6 min read
Separate audio from video: the assistant editor's first job
Separating audio from picture is not a conversion trick. It is the step that turns rushes into something a cutting room can work with.
Picture and sound travel together in a camera file and apart everywhere else. Sound is recorded on a separate device, edited on separate tracks, mixed in a separate room and delivered as a separate set of stems. Separating audio from video is the moment that pipeline starts.
What separation actually means
It means pulling the audio stream out of the container without touching the video. With a stream copy nothing is decoded at all; with a browser tool the audio is decoded and written back as uncompressed WAV, which adds no lossy generation. Either way the picture is never re-rendered and the operation takes seconds even on long files.
Why the cutting room needs it
- Sync: the camera scratch track is the reference for aligning recorder files.
- ADR: a clean line pulled from the cut becomes the guide track for the session.
- Music and effects: temp material has to live on its own tracks to be replaced later.
- Transcription: a separate audio file feeds transcription tools without shipping the picture.
Keep the relationship documented
The moment audio and picture are separate files, the only thing linking them is naming and timecode. Keep the source clip name in the audio filename, keep the start timecode intact, and never renumber takes after separation.
Multichannel camera audio
Cameras that record two or four channels write them interleaved. After separation, split them to mono stems so the boom, the mix and any wireless feed each land on their own track rather than arriving as a locked stereo clip.