Remove the vocals from a stereo song and download an instrumental — processing stays in your browser.
Drop a stereo song here
or click to choose from your computer
MP3 · WAV · FLAC · AAC · M4A · OGG
Working…
Vocal Remover turns a regular stereo song into an instrumental version. Drop in a track, press play on the Instrumental button, and you can hear the song without the lead singer before you download anything. It is made for the moments when you need the music but not the voice: a karaoke night, a cover recording, a dance or video background, or learning a part by playing along.
Because everything runs in your browser, there is no upload queue and no waiting for a server. You can compare Original and Instrumental with one click, watch both on the waveform, and change the settings until it sounds right.
In most studio mixes the lead vocal is placed in the center, which means it is equally loud in the left and right channels. Guitars, pianos and synths are usually panned to the sides. That difference is what this tool uses.
The song is cut into small overlapping slices and each slice is split into frequencies. For every frequency the tool checks whether it is “centered” (the same in both channels) or panned. Centered parts are turned down, panned parts are left alone, so the stereo image of the band stays intact.
The right channel is subtracted from the left. Everything identical in both channels cancels out. It is instant, but the result is mono and the cancelled part also includes other centered sounds.
Kick, bass and cymbals are centered too, so plain removal can make a track sound hollow. The tool puts the low end (below the cutoff you set) and, optionally, the very top (above “Keep above”) back from the original.
A quick test: if you listen on headphones and the voice sounds like it is “inside your head” rather than on one side, the song is a good candidate. Tracks from the 70s to today that were mixed with a centered lead vocal usually give a clear instrumental.
For the cleanest result, start with the highest quality source you have (WAV, FLAC or a high-bitrate MP3). A low-quality file leaves more artifacts after removal.
Use Original and Instrumental to switch between versions while the song plays, and listen for the chorus, where the voice is the loudest. If you can still hear it there, increase Center width a little, then check the drums: if they have become weak, raise “Keep below”.
If you plan to sing or play along, lower the speed or pitch afterwards with the Speed Changer or Pitch Shifter, or boost the final file with the Volume Booster.
Not from every song, and it is better to know that upfront. The tool relies on the stereo image, so it works best when the lead vocal sits in the exact center of the mix, which is true for most pop, rock, hip-hop and R&B releases. Songs with wide stereo vocals, heavy vocal reverb, or a mono master will keep some trace of the voice.
Reverb and delay on a vocal are usually spread across the left and right channels, so they are not “in the center” and are left behind. Try a slightly larger Center width, or switch to Full removal. A faint ghost of the voice is normal for this kind of tool.
Kick, bass, snare and sometimes a solo instrument are also mixed in the center, so they get reduced together with the vocal. Keep bass and kick (on by default) and the Karaoke preset restore the low end from the original, which fixes most of it.
Classic subtracts the right channel from the left. It is instant, but the result is mono and anything centered goes away. Spectral analyzes the song in thousands of tiny time and frequency slices and lowers only the centered ones, so guitars, keys and backing parts keep their stereo width. Spectral takes longer but sounds noticeably better.
No. AI tools such as Demucs or Spleeter are trained to recognize voices and can pull out a clean vocal or drum stem, but they need heavy models or a server. This tool uses signal processing that runs entirely in your browser. It is faster to open, needs no upload, and is very good for karaoke and practice tracks, but it cannot isolate vocals that are not centered.
No. The file is decoded and processed on your device, and the instrumental is encoded in the browser as MP3 or WAV. Nothing is sent to Sound Booster servers.
MP3 at 192 or 320 kbps is perfect for karaoke and sharing. Choose WAV if you plan to keep editing the instrumental in a DAW or run it through another tool, so you do not lose quality twice.
Files up to 100 MB are accepted. Longer songs simply take more time to process, especially with the Spectral method.
Keep editing in the browser after you remove the vocals.
