Ultimate Vocal Remover – Neural networks at your service for your instrumentals

I’ve told you many times about online services that can separate the instrumental audio from the vocal track of any song. Things like Acapella Extractor and Vocalise for example or even this kind of integration in Ableton.

These are pretty decent tools, but they don’t work miracles either.

But that was without counting on Ultimate Vocal Remover GUI (UVR), a Windows interface for a track splitter that uses neural networks to achieve its ends. I tested it and the result is really excellent.

So UVR is really a large software of several gigas which embeds everything necessary to operate and which above all requires a lot of machine power. If you have an Nvidia graphics card with at least 8 GB of V-RAM this is ideal.

As input, you put your audio file (MP3) then you select an output directory and above all a model, the best being MDX-Net + UVR MDX Net Main or Ensemble Mode + Multi AI. You can also force the GPU or increase the noise reduction.

Thus, after a few minutes, you will recover an instrumental version and a vocal version. Even ad libs can be extracted or left to choose (karaoke template). Enough to make all the remixes, karaoke versions and samples of your dreams. It is magic !

If you are interested in testing the tool, it is available here on Github. You will first have to download the software, install it and then apply its patch. Read the instructions on the site carefully..

5/5 - (1 vote)
SAKHRI Mohamed
SAKHRI Mohamed

The blog of a computer enthusiast who shares news, tutorials, tips, online tools and software for Windows, macOS, Linux, Web designer and Video games.

Articles: 3747

Leave a Reply

Your email address will not be published.