I discovered the method where you can fix sound separations and other unwanted sounds from stems using at least 2 different AI based stems providers. Lets say we have a random track and you just extracted instrument stem from it using Engine DJ Pre-rendered processing but the issue is you can still hear high layer from the snare and clap in some cases depending of the specific track.
You are still able to extract stems from other stem using another AI separation. Like you used Engine DJ Software then Virtual DJ to fix other unwanted sounds.
Why?
Because not eveything is perfect and every audio companies has own developed AI stem features so its totally different performance and quality of stem processing. I calling this “Substem” technique where you use 2 or multiple AI separation softwares over stem to do another stems from it.
This might be sometimes very useful when Engine DJ not separated that perfectly, so you will use other third patry separation for stem to create substem.
This may fix and remove some other sounds in the background to make sure it’s cleaner for you.