In this guide
First check whether you have the original music or a separate project track. Exporting a video's soundtrack keeps any speech mixed into it. If the music is mixed with dialogue, test a music isolation and review Isolated sound for voice bleed, missing instruments and truncated transitions.
To extract music from a video, first check whether the original music file or an independent music track is available. If the video contains only music, exporting its audio track may be enough. If speech and music are mixed together, exporting the soundtrack keeps both: try isolating the music and listen for remaining dialogue and lost musical detail.
Keep the original video and work with a recording you have permission to use. A separated audio file is an estimate of the sound in that recording, not the original licensed music track.
Choose the right extraction method
| Source | What to do | What you get |
|---|---|---|
| Original song file or video project with a separate music track | Use or export that original track. | The most direct source to compare; check whether project effects are included. |
| Video soundtrack contains only the music you need | Export the audio track in an editor. | The whole selected soundtrack, including any recorded ambience. |
| Music and talking are mixed in one track | Test source isolation targeting the music. | An estimated music track that may retain speech or lose notes. |
| You want to know the name of a song | Try identification using a clear passage and check the result. | A possible title or artist; isolation itself does not identify the recording. |
If your video contains only the soundtrack you need, the free video-to-WAV tool can extract a short supported MP4 or MOV locally. It keeps all sounds already mixed in the soundtrack.
Exporting a track is different from unmixing its contents. The audio extraction guide covers whole-track export, WAV versus stream copy and videos with multiple selectable audio streams.
Extract music from a video with speech
- Choose a representative excerpt. Include a section where speech overlaps music, plus a quiet instrumental moment and the natural ending of a phrase. Save the original start time if you will return the result to an edit.
- Upload the file. Open the background music extractor. It accepts MP3, WAV, MP4 and MOV up to 50 MiB (52.43 MB). If a video is larger, export a short excerpt from your editor for the test.
- Describe the sound to keep. Try “The background music” or describe an audible musical source more specifically. Select Isolate sound. Keep the description focused on one sound; it is a starting point, not a guaranteed setting.
- Review Isolated sound. This is the requested music candidate. Background audio is the remainder after that target is separated; despite its name, it is not automatically the music track.
- Compare before downloading. Listen at a similar level to the original. Check quiet notes, drum attacks and speech overlap. The download dialog shows the credit cost; AI downloads require an account and sufficient credits.
For example, if a presenter speaks over a piano bed, listen for the presenter during the loudest sentence and for piano notes at the same moment. A result with less speech may also have missing notes. Compare the whole phrase rather than choosing solely by how quiet the gaps sound.
Check whether the extracted music is useful
- Voice and effects: Listen for words, breaths, applause or other sounds that remain attached to the music.
- Arrangement: Compare bass, percussion, melody and sustained notes with the original. A quieter instrument may have been reduced with the speech.
- Transitions: Include the complete attack and decay of the passage; a trimmed start or tail cannot be reconstructed by isolation.
- Artifacts: Listen for watery textures, pulsing or abrupt changes where the speaker enters. Compare at the same comfortable playback level.
If the result loses an important part, return to the original and try a clearer description once as a comparison. Reprocessing a previous result is not a reliable way to restore missing information. A cleaner source, official soundtrack or original production track may be more useful.
What if the video is on YouTube?
Saturalabs accepts supported uploaded files; it does not import a YouTube link directly. If the video is your own, use an available original or a copy obtained through the platform's permitted options. For someone else's work, obtain an authorized file and the rights needed for your intended use. YouTube describes how creators can download videos they uploaded; its options are not a general license to reuse music.
Once you have a supported local file, the same source checks apply. If you want the title rather than an audio file, try a song-identification service with a clear excerpt and confirm the match; see the sound-finding guide for the distinction between identifying a source and separating it.
Use the result in an audio or video edit
Import the chosen WAV into your editor. If you used a trimmed excerpt, place it at the noted original start position and listen across both boundaries. Mute any original mixed soundtrack during the comparison to avoid doubled music. If the music should continue under speech, check its level and the finished video export, not just the isolated preview. The audio replacement guide explains timeline alignment and sync checks.
Common questions
Can I extract music from a video without speech?
Try targeting the music and review Isolated sound. Speech can remain and music can be altered, especially when both overlap. A separately recorded music track is a better source to compare when available.
Does extracting audio from an MP4 remove the talking?
No. Exporting the audio stream retains the sounds already mixed into it. Separating the music from speech is a different processing step.
Is the extracted music copyright-free?
No. Separation does not change the rights associated with the recording or composition. Check the permissions and terms for your intended use.