To remove music from audio and keep speech, upload the recording to the audio splitter, separate voice and background, and listen with the voice track selected. Download the voice audio when it sounds suitable. This workflow does not require converting the file into a video.
What an audio-only workflow changes
An interview, podcast clip, language lesson or recorded voiceover may already be a standalone audio file. Using an audio-only tool keeps attention on the recording: there is no picture to preview and no combined video to export.
The separation goal is still the same. A third-party model estimates a voice track and a background track from the mixture. Speech and singing can both appear in the voice track. The background track can contain instruments, ambience and effects, rather than music alone.
Before processing, check whether you have the original editing project. If speech and music remain on separate tracks, you can lower or remove the music in that project. If all sounds are already mixed into one file, a volume slider cannot turn down music independently; it turns down the whole recording.
How to remove music from audio
- Choose an audio file. The audio splitter accepts MP3, WAV, M4A, AAC, FLAC. It is intended for audio files; use the video workflow for a video recording.
- Listen to a representative section. Note a few words over the music, a pause and a speaker change. Keep the source for comparison.
- Sign in and submit the task. Check the format and the cost shown by the tool before starting. Keep the browser open until upload completes and the task is queued.
- Select the voice track. Turn background off to hear the music-removal result. You can select both tracks and adjust each volume when a little background sounds more natural.
- Download the individual audio. Choose the voice track for speech-led editing or the background track if that is your intended output.
You can return to My Files after leaving the page. Uploaded files and processed downloads remain available for 30 days. Save a copy on your device for a longer project.
Compare the result in context
Do not judge an entire recording from its quietest moment. Listen where the music and voice overlap most strongly, where words end, and where the speaker pauses. Compare the same time position in the source, voice and background tracks.
If a word sounds thin or broken, listen to the background track at that point. Parts of the voice may have moved into it. You can try a quieter background mix, keep the source for that section in your editor, or decide that the processed version is unsuitable.
Separation is also different from repairing wind, clipping, room echo or a poor microphone recording. Those problems may remain after the music is reduced. Repeatedly processing a damaged result is not a reliable way to recover the original voice.
For transcription, use the cleaner track as an aid and verify names, numbers and uncertain words against the source. The tool separates sound; it does not produce a transcript or verify what the speaker said. See the transcription preparation guide for a review workflow.
Choose an audio format
Available separated audio exports are AAC, MP3, WAV, M4A, FLAC. Use a format accepted by your next editing tool. WAV can fit a workflow that needs uncompressed audio; compressed formats may be more convenient for listening and sending files. A larger export does not by itself mean the separation is more accurate.
Download a named voice file and keep it next to the source. For example, use a filename such as interview-voice so another editor can distinguish it from the mixed recording. If you need to combine it with other material, check the starting position before making any cuts.
Common questions
Can I remove music from an MP3?
MP3 is a supported audio input. Processing aims to separate voice from background, but exporting the result as WAV cannot recover information lost before the upload. Listen to the actual result instead of judging it by the file extension.
Can I remove singing and keep the instrumental?
Use the vocal remover with background selected. It starts with the opposite preview choice to keeping speech. Instruments or reverberation can still be affected, so review the output.
Does the tool separate drums, bass and every instrument?
This workflow provides voice and background tracks. It is not a multi-instrument arrangement editor. If you need individual drum or bass tracks, do not assume they will appear as separate downloads.
Can I mix the voice and background afterward?
Yes. Select either or both tracks and adjust their individual volumes in the preview. For further cuts, fades or sound design, download the audio and continue in your editor. The result is still subject to the source recording's quality and your rights to use it.
Read the privacy policy before submitting sensitive recordings, or contact support about a task.