2 February 2026 | 7 min read
How to sync audio and video: a practical guide for dual system sound
Three ways to sync, one right answer for each situation, and the preparation that makes all three faster.
Dual system sound means the good microphone and the good camera are recording to different devices. Someone has to marry them again. Here is how the three standard methods actually behave in production, and how to prepare files so any of them works on the first attempt.
Method 1: timecode sync
The camera and the recorder share a common clock, either jam synced from a master or locked wirelessly. Every file is stamped with the same time reference, and the NLE aligns thousands of clips automatically.
It is the fastest method by a wide margin and the standard on any professional set. It fails only when frame rates disagree or when a device drifts after a long unsynced stretch.
Method 2: waveform sync
Software compares the scratch audio recorded by the camera to the good audio from the recorder and aligns them by correlation. It is remarkably reliable as long as the camera track contains a recognisable version of the same sound.
It struggles in wind, in very reverberant rooms, when the camera mic is a long way from the subject, or when a scene is nearly silent. Keep a scratch track running even if it sounds terrible. Terrible is enough for correlation.
Method 3: clap sync
One sharp transient, matched by hand. Slow, but it never fails and needs no metadata. Always slate, even on timecode shoots. It is the insurance policy that costs one second per take.
Prepare the files first
Sync problems are usually preparation problems. Before importing anything, do three checks:
- Confirm the sound files are the sample rate and bit depth the project expects.
- Confirm the frame rate the recorder logged matches the camera and the timeline.
- Split poly WAVs into named mono stems if your NLE expects mono, and do it before the assembly, not after.
The last one saves the most time. Splitting a day of poly recordings into named boom and lav tracks up front means dialogue editing starts with a readable timeline instead of a wall of numbered channels.
A workflow that holds up
- Copy cards, verify checksums, keep the originals untouched.
- Inspect one file per card to confirm sample rate, bit depth and timecode.
- Split poly files into mono stems using the track names from the metadata.
- Import, sync by timecode, spot check three takes across the day.
- Fall back to waveform for any take the timecode missed, and clap sync anything left.
Everything in steps 2 and 3 can be done in this browser, locally, without uploading rushes anywhere or waiting for a render farm.