Timecode sync for multicam shoots

When the gear is jammed, timecode is the best reference there is

If your cameras and recorder run on one clock, from Tentacle Sync boxes or another LTC source, timecode is the most precise thing you have, and CamHeard leans on it. On the jammed shoot in our own test material our placement agreed with the hardware timecode to a median of 40 ms across 158 clips, which is one frame at 25 fps.

So this page is not an argument against timecode. It is about what happens on the rest of the day, when the timecode in the files is not what it looks like.

A recorder counter often says position, not time of day

In our test material of 1121 files from real shoot days, 9 out of 10 files carry timecode, and more than half of those counters run only while the device records. A number like that tells you how long the previous take was, not when this take happened. In the file it looks exactly like a time of day clock.

A tool that simply trusts timecode will put that material in the wrong place, and you find out on the timeline. CamHeard tells the two kinds apart by itself, and you are never asked to pick a mode.

A jam is rarely the whole shoot

On our jammed test shoot, 8 devices went out for the day and 5 shared the clock. One camera was jammed at a different moment and sat at a constant offset. One lost its timecode when the power was cycled. One was never jammed at all.

That is the normal shape of a shoot with timecode: mostly jammed, not entirely. CamHeard takes timecode where it is real and sound where it is not, in the same pass and in the same project.

Timecode as a hint, sound as the check

CamHeard takes the timecode where it is real and the sound of the day where it is not, in the same run. When the two agree, the clip lands tight. When they contradict each other, the sound wins.

This is also why you do not choose a sync method before the run. There is no rec run switch, no free run switch and no waveform switch to get wrong.

Cameras without timecode join the ones that have it

A shoot mixes bodies. Two cinema cameras were jammed, the third camera and the second unit never were, and the action camera has its own idea of the clock. All of it comes out as one timeline, in one run, without you sorting the cameras into groups first.

An LTC signal recorded onto an audio channel is read as timecode too, so a camera fed from a box that writes audio instead of metadata is covered.

What happens to a clip it cannot prove

A clip CamHeard cannot place with confidence is parked at the end of the timeline with a mark, or stood at an approximate place by the logic of the shoot if you prefer that, still marked. Nothing in the body of the timeline is presented as synced without proof.

Export goes to Premiere Pro and DaVinci Resolve as FCP7 XML and to Final Cut Pro as FCPXML. Multitrack recorder channels keep their own tracks, and camera sound stays with its camera. Everything runs locally and offline, with no account and no upload.

Frequently asked questions

Do I have to tell CamHeard whether my timecode is rec run or free run?

No. It works that out per device and uses the timecode only where it means time of day.

Must all clips carry identical timecode?

No. Partly jammed shoots, cameras jammed at different moments and cameras with no timecode can sit in one project.

Can I use CamHeard without any timecode gear?

Yes. With no usable timecode in the files, the sound decides, and that is the base case the app was built for.

What if a camera lost its timecode in the middle of the day?

Those clips are placed by sound with the rest. If there is nothing to place a clip by, it goes to the end with a mark.

Does CamHeard need Tentacle Sync or any other hardware?

No. Timecode from a jammed rig makes the result tighter, and nothing about the app depends on owning a box.

Which platforms does it run on?

macOS 15 (Sequoia) or later on Apple Silicon, and Windows 10 or 11, 64-bit.