Has anyone noticed if there is a difference in quality between separating a mono or stereo track? Back in the early days of stereo, instruments were often hard panned. I'm wondering if this helps a separate or not. I'm wondering if some apps seem to prefer mono or stereo or maybe it doesn't matter.

Note: At present, I have the flu, my ears are ringing so loud that I barely hear anything.

 

Mono should be the same on each side. To save cpu, if it's designated a "stereo" track, you can bounce it as mono.

Different daws will export things differently (like Ableton likes to consider everything stereo etc... so sometimes I use a third part app like "Stereomonizer" or whatnot to not only diminish confusion from others receiving tracks (I don't want them trying to figure it out or wasting any time on that), or inline with that, cpu reasons. (which are diminishing in terms of audio track playback counts these days but)...

 

What I'm wondering is . . . dropping a track into a stem separator . . . would stereo be better or worse. I ask because I'd like to convert all my files to mono to make them half the size. But, would separating a stereo file have been better?

 

Don't convert to mono to save file space.
What DAW do you use? Since I use Studio One which supports FLAC, I've converted my standard library to lossless FLAC. Saves 50-60% space.

 

Is there a downside to using .flac vs. .wav?

 

I only ever separate stereo files. Almost every file I ever separate is a completed track. It has nothing to do with file size and storage, or anything else like that. I do not want to edit the file before separation, if I can avoid it. If I determine I want any of the stems in mono, I will generally just load them into the DAW (Logic), and I will insert a utility Gain plugin which has a button to mono the track. I use RipX, and it only kicks out a few stems. Storage and CPU use of those separated tracks don't matter with such a low number of stems.

I try to not spend prep time on something I may not even use, or keep. I'll convert it online in the DAW when it needs to be done. I'm not batch processing every sample I have in advance. Big waste of time, storage is cheap.

 

Yeah, I guess I wasn't sure what you were asking, as I wouldn't use a stem separator tool like RipX or whatever on my own sessions, I would just export from my own sessions what I wanted... and if I'm bouncing all of the tracks for a mix for someone else or in another daw etc, I don't want to give them a bunch of chafe or make them 'wonder' if crap is stereo or not as my daw may export it as that, so I try and keep it neat and clean...

My 'stems' I guess would normally be stereo, if I were approaching it that way.. and would export accordingly.

I wouldn't f with Flac or whatever, I would just go .wav's as hdd space is cheap and I don't want to worry about a recipient not being able to open it in their system... and in the world of "deliverables", "Flac" isn't a thing... it's fine for the music I listen to, but not what I would send to another professional client or collaborator etc... no way...

if working with other people, you usually have a system, or working with movie houses or whatever, per courtesy and to keep things clear and concise, they usually tell you exactly what they want, and I've never seen Flac as an acceptable standard for that kinda thing myself.

 

My initial reaction strictly to the technical notion of "does it matter if you're extracting stems from a mono or stereo source" is I can't imagine why it would make a difference but it actually is kind of an interesting question... does it?? Would it? Why?

 

In the case of hard panned elements of a mix, it does matter, as that is a form of inherent separation making it easier for the algo to identify. If you collapse one of those old "Big Mono" mixes into mono you'll get phase cancelling of elements and freq masking in the centre channel making the task for the algo even more difficult, even just listen to them in mono and you'll hear it.

Most modern stereo mixes are mono compatible by design so the lost info by collapsing to mono isn't as drastic. I value the spatial information so if there's a choice I'll use the stereo file.

I do various things to pre-process the material I'm trying to seperate, like eq'ing, spectral enhancement/removal, using just the L or R channel, adding fx and so on. Every track is unique and it depends what I'm after in the separation, but I have found you can manipulate the material to get different results from different algos, it's a bit trial and error.

Hope you feel better soon.