own latency too. Sample rate conversion, effects and some wireless links each add a small amount, which is harmless alone but noticeable when several are stacked. Clock settings matter as well: if devices are not locked to one reference, the offset slowly drifts, and operators find themselves nudging a delay control every few days.

Knowing which of these applies decides whether the remedy is a setting, a re-route or a small equipment change.

• Added video processing holding the picture back
• Stacked audio latency across several devices
• Devices running without a shared clock reference

Timecode and the problem of agreeing on one clock

Timecode causes quieter faults. If the camera, the recorder and the playout system each carry their own idea of the time, material that looked aligned in the studio arrives out of step in the edit or on air. We check that every source takes its timecode from one generator and that the frame rate and drop-frame settings match everywhere.

A surprising number of problems come down to one unit set to a different frame rate from the rest, or running free after a power cut. We write down the settings of every device so the studio has a reference page, not just a repaired system.

Where a studio records isolated audio alongside video, we also check that file start times and stamps agree, since this is what lets an editor line the two up quickly.

• One timecode generator for all sources
• Matching frame rate and drop-frame settings
• Isolated audio files stamped consistently

Making the fix and proving it stays fixed

Once the cause is clear, the correction is usually small: a delay value on the mixer, a corrected clock source, a changed route. We apply it with the studio's operator beside us so they see what changed and why. Nothing is altered without a note of the previous setting, so it can be restored.

We then repeat the test measurements and run a normal programme through the path, leaving the system on air for a while before we call it done. A fix that holds for ten minutes does not tell us much; one that survives a full rehearsal and a restart tells us more.

The operator keeps the before and after readings, which turns a vague complaint into something they can compare against in future.

• Previous settings noted before any change
• Measurements repeated after the fix
• A full rehearsal and a restart as the final check

What the studio receives afterwards

The written summary is short and plain. It states the offset we found, where it came from, what we changed and what the studio should check if the symptom returns. We avoid long equipment lists and keep to what the crew needs on a busy production day.

We also suggest a simple routine, such as a quick flash and click test at the start of each week, so a new fault is caught early. Studios that adopt it tend to call us for planning rather than for emergencies.

If the report points to a larger change, such as replacing a link or reworking a route, we say so and leave the decision with the studio.