I just discovered MVSep thanks to a A.S. member. Any thoughts on how this separator compares to UVR and the others for drums?

 

Does their free mode allow many 10 minute songs or just one and done forever?

 

Is there a limit per day, month, etc?

 

MVSEP is the best service! Been using it for years, so many algorithms are available! And there are also credits with which you can get results, if you don't want to wait 10-20 minutes for the queue (there is a lot of queuing on certain days, and sometimes less).

 

Never heard of MVSEP.
How does it stack up against Spectralayers or RipX??

 

Same interested.

 

In many cases UVR and MVSep are using the exact same models. All of the models available on MVSep are available to use in UVR, with the exception of a few. There's nothing else available that gives you the models available for free like UVR and MVSep. They are the best as far as the bleeding edge models. Everything else is mostly just using a version of Demuc 4, which is pretty old at this point when it come to the models available in UVR.

 

For Intel users I'd go with OpenVINO™ Music Separation (part of OpenVINO™ AI Plugins for Audacity) which relies on Meta's Demucs v4 model - same used by MVSEP - and produces comparable results to (or sometimes even better of) Izotope RX 11's Music Rebalance feature but for free.

Last but not least, for those who wanna experiment, my collection of open source audio separators can be found at:

HyMPS project \ AUDIO \ AI-based \ Separating

Enjoy !

 

Last edited: Apr 24, 2025

Is this part of Audacity? An add-on? Do I need an old Intel Mac?

 

No, it's an Intel project, so an Intel CPU and/or GPU needed to run it.

Works on my i5-12600 on both CPU and iGPU.

 

Will it run on my Mac Mini M2?