Everyone knows how annoying it is when a system has an AV sync problem.

So how should I test a system?

A simple test would be to play content known to be in sync and measure it.  The content will be a lots of individual beeps and flashes perfectly in sync.  Ideally let the test file play and measure the error of a number of beeps and flashes.  Ideally at least 5 but 10 or more is better.

Then look at the stats of these readings.  The span (difference between the largest and smallest error) should be 0 in a ideal system.  However, outside of a post production facility that may be hard to achieve.  Next check the average of the readings taken, this will be the error in the system.  Use this average as the correction in the system and repeat the test.  You should now have an average close to, or actually, 0.  For that test file and system configuration your done.  Naturally if you can run in multiple frame rates, you will need to test those too as needed.

Why take the average? Most systems will have some jitter in them, just measuring a single beep flash will not be a true representation.  So to account for the jitter a number of beep flash readings are taken then averaged.

Just Eyeball it?

Some still try to get accurate sync doing this, even though it’s prone to a multitude of errors.  The largest of all being that the test is entirely the subjective view of the person doing the testing.  It also wastes time.  Guessing the error then trying adjustments in the hope of getting the ideal number.  It isn’t accurate, but it is cheap.  If you don’t factor in the time of the person and the system under test.

After all the time spent tinkering, someone else comes in and says “I thought you going to sort the sync out in here?”.

Use a Mobile App?

There are a few, and they can get you close if you have the patience to use them.  Using a smartphone will always introduce an unknown element into the measurements, as the app cannot directly interact with the hardware.  There is all sorts of processing going on in the middle the apps have no control over.  So the readings derived will have an unknown margin of error.

It can take a while to scrub along to a number of the beeps and flashes taken during the video of the test file playing. Then note the approximate error and manually calculate the average.  Using just one of the beep flash measurements would be fine, if the span of the readings were 0.  But you don’t know that, until you measure them all and do the maths.  So you have to always do them all just to get an approximate reading.

No doubt about it, apps are cheap.  Providing you gloss over the cost of the smartphone they are running on and the time it takes to test a system.  You did check the correction factor is correct for you device and OS version, didn’t you?

Use a Sync-One2

If you want to be accurate and not waste time, use a Sync-One2.  It only takes a few minutes to test, correct, and re-test a system with Sync-One2.  Readings are given in real time and the stats are automatically calculated for you.

Sync-One2 is used by those who want to know the true sync of a system, and don’t want to spend a lot of time working it out.  It is the tool of choice of professionals around the world.

Full details of Sync-One2 are available from the home page.