Need a vocal-free backing track, or separate parts of a song for a remix? Ultimate Vocal Remover uses deep-learning audio models to split vocals from music, with options for additional stems and audio adjustments.
How do you separate a song’s vocals?
Load an audio track, choose a separation model, and start processing. UVR produces separate vocal and instrumental files, letting you keep the backing track, isolate the voice, or bring the results into another audio editor. The output depends on the source recording and the selected model; dense or distorted mixes can leave artifacts in the separated parts.
Which stems can UVR create?
Along with vocal-and-instrumental separation, the available models can split a track into bass, drums, and vocals. UVR includes model families such as MDX-Net and Demucs, so the choice of model affects the separation approach and result. Try another model when the first output does not isolate the parts cleanly.
Can you change pitch or timing in UVR?
UVR also provides time-stretch and pitch-shift tools for adjusting audio. These operations rely on the Rubber Band library. The app uses FFmpeg when processing non-WAV audio files, and it remembers settings after closing, so a later session can retain the previous configuration.
Pros
- Separates vocal and instrumental tracks with selectable AI models
- Can produce bass, drums, and vocal stems
- Includes time-stretch and pitch-shift tools
System requirements: Windows 10, Windows 11 · 64-bit only · Nvidia GTX 1060 with 6 GB VRAM is the minimum for GPU conversions. · Nvidia GPUs with at least 8 GB VRAM are recommended. · FFmpeg is required to process non-WAV files. · Rubber Band is required for time-stretch and pitch-shift.
Can Ultimate Vocal Remover make karaoke tracks?
Yes. Separate the vocals from the instrumental backing and use the instrumental output as a karaoke track.
Can it separate drums and bass too?
Yes. Some models can produce bass, drums, and vocal stems.
Why does a separated vocal or instrumental still contain artifacts?
The result depends on the recording. Compression, distortion, and other characteristics of the source can affect separation quality.
