Remove Music from a Voice Recording
When a voiceover, interview or clip has music underneath, this tool removes the music bed and keeps the speech. The Demucs engine separates the voice from the backing track.
How it works
Demucs treats the spoken voice as the "vocal" stem and the music as accompaniment, returning the isolated speech. This rescues dialogue that was mixed over a song or score.
What it's good for
- Clips with copyrighted music beds
- Interviews over background music
- Cleaning music from voiceovers
- Salvaging dialogue from videos
Details
- Engine
- Demucs
- Formats
- MP3, WAV, M4A, FLAC, OGG, AAC, MP4, MOV
- Price
- Free to try
Frequently asked questions
Same engine, different intent: vocal isolation is framed for songs; this is framed for spoken voice over music. Both return the voice stem; pick the one matching your material.
Speech comes out clearly with the music greatly reduced. Loud music in the same range as the voice can leave faint traces; follow with the background-noise tool if needed.
Very loud music that overlaps the voice's frequencies is the hardest case, but Demucs still pulls the speech forward in most clips.
No. Removing a music bed only changes the audio; if the clip used copyrighted music, taking it out does not grant you any rights to the original recording.
A clean source where the voice is reasonably forward in the mix works best; you can upload common audio formats, or extract the audio track from a video first.
Short voiceovers and interview clips usually finish in seconds to a minute, scaling with the length of the recording rather than how loud the music is.
Some residue can survive when music sits in the same frequency range as the voice; running the cleaned speech through the background-noise tool afterward smooths over most of it.