Remove the Background Music and Keep the Person Talking
Music Remover splits a file into voice, music, and other sounds. When a person is talking over a bed, EraseNoise lowers the music and keeps that voice.
Remove background music from a video when a track sits under someone talking. Keep the tool that lowers that bed. Switch tools when the file is a finished song and you want the band without the singer.
Music Remover says it separates Vocal, Background Music, and Other, including a background song that itself has sung vocals. EraseNoise does not hand you those three stems. It lowers music under a speaking voice and, for a video, gives back an MP4 with the same picture.
Key Takeaways
- A vlog, tutorial, or interview with a music bed is a speech job. A record you want as karaoke is a song job.
- Music Remover's page, read on November 1, 2026, says it separates three stems and accepts files up to 3 hours.
- EraseNoise keeps the speaking voice, returns video as MP4, and does not split a song into a vocal and an instrumental.
- Judge the preview on the words. If the speaker thins out, the quieter track is the worse file.
Which job is the file?
| What you have | What to open | What you should get back |
|---|---|---|
| A person talking over stock music or a cafe speaker | Remove background music | The voice stays. The bed drops. Video comes back as MP4. |
| A finished song, and you want the band without the singer | Remove vocals from a song | Singing comes out. Bass, drums, and chords stay. |
| A noisy room, and the music is not the problem | Background noise remover | Hiss, hum, or room noise drops. A music bed is a different page. |
Photo: Myotus, CC BY 4.0, via Wikimedia Commons.
What Music Remover says it does
On November 1, 2026, the Music Remover homepage said the AI separates Vocal, Background Music, and Other. It said Background Music can be an instrumental score or a song that includes sung vocals. The same page said you can switch among Original, Vocal, Background Music, Other, and a custom mix, then download one stem or all of them. The upload line said "Up to 3 hours" and listed upload, browser recording, screen recording, or a direct media link.
That is a stem workflow. It fits when you need the voice as its own asset for a new edit, a transcript, or a replacement soundtrack. It is a poor fit when you only want the original clip with the bed turned down and the picture unchanged, and you do not want to remix three tracks.
The page also links to a vocal remover and a karaoke maker. Those are song tools. They are not the same button as "lower the bed under this interview."
When someone is talking over a bed
Use the background music remover for a vlog, a lesson, or an interview. Upload the video itself. MP4, MOV, M4V, WebM, and MKV are accepted, and audio can be MP3, WAV, M4A, AAC, OGG, Opus, FLAC, M4B, or MKA, up to 2 GB.
You hear a free 30-second preview before anything is billed. Without an account that preview is available 3 times in 24 hours. After you sign in, it is 20 times. The full file is billed in whole minutes, rounded up, with a minimum of 1 minute. Audio comes back as WAV. Video comes back as MP4, and the frames stay as they were. EraseNoise deletes the upload and the result 24 hours later.
This page does not pick one person out of two, and it does not turn a song into an instrumental. If the words start dropping with the music, stop. A louder bed that still has the speaker is more usable than a clean file missing the sentence.
A fan, a hiss, or a room wash is not this job. Those go to the noise remover, or to echo removal when you can count a repeat. The difference between making a voice louder and taking a noise out is in voice enhancer vs noise remover.
When the file is a song
Don't run a song through a speech music-remover and expect karaoke. Remove vocals from a song takes the singing out, including the harmonies, and leaves the bass, drums, and chords. The preview is 30 seconds of the file you uploaded.
Music Remover can still be the right comparison when you specifically want three stems and a mix control. EraseNoise does not offer that mixer on the speech page. Hear both previews on the same chorus or the same sentence, and keep the one where the part you need is still there.
What to listen for
- Play the busiest sentence, not a quiet gap.
- Check that the speaker's name, the number, or the punchline is still intelligible.
- On a video, confirm the picture did not get re-exported into a different frame.
- If you needed a stem for a new soundtrack, confirm you actually received the voice alone, not a full mix that is only quieter.
Frequently asked questions
Can I remove background music from a video and keep the picture?
Yes on EraseNoise. Upload the MP4, MOV, M4V, WebM, or MKV. The download is an MP4 with the same frames and a soundtrack where the bed under the speaking voice is lower.
Will this make a karaoke track?
No. A finished song belongs on remove vocals from a song. The background-music page is for a person talking.
What if the voice gets thin?
Keep the original. Lowering a bed that shares frequencies with speech can pull the voice with it. The preview is there so you can reject that result before you spend minutes.
If you are choosing among noise-remover sites rather than a music bed, the same file is compared in five audio cleaner sites.