Working with video

erm was built for audio, but it handles video inputs too — either pulling the clean audio out, or rendering a tightened picture that stays in sync. This page is the user-facing guide; for the A/V-sync mechanism see the internals doc, video render & A/V sync.

Audio-only or keep the picture?

Feed erm a video file and by default you get the cleaned audio only (.wav) — the fast path when you just want the audio, and identical to how erm has always behaved. Add --video to render a synced video output (container inferred from the input):

You want…	Run	Result
Just the audio from a video	erm talk.mp4	talk-cleaned-*.wav, no picture
A tightened video, fillers cut	erm talk.mp4 --video	re-encoded video, crossfaded splices, A/V in sync
Same length, captions intact	erm talk.mp4 --video --mode silence	picture stream-copied (lossless), fillers muted in place
Hard jump cuts	erm talk.mp4 --video --video-splice cut	concat splices on both streams

Container and codecs

• The output container is inferred from the input (mp4→mp4, mov→mov…); an -o extension always overrides. Naming an -o video container without --video is an error — add --video, or choose a .wav path.
• Audio is stored losslessly where the container allows (PCM in mov/mkv/avi). mp4 gets AAC 256k, webm gets Opus 160k — there's no universal lossless audio codec for those containers.
• Without --video, the picture-related flags warn and are ignored.

Splice style: crossfade vs. cut

--video-splice crossfade (the default) dissolves each splice so the join is soft; --video-splice cut makes hard jump cuts on both streams. The audio and picture always make the same choice at every splice, so they can't drift apart regardless of style.

Keeping captions and lip-sync aligned

If something downstream depends on the timeline staying the same length — caption timestamps, lip-sync, a multitrack edit — use --mode silence. The filler's sound is muted but its slot stays, and with --video the picture is stream-copied untouched (lossless, frame-exact). See which --mode for the full decision.

Min-gap plays the footage through

With --video --min-gap-ms N, the injected pause plays the removed footage through (muted) rather than freezing the frame — the excised disfluency rolls under the pause instead of a frozen still. Audio and video inject identical, frame-snapped lengths, so sync holds within ~1 frame. The mechanism is in video render & A/V sync → min-gap plays through.

Where to go next

• The full flag list and defaults → the README.
• A copy-paste command for a common job → recipes.md.
• How sync is guaranteed (CFR, frame-snapped fades, the tail conform) → video render & A/V sync.