Mute a clip or extract its audio? Trim audio drop vs audio detach

They answer different needs. To keep the picture and lose the sound, send audio: "drop" to Sume video trim; you get a new MP4 for the range you cut. To keep the sound and use it elsewhere, call audio detach, which returns a new wav or mp3 and leaves the video untouched. Neither changes the source file.

Side by side

Typical uses of each

Mute when you will replace the soundtrack: a screen recording that you narrate with your own voice, or generated footage whose built-in audio you do not want. The result goes into Timeline 1.0 video[] and the timeline's audio spine supplies the sound. Extract when the audio is the asset: a talking-head clip whose voice you want to transcribe, or a track you want to split with timeline audio into reusable pieces. For many ranges from one clip, detach once and split afterwards.

What happens with a silent file downstream

A mute clip passed to timeline compose raises the warning compose_video_has_no_audio, which does not fail the job: the compose output uses the audio of the video, and the Timeline spine supplies the sound at assembly. A silent clip passed to audio detach fails with detach_source_has_no_audio, and a silent clip given to a captions job that expects speech fails with caption_no_speech. If you are not sure the clip has a track, read probe.has_audio with a frames: false inspect before you pick a tool.

{
  "video_url": "https://media.sume.com/artifacts/artf_demo/screen.mp4",
  "start": 0,
  "duration": 45,
  "audio": "drop"
}

One more note

With exact precision, kept audio is remuxed as AAC; with drop there is no audio stream at all in the result. If your next step is a platform upload that requires an audio track, add one on the timeline rather than uploading the muted MP4.