video-audio-replace — Traditional AI Model
Audio Replace keeps a clip's picture exactly as it is and gives it a new soundtrack — swap the original audio out, or lay the new sound over it. The video is copied, not re-generated or re-encoded, so a clip you like stays the same while you redo only its sound. Set length to Sound and the video lasts exactly as long as the new sound instead — a single still with a music track ends when the music does. FFmpeg.
Details
- CategoryTraditional
- Year-1
- License
Inputs & Outputs
- VideoInputvideo
The clip whose picture is kept
- SoundInputsound
The new audio (TTS, music, or an uploaded track)
- VideoOutputvideo
The same clip with the new audio
Parameters
- modestringdefault replaceoptions: replace, mix
Replace drops the clip's own audio; Mix keeps it under the new sound.
'replace' (default): the output carries only the new sound. 'mix': the clip's original audio stays and the new sound is laid over it; set the balance with original_volume and sound_volume. A clip with no audio track behaves as 'replace'.
- sound_volumefloatdefault 1range 0–2
Volume of the new sound. 1 keeps it as is, 0 mutes it.
Gain applied to the new sound, 0-2 (1 = unchanged, 0 = muted).
- original_volumefloatdefault 1range 0–2
Volume of the clip's own audio in Mix mode. Ignored in Replace.
Gain applied to the clip's original audio in 'mix' mode, 0-2 (1 = unchanged). Ignored in 'replace'.
- sound_offsetfloatdefault 0range 0–60advanced
Seconds into the clip before the new sound starts.
Delay, in seconds (0-60), before the new sound starts playing in the clip. Use it to line the sound up with the picture by hand.
- lengthstringdefault videooptions: video, sound
Which input sets the output's length. Video keeps the clip's length; Sound ends when the new sound does, holding the clip's last frame if it is shorter.
'video' (default): the output is exactly as long as the input clip — a shorter sound is padded with silence, a longer one is cut — and the picture is stream-copied untouched. 'sound': the output ends where the new sound ends (sound_offset + the sound's length). A longer clip is cut there (still stream-copied); a shorter one holds its last frame until the sound ends, which re-encodes the picture. Lengths are measured from the streams themselves, so a clip whose audio runs past its picture is held rather than cut short. The output is capped at 600 seconds in this mode. Use 'sound' for one still + a music track: a 1-second Video-KenBurns 'Hold' clip + a 40s track gives a 40s video (to within a frame), with no silent tail when it loops. Video-Narration is the wrong node for that — it plays the voiceover over a frozen first frame and then appends the clip, so the result is narration + clip long.
Tags
Alternatives in Video Editing
- Audio Edit
Trim a sound, fit it to a video or a length, dip or mute parts, and fade it.
- Audio Finish
Normalize a video's loudness to a broadcast target (2-pass loudnorm + peak limiter).
- Audio Mix
Mix narration, music and sound effects into one track, with ducking and loudness.
- Chroma Compose
Cut a green-screen person out of one video and place them over another.
- Hook Teaser (Cold Open)
Prepend a cold-open teaser of each clip's hook moment to the front.
- Ken Burns (Still to Video)
Animate a still image into a video with a moving crop window.
- Narration Mux (Cold Open)
Prepend a voiceover cold-open (frozen first frame + narration) to a clip.
- Remove Dead Zone
Remove silent gaps from each clip in a video list to tighten pacing.