Audio Sync Checks to Make Before the Winter Schedule Starts

Why autumn is the right time to look

Most studios I speak to share the same pattern. The summer is spent on maintenance, a few firmware updates and the odd rerouted cable, and then the schedule fills with winter entertainment, election-style specials and the long run of live evening shows. A sync fault that was tolerable in a quiet month becomes a complaint once the audience grows.

This is the second of our notes on audio sync audits, and it deals with timing rather than method. The first note covered how to trace a fault. Here the question is when to do the work: before the heavy weeks, while the gallery can still be taken offline for an afternoon, and while any change can be tested without a live programme depending on it.

I would rather a studio spent one calm session now than several rushed ones in December. Changes made in a hurry tend to be undocumented, and undocumented changes are the usual cause of the next fault.

What has changed since the spring

Sync problems rarely arrive from nowhere. They follow a change: a new frame synchroniser, a replacement monitor with a different processing delay, a software update on a graphics machine, or a remote contributor added over a new link. Begin by listing everything that has been touched since the last time the chain was known to be good.

Ask the people who work the room, not only the people who maintain it. A vision mixer will often say that a particular camera feels late, or that the guest on the remote feed always looks slightly ahead of the voice. Those remarks are worth writing down verbatim, because they point to a source before any equipment is plugged in.

Keep the list short and dated. If nothing has changed and the chain still drifts, that is information too, and it points towards a clock or reference problem rather than a single device.

• New or replaced displays and converters
• Software or firmware updates on any processing device
• Remote contributor links added or rerouted
• Changes to genlock or reference distribution

Measure the chain, not just the endpoints

A common shortcut is to compare the picture and sound at the final output and call the difference the fault. That tells you how large the error is but not where it lives. Measure at several points along the chain: after the cameras, after the vision mixer, after any graphics insertion and at the output feed. Each stage adds a small, usually fixed delay.

Use a test signal with a clear marker in both picture and sound, and record the offset at each point in milliseconds. A neat table of stage against offset will show immediately which device adds more than its share. It also gives you a baseline to return to when something changes in the winter.

Do the same for each source type separately. A studio camera, a playback server and a remote link will not share the same delay, and an average hides the one that causes the trouble.

Cold weather and long runs of live material

Autumn into winter changes the working conditions in small ways. Outside broadcast crews carry kit between cold vehicles and warm venues, and condensation and temperature swings can affect equipment that has been behaving all summer. If your chain includes outside sources, test them after they have settled to working temperature, not straight from the van.

Longer live runs matter as well. A fault that appears only after two hours of continuous operation, a slow drift between audio and video, is often missed in a short check. If you suspect drift, leave the test signal running for the full length of your longest programme and compare the offset at the start and the end.

Drift points to clocking, and clocking problems are cheaper to fix in a quiet week than during a transmission.

Writing the fixes down

The part of an audit that studios skip most often is the record. A fix that lives only in one engineer's memory is lost the moment that engineer is on leave. For each change, note what was altered, the measured offset before and after, and the setting to return to if something goes wrong.

Keep the document short enough that someone will actually open it in the middle of a show. One page of current delays per source, a note of any compensation applied and the name of whoever approved it is usually sufficient. Store it where the gallery staff can reach it without asking.

When we write up an audit for a client, this is the part they use most. The measurements are useful once; the record is useful for years.

• Offset per source in milliseconds, before and after
• Where compensation is applied and by how much
• Who approved the change and when

A sensible order for the next few weeks

If the time available is short, work in this order. First, list what has changed since the spring. Second, measure each source type through the chain and build the table. Third, correct the largest single offset and measure again before touching anything else. Changing several things at once makes it impossible to know which one helped.

Leave a margin for retesting after any update that arrives later in the season. Software updates have a habit of arriving just before a busy week, and a short repeat of your baseline measurements will tell you at once whether anything has shifted.

None of this needs special equipment beyond a decent test signal and a means of reading the offset. It needs an afternoon, a notebook and some patience.

Sync faults are easier to deal with before the schedule gets busy than during it. Use these mid-autumn weeks to list what has changed, measure each source through the chain, correct one thing at a time and write the result down where the gallery can find it. A short, dated record will serve you well through the winter and well beyond it.