deMIX Pro is pretty damn good at vocal separations! Check this out

QuadraphonicQuad

Help Support QuadraphonicQuad:

JonUrban

Forum Curmudgeon
Staff member
Admin
Moderator
Since 2002/2003
Joined
Mar 2, 2002
Messages
17,236
Location
Connecticut
Nice! I am glad you are happy with it John. I hate to talk-up something that costs quite a bit and cause someone to purchase it and then be disappointed. I think this is great stuff, and it might even get me back into playing with files on the PC. Quite honestly, I hadn't touched this stuff in months. I haven't listened to what others have done either, which is a shame. I am sure there have been very amazing upmixes done by QQ'ers that I really should be enjoying. Maybe it's time for me to get off my lazy butt! :)
 

ar surround

2K Club - QQ Super Nova
QQ Supporter
Joined
Apr 3, 2010
Messages
3,961
Location
New Joisey
I hate to talk-up something that costs quite a bit and cause someone to purchase it and then be disappointed.
It gets even better. Here are the vocals for Please Please Me. The individual voices are so clear, distinct and well separated that I may actually need to do some blending.

Please Please Me.jpg


Separation of the voices on Twist and Shout didn't work out so well. However, the "ahhhh, ahhhh, ahhhh, ahhhh" parts did separate nicely and will go into the rears as well as the fronts. I was also able to isolate the "Ooooo"s. They will sound good added to the back reinforcing the "Ooooo"s in the fronts.

Twist and Shout.jpg
 

JonUrban

Forum Curmudgeon
Staff member
Admin
Moderator
Since 2002/2003
Joined
Mar 2, 2002
Messages
17,236
Location
Connecticut
For you Joni fans, I've always said that the vocals in "Raised on Robbery" should have been split up across the quad channels for the quad release. Well holy crap, check this out. Joni lead in the center, and stereo backgrounds in the rear.

AMAZING SHIT!!!
 

Attachments

  • SampleJoni.flac
    5.6 MB · Views: 0

Wagonmaster_91

600 Club - QQ All-Star
QQ Supporter
Since 2002/2003
Joined
Mar 11, 2002
Messages
662
Location
Dallas,TX
This is impressive stuff. I may have to take the plunge and buy it, too.
Is the separation adjustable (within the program) if you are getting artifacts in the results? I would usually opt for higher quality over absolute separation when I am doing upmixes. Sometimes too much isolation of instruments or vocals feels to me more like an experiment than a good surround mix. Everything has to 'hang together' or sometimes it feels like 4 - 6 sounds just happen to occur at the same time.
 

ar surround

2K Club - QQ Super Nova
QQ Supporter
Joined
Apr 3, 2010
Messages
3,961
Location
New Joisey
This is impressive stuff. I may have to take the plunge and buy it, too.
Is the separation adjustable (within the program) if you are getting artifacts in the results? I would usually opt for higher quality over absolute separation when I am doing upmixes. Sometimes too much isolation of instruments or vocals feels to me more like an experiment than a good surround mix. Everything has to 'hang together' or sometimes it feels like 4 - 6 sounds just happen to occur at the same time.

Like I said above in post #22, you might find the need to blend some separated voices afterwards for a more convincing mix. Post facto blending is easy. Separation is not.

I can see myself spending infinite hours mixing almost the entire early Beatles albums into 5.1 using this software. I will get nothing done around the house. My wife, she's is going to kill me!!!
 

ar surround

2K Club - QQ Super Nova
QQ Supporter
Joined
Apr 3, 2010
Messages
3,961
Location
New Joisey
Has anyone tried this online service?:
mvsep.com
It also seperates tracks. Seems to be free.
That's what I've been using until I got DeMiX Pro. Use the Demucs 4 HT model. I also use LALAL.AI from time to time. For example, the Demucs models have difficulty separating the bass in much of the early Beatles stuff, even after boosting it substantially. The LALAL.AI Phoenix model does separate the bass.

[Edit: Sometimes you have to use more than one model, because one might do something better than another and vice versa.]
 
Last edited:

Blackwood

Active Member
Joined
Oct 6, 2022
Messages
75
Location
U.S.
Use the Demucs 4 HT model. I also use LALAL.AI from time to time.
Good to know. I just started playing with it a month ago and found Demucs3 Model B pretty good, but it has been a bit of a mystery to me which to use.

Looking into getting deMix Pro, though. If it simplifies things, that’s worth the the cost.

And the samples @JonUrban has been sharing are stunning.

With the Thanksgiving holiday coming, that might be a good time to try the demo.
 

ar surround

2K Club - QQ Super Nova
QQ Supporter
Joined
Apr 3, 2010
Messages
3,961
Location
New Joisey
Good to know. I just started playing with it a month ago and found Demucs3 Model B pretty good, but it has been a bit of a mystery to me which to use.

Looking into getting deMix Pro, though. If it simplifies things, that’s worth the the cost.

And the samples @JonUrban has been sharing are stunning.

With the Thanksgiving holiday coming, that might be a good time to try the demo.
If I've already completed something using Demucs3 Model B, I don't bother doing it again with Demucs 4 HT. Also, I've found a couple of instances where Demucs 3 and 4 have done a more complete extraction than the LALAL.AI Phoenix algorithm or DeMiX Pro. So it's best to keep all of these in your quiver. Compared to the human brain, AI currently still stands for "Artificial Idiot."
 

ar surround

2K Club - QQ Super Nova
QQ Supporter
Joined
Apr 3, 2010
Messages
3,961
Location
New Joisey
Here is a snippet of The Beatles P.S. I Love You created from the mono recording with DeMiX Pro. It is a mixdown to 2.0 from a 5.1 mix that I created.
 

Attachments

  • P_S_I_Love_You_2.0 Mixdown Snippet.flac
    2.3 MB · Views: 0
Last edited:

JonUrban

Forum Curmudgeon
Staff member
Admin
Moderator
Since 2002/2003
Joined
Mar 2, 2002
Messages
17,236
Location
Connecticut
Just joined in on the DeMIX Pro gravy train, will be evaluating results this weekend; any tips for best results will be much appreciated!

Oh no! I'm a bad influence! ;)

It's pretty easy to figure out how to get right into it. So far, I have no tips, just have fun.

PS - MixerRog is the first one I know of to try this software, and he's been pretty happy with it so far. That was a big part of me taking the plunge. Heck, we're all in this together!
 

J. PUPSTER

💿🐕 Senior Disc Chaser 🎸
QQ Supporter
Joined
May 30, 2017
Messages
11,059
Location
CALIFORNIA (CENTRAL)
Oh no! I'm a bad influence! ;)

It's pretty easy to figure out how to get right into it. So far, I have no tips, just have fun.

PS - MixerRog is the first one I know of to try this software, and he's been pretty happy with it so far. That was a big part of me taking the plunge. Heck, we're all in this together!
I've had nothing but fun with my Up-mix journey; but yeah you're the ultimate OG influencer on here Jon, since you got me into this in the first place :LOL:
 

JonUrban

Forum Curmudgeon
Staff member
Admin
Moderator
Since 2002/2003
Joined
Mar 2, 2002
Messages
17,236
Location
Connecticut

I played around with it a bit more this morning. Not every audio track results in spectacular results. As will all upmix technologies, your results will depend a lot on the source file. I did a few random stereo wav files I had on my PC this morning while I was waiting for my wife to get ready (that gave me a lot of time :) ), and some were good and some were not good at all.

I posted this elsewhere, but the Steely Dan (Fagen) vocals on 'My Old School' were amazing, IMHO. I am not sure Don would approve!

 
Top