MOTUnation.com

mhschmieder wrote: ↑Tue Apr 23, 2024 9:59 am

HCMarkus wrote: ↑Mon Apr 22, 2024 2:21 pm I'm just gonna' try to mix it well the first time around.

Famous last words.

Of course, most of us are dealing with projects we did not control, kind of like how most software jobs are cleaning up the last person's mess.

The other main category is premature digital transfers that forced us to do track reduction from the original analog, sometimes to stereo (mixed, not even dual mono!), followed by the loss of the multitracks and/or the gear that could play them.

Amen to that!

Some gearheads at my church think that it will be a great summer project to mount an expedition into THE BASEMENT WHERE ANALOG GOES TO DIE!!! pull out my analog machines and transfer my boxes of tapes to digital. I believe that I will let them.

Donnie Osmond once told me that a beef jerky drying oven is the perfect appliance for baking polyester backed tape. Amazon has a few with 5-layer round trays around 11"–12" diameter. Perfect.

All my old analog (and a lot of digital) tapes burned up in 2003 Cedar Fire.

So there is no basement here.

stubbsonic wrote: ↑Tue Mar 05, 2024 9:04 am I'm hoping they add new instrument classes to their AI picks for unmixing. Organ would be an obvious one. Since they have one for drums, perhaps after other types are unmixed, then they could have a second step to unmix parts.

Spectralayers 11 came out this past week. and it adds unmixing of brass instruments, unmixing of lead and background vocals, and unmixing of "foreground music" vs. "crowd" among other things...

scott_bernstein wrote: ↑Mon Jun 24, 2024 12:21 am Spectralayers 11 came out this past week. and it adds unmixing of brass instruments, unmixing of lead and background vocals, and unmixing of "foreground music" vs. "crowd" among other things...

I was lucky that I got SL 10 on sale, and SL 11 ended up being included. Still, hopefully they'll provide even more degrees of unmixing. As it is, it's mostly quite good. Perhaps paired with some tools in Melodyne's software, parts could be extracted as well.

My upgrade price was under $68 so I went for it. I did not find v.10 that useful to me but will see what v.11 brings to the table.

I am hoping that EQ match is better than that found in Ozone which only works on mixes. I need something more along the line of RX's Dialog Match — which I have but it requires ProTools that I do not have.

scott_bernstein wrote: ↑Mon Jun 24, 2024 12:21 am

stubbsonic wrote: ↑Tue Mar 05, 2024 9:04 am I'm hoping they add new instrument classes to their AI picks for unmixing. Organ would be an obvious one. Since they have one for drums, perhaps after other types are unmixed, then they could have a second step to unmix parts.

Spectralayers 11 came out this past week. and it adds unmixing of brass instruments, unmixing of lead and background vocals, and unmixing of "foreground music" vs. "crowd" among other things...

Oh wow, how I need that for sure! I did the upgrade this past weekend but was too busy today to check out the new version or even see what has been added, so Unmix Brass is actually my strongest need that I feared would never be met!

How does the track separation in SpectraLayers compare to something like the pro version of Moises (https://moises.ai/products/moises-web-app/)? I qualify for the $199 crossgrade to SpectraLayers versus the $90/year for Moises pro, so it wouldn’t take long for Moises to cost more than SpectraLayers.

Ultimate Vocal Remover 5 is free. If you dig around in the library of additional model downloads there are models that will de-mix vocals, and de-mix crowd noise from vocals. Slightly crude interface, not much support, no real auditioning, but....free and works.

FWIW, SL11 didn't do a very good job splitting up vocal harmonies when I tried it.

stubbsonic wrote:FWIW, SL11 didn't do a very good job splitting up vocal harmonies when I tried it.

I haven’t used Spectral Layers 11 that much, but isn’t splitting harmonies more of a job for something like Melodyne? As I recall Spectral Layers splits up types of instruments, like separating vocals from drums or bass… not separating different vocal parts from each other?

James Steele wrote: ↑Mon Aug 26, 2024 12:27 am I haven’t used Spectral Layers 11 that much, but isn’t splitting harmonies more of a job for something like Melodyne? As I recall Spectral Layers splits up types of instruments, like separating vocals from drums or bass… not separating different vocal parts from each other?

It's a feature they added in v11, but it's kind of a bust at this point. Perhaps it will improve in future revisions.

Ahhh thanks! I had no idea they were attempting to do that now!

Fortunately, so far I have not had a super-strong need for splitting vocal harmonies into stems, as most of the all-on-one-mic recordings delivered to me (or ancient two-track masters) have little if any harmony vocals, but I do have some old projects that are waiting for such functionality to exist (and to work well), even if a two-pass approach is required.

I haven't had a chance to use v11 since upgrading. Super busy year so far, with no time for music production, though that is about to change. I've been hyper busy as a musician outside of my day job and the money pit house that takes so much of my time and retirement funds, but it has been as a live performer rather than studio work per se, for this calendar year.

Just adding to this general Spectralyers thread.

I was using a 8dio solo brass patch for Kontakt, and turned on only the close mic, turned off all reverb that I could. But unfortunately, it was either recorded in a large space, or they baked whatever reverb they used into the samples. BTW, I really don't like that library, it's really annoying to use, always ignores release, and ugh, I could go on.

OK, then I tried using Spectralayers to remove the reverb, but it really didn't work at all, even though the direct signal was high in relation to the reverb.

I ended up using Acon's De-Reverberate, and that worked pretty well-- though applying it destructively moved the audio of the track earlier by a significant value (400 ms?).

Interesting, it hadn't occurred to me to try to use de-reverb to deal with sample library deficiencies, as I probably assumed it wouldn't be the same as dealing with live miked audio, but I guess it's close enough that note to note relationships and harmonics would trigger the same AI algorithms so that in theory it should work OK.

I agree with you about 8DIO libraries, but of course that is an entirely different topic.