Ok, I need to extract the pure acapellas from my older songs for remixing,
so, what are the best, and simplest ways to do this?

 

You just create a new Automator Quick action and add a bash script:




Make sure to select Pass input as arguments:



See it will add a for loop for you which will run over all passed files

Spoiler

batch normalizing works either with help of sox

So now lets add the spleeter command to that loop :

Code:

for file in "$@"
do
   cd "$(dirname "$file")"
   spleeter separate -i "$file"  -p spleeter:2stems -o output
done

This already works, but it will download the pretrained_models for each new session.
By defining a MODEL_PATH we can avoid this. Here I just set it below my a folder.
Adjust to your liking. same with the model (spleeter:2stems, spleeter:4stems etc...)

Code:

export MODEL_PATH=/temp/pretrained_models/

for file in "$@"
do
   cd "$(dirname "$file")"
   spleeter separate -i "$file"  -p spleeter:2stems -o output
done

Save this workflow to ~/Library/Services. After that you can apply it to even multiple files at once, just select a file/multiple files and select Services, I called mine Spleeter2:

 

Try this:

Code:

(base) C:\Users\me>spleeter separate -i C:\Users\me\Desktop\audio\some_song.wav -p spleeter:2stems -o output

instead of this:

Code:

(base) C:\Users\me>spleeter separate -i spleeter/C:\Users\me\Desktop\audio\some_song.wav -p spleeter:2stems -o output

or run the command direct in the folder where you have your: some_song.wav

Code:

spleeter separate -i some_song.wav -p spleeter:2stems -o output

 

define best and simplest. macOS or PC ? do you know how to install python stuff ?
The easy way is to just use acoustica, but it is not free

 

Hi, thanks. I'm totally in the dark. Tried everything you suggested, I don't think the problem's spleeter not seeing the the audio file. From this:

Spoiler

, it feels like it's not installed correctly -- the .dll is sitting in that directory, and gets found if I set "set CONDA_DLL_SEARCH_MODIFICATION_ENABLE=1," at which point it grinds much longer before spitting out

Spoiler

(base) C:\Users\NewLoRez>set CONDA_DLL_SEARCH_MODIFICATION_ENABLE=1

(base) C:\Users\NewLoRez>spleeter separate -i spleeter/1.mp3 -p spleeter:2stems -o output
Traceback (most recent call last):
File "c:\programdata\anaconda3\lib\runpy.py", line 193, in _run_module_as_main
"__main__", mod_spec)
File "c:\programdata\anaconda3\lib\runpy.py", line 85, in _run_code
exec(code, run_globals)
File "C:\ProgramData\Anaconda3\Scripts\spleeter.exe\__main__.py", line 7, in <module>
File "c:\programdata\anaconda3\lib\site-packages\spleeter\__main__.py", line 54, in entrypoint
main(sys.argv)
File "c:\programdata\anaconda3\lib\site-packages\spleeter\__main__.py", line 46, in main
entrypoint(arguments, params)
File "c:\programdata\anaconda3\lib\site-packages\spleeter\commands\separate.py", line 43, in entrypoint
synchronous=False
File "c:\programdata\anaconda3\lib\site-packages\spleeter\separator.py", line 122, in separate_to_file
sample_rate=self._sample_rate)
File "c:\programdata\anaconda3\lib\site-packages\spleeter\audio\ffmpeg.py", line 63, in load
probe = ffmpeg.probe(path)
File "c:\programdata\anaconda3\lib\site-packages\ffmpeg\_probe.py", line 20, in probe
p = subprocess.Popen(args, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
File "c:\programdata\anaconda3\lib\subprocess.py", line 800, in __init__
restore_signals, start_new_session)
File "c:\programdata\anaconda3\lib\subprocess.py", line 1207, in _execute_child
startupinfo)
FileNotFoundError: [WinError 2] The system cannot find the file specified

Am totally clueless; gonna try to figure out how python works & stop posting logs/asking "what's this button do?"

