A friend talk to me about the SpectraLayers as a tool capable to separate individual instruments.
Let’s assume I have an mp3 or whatever music file, with four instruments involved (for instance: piano, violin trumpet and flute), I want to produce four independent files each one with only one instrument’ sound track.
Is SpectraLayers able to do that? If yes, is there some doc that explain how to do it?
Thanks for your help,
There are good video tutorials official and not, like this:
Directly you may unmix intruments
See the online manual
Plus you there is a separate process that allows to unmix components from individual layers (tonal parts, transient and noise ones).
Therefore you have two main unmix processes together with several tools to perfect and fine tune the results.
Starting by several editing precision tools to dial in/extract/etc. the desired frequencies with fine gradable, complex selection tools akin to Photoshop ones.
Since spectral processing works on “visual” aspect of sound, SL Pro further allows to change the resolution of the spectral image (and the associated tools) to focus its processing, iex; on the differences related to the particular instrument you are dealing with (there are all kinds of flutes, pianos, trumpets, etc). This alone is unique and one of the most powerful tools available.
Additionally SL Pro has sensitivity controls and several algorithm choices for further fine selection.
All of those tools are integrated into a powerful multitrack layer management, ARA2 connected to your DAW. These have been perfected for near a decade.
///So, yes SL Pro has a press and play unmix process, like several other software out there, however in this case accompanied with a laboratory built around to expand the possible results, quite further than what the AI already achieves.
…"Let’s assume I have an mp3 or whatever music file, with four instruments involved (for instance: piano, violin trumpet and flute), I want to produce four independent files each one with only one instrument’ sound track…’
The answer is “no”, Spectralayers will not split your mp3 into separate piano-violin-trumpet-flute…not in the automated way you’re expecting.
Why? Because the underlying demix technology in SL is based on Spleeter and Spleeter, to this date, is trained to look for and split…drums-bass guitar-vocal-piano-other.
“Other” is likely where your violin-trumpet-flute would land…still mixed together.
Why is this the way it is? Because here in 2021, that’s the state of the “automatic demix” technology.
Which is way better than where we were five years ago. And which will get better.
Well, once SL splits your piano by itself automatically, you could then manually extract the flute-trumpet-violin components from within SL or Melodyne or etc.
But that presents a learning curve for you and the possibility of artifacts.