How to remove vocals from a song and make a karaoke backing track
To make a karaoke track, separate the recording into vocals and instrumental, then keep the instrumental file. A vocal remover estimates these parts from a finished mix; it does not retrieve the artist’s original studio tracks. Listen before committing to a full-song result, because removing the singer can also affect sounds that overlap the voice.
1. One-minute audio test
It's OVER!!! — Admiral Bob · 0:30–1:30 · 60s
Compare the same passage at the same position. Listen for voice left in the backing track and instruments missing from it. These samples do not establish a universal model ranking.
Click or drag a waveform to seek. On a focused waveform: ←/↓ back, →/↑ forward; Tab next track, Shift+Tab previous track. Space plays or pauses.
Loop
Test settings and processing record
Tested on 7 November 2026: standard mode, 4-second segments, NVIDIA RTX 4060. Times cover local inference and export, excluding model loading, queue and transfers.
| Model | Processing time | Peak VRAM |
|---|---|---|
| kim-melbandroformer | 7.63 s | 1,451 MiB |
Model loading, queue and transfer time are excluded. All versions use the same input and standard mode. MP3 previews: 192 kbps; references: 96 kbps mono, 16 kHz.
Author reference audio — not AI output
Reference recordings were aligned to the original mix to check that the target sound occurs in this passage. This verifies activity, not subjective separation quality. Reference levels were adjusted; these are not the original mix balance.
Full test record ↓2. Define the goal and prepare your audio
If you mean the sung words, vocal removal is the relevant task. If you mean subtitles or printed lyrics on a video, audio separation will not erase that text. An instrumental version keeps the musical backing without the main sung performance, though some vocal traces or backing harmonies may remain. To keep the voice and remove music instead, use the acapella workflow.
Use a recording you own or are authorized to process. Choose the best source file you already have: MP3, WAV, FLAC or M4A. Avoid making several new lossy copies before uploading. Converting a compressed MP3 into WAV does not bring back missing detail; an MP3 to instrumental workflow can start with the MP3 itself. A cleaner source gives you a fairer basis for comparing results.
3. Process the song and choose a download format
- Upload the audio, then choose the vocal/instrumental processing option. Check the displayed account limits and cost before starting; they can change with your account and site settings.
- After the full result is ready, compare the difficult passage with the original. Check both the wanted detail and unwanted leakage before choosing a track.
- Play the Instrumental output first. The Vocals file is useful for diagnosing what was removed, but it is not the karaoke backing track.
- Compare against the original at a comfortable, similar listening level. Keep the same timeline position when switching, and loop a phrase rather than repeatedly seeking three different players.
- Select an available download format and download the instrumental before the displayed file expiry. Keep your original separately.
MP3 is convenient for a small rehearsal file. Where available, WAV or FLAC avoids adding another lossy encoding stage during editing; it cannot repair separation artifacts. If you will edit or combine tracks, retain a common start point rather than trimming each file independently.
4. A practical three-pass listening check
First, listen to the quiet gaps between phrases. A ghost of the singer, a repeated syllable or a long reverb tail is easier to hear there. Next, follow a guitar, keyboard or string part through the sung section. If that part becomes hollow only when the singer enters, the removal may be taking too much accompaniment with it. Finally, sing over the result at the volume you expect to use. A faint trace that is harmless during rehearsal may still be distracting in an exposed performance.
Write down the exact phrase and the problem: “last word of chorus remains” is more useful than “bad quality.” Compare another available model on that same phrase if needed. Do not assume that a bigger model, extra processing or more stems will automatically make the backing better. For a performance where the remaining artifacts are unacceptable, seek an authorized instrumental release or a newly recorded backing instead.
5. Frequently asked questions
Can I remove vocals from an MP3 online?
Yes, MP3 is an accepted input in uvr online. Start with the best existing copy, then evaluate the instrumental; changing the extension or exporting WAV cannot guarantee a cleaner separation.
Does a karaoke maker remove backing vocals too?
A general vocal model may group lead and backing voices together or leave some behind. Keeping harmonies while removing only the lead is a separate specialist task.
Which is the best vocal remover for my song?
There is no supported universal winner here. Compare the same difficult phrase, the vocal leakage and the damage to accompaniment, then choose the result that fits your use.
Can I use a vocal remover online for free?
uvr online offers a free vocal-removal route subject to the current usage, size and duration limits shown before upload. A free vocal remover is not unlimited processing: check those limits, any verification step and file expiry first.
Does a voice remover erase all background voices?
A music voice remover usually targets the vocal group, not only a chosen person. If you want harmonies preserved, use the lead/backing guide rather than assuming any AI vocal remover will keep them.
Audio sources and usage
It's OVER!!! — Admiral Bob · CC BY 3.0
The excerpt was trimmed, separated into audio stems and encoded as MP3.
If a sample affects your privacy or rights, contact support@uvr.online or open a Support ticket with the sample link to request a review and removal. Contact support →