 

I am currently installing it with anaconda on my windows system , lets see how it goes


[EDIT]
I still do not have it working under Windows. I get anytime errors when installing spleeter with anaconda. Some version conflicts huh ?!
On macOS it took just a few minutes and all was running.
I wonder if Spleeter woudl work under windows / Ubuntu subsystem?!
[/EDIT]

 

Last edited: Mar 17, 2020

PC.And are there options to send the WAV/mp3 to some resources and get back acapellas that are splitted off the tracks?

 

Like this?

 

Just tried that and compared it with results from melody.ml, both are pretty useful.

 

i just tried the algorithm with a track that had almost none of the instruments listed that it can seperate and all that came out was a bunch of chirping. it did have drums but they had severe chirping artefacts too.

edit:
just tried it on a track with the typical elements (vocal, drums, bass, etc) and the bass track came out usable, but the rest of the tracks were full of chirping and completely unusable.

in my experience so far izotopes seperator is so much better.

 

Last edited: Mar 17, 2020

i just tested spleeter (as it was recommended by fellow member in his post)
spleeter results were great in isolating one of the recent pop songs with 5 stems ,and easy use i can say )
just make a folder to drop the wav files at selected drive and drop the song u want its stems..results comes in less than 3 mins

i got 5 stems of vocals/drums/bass/other/piano in wavs (and since there was little piano in the background the piano file has bad result
as well as the bass (muddy result)and all this was fine with me because i needed the vocals only(good result)
and i can get rid of other muddy effect from vocals fx by spliting it to grid and clean it up

i might try spliter.ai soon on same song and see which gave better result

 

Really depends on the song... I was testing a couple of vocal focused songs, I imagine something very muddy can't be done very well. Those that are muddy I just add more mud onto the remix till I build a hill. Then I overdrive everything so it's just a smooth mess.

But no crackling, that's where I draw the line. Like sledding down a bumpy hill, with 90 degree bumps.

Smooooootttthhhh mess.

 

Obviously, this is some environment variable shit. Why don't you try CPython? (download link)

EDIT: Obviously you need to download all the extra required libraries with pip.

 

It got me (2 whole days!) into Python. Figured out just enough to read & sorta starting to understand it (enough to copy/paste & cobble stuff together). Playing with other people's scripts is fun, but spleeter isn't spleeter.py, it's spleeter.exe (it compiles/builds binaries during install, I guess?), and I can't figure out what it's looking for from

Spoiler

File "c:\programdata\anaconda3\lib\subprocess.py", line 1207, in _execute_child
startupinfo)
FileNotFoundError: [WinError 2] The system cannot find the file specified

Google (=>StackOverflow) are not being friends (kidding, doing this for fun, otherwise I'd just use one of the websites in this thread).

Thanks, maybe. Installed Anaconda for no other reason than (from their readme):

Spoiler

## Quick start

Want to try it out ? Just clone the repository and install a
[Conda](https://github.com/deezer/spleeter/wiki/1.-Installation#using-conda)
environment to start separating audio file as follows:

```bash
git clone https://github.com/Deezer/spleeter
conda install -c conda-forge spleeter
spleeter separate -i spleeter/audio_example.mp3 -p spleeter:2stems -o output

Easy! it'll be fun they said...

 

Yeah... About that... There's fine print. But there's some immense self satisfaction coming.

 

This is what finally worked for me:

1. (base) C:\Users\me>conda create --name spleetertest spleeter

Spoiler

Now it's happy, likes having its own environment, lists what it plans to do & asks if I'm cool with it

2. (base) C:\Users\me>conda activate spleetertest
3. python -m spleeter separate -i spleeter/audio_example.mp3 -p spleeter:2stems -o output
(on the first run, downloads & extracts 2stems.tar.gz, spits out a screenfull of deprecation warnings, and, finally,
4. joy!


TL;DR: installer doesn't know how to deal with base environment, I didn't know notin' 'bout no environments.