Anyone here doing Atmos / Spatial Audio mixes?

Hi,
I've been involved with Atmos since very near the beginning. I made reverb plugins (probably the first) that supported Atmos and competing immersive formats. And that meant I had to learn my way around it. I've been making mixes in Atmos since that time and still do--more since I retired from the development business. Because these mixes are classical, they're typically for my own enjoyment or that of the performers. But it's a great format. My room has been 7.1.4 for quite some time.

Your big question about master buss compression has a simple answer. There isn't one that I'm aware of for Atmos and you wouldn't want one anyway. That latter statement is probably hard for you to believe, but the more time you spend in the format, the more you'll understand the reasons. Because objects can move around, and because the true channel count is only determined by the playback system, there isn't any real guarantee of a fixed channel relationship. Master bus compression would do things that surprise you and generally not in a good way. As a matter of fact, it's really best not to think of channels and is better to think of the 3D-ness of it. I remember hearing Steve Genewick (Capitol) asked about compression at some sort of gathering. Steve's been doing Atmos for quite a while and his answer was a simple "you just mix". You really do develop a new way of thinking about energy, loudness and all that. No problem with limiting for a particular problem object or perhaps compression for effect. You can still run those in mono or stereo and that's generally all you need. It's also best to forget there's such a thing as a "bed". That's kind of a crutch for people coming over from stereo or surround and it's full of problems. Put stuff into objects as soon as you know how.

Dolby is a great supporter of the format and they have lots of information (including an ongoing series of webinars with room for Q&A). Start here.

Be a little careful with EarPods. The binaural model doesn't really account for height (and doesn't always represent rear or center very well). It's what you've got now and I understand that. Real speakers will help you get a better idea of how to translate your work in buds. And be very careful of Apple Spatial Audio. Even though they gave Dolby enough money to use the Dolby logo, Spatial Audio isn't the same thing as Atmos and is generally inferior. Just so you know.

But the main thing is to jump in and start mixing. Just like stereo, the easiest way to learn is by doing and following. You'll have fun. You'll need the Dolby Renderer to work in Pro Tools, but it's not expensive. It will also give you a little better insight into what you're doing in Logic. And welcome to a growing group of mixers!

I love mixing in any immersive / surround environment. LOVE IT!!!! I have done huge multi-channel mixes for Disney, Universal, etc.. Spent a big wad of money equipping my previous studio to do it efficentially and elegantly. But IMO, it's a complete waste of effort and $$$$$$$$$. The only monitized source of income for me was LARGE feature films. No artists wanted to pay for it even after showing them how awesome their music sounded mixed in surround. Not ONE. Films even are balking now that they are trying to pinch every penny. Smaller films are fine with stereo stems. The interest in surround/immersive is hyped in the media, and virtually non-existent in real life situations IME - with the aforementioned exception of large feature films. I abandoned doing it when I built the new CRM so that I could narrow down the mixing environment to be as good as it could be. As hard as the tech companies try various types of surround / immersive audio, it never seems to catch long term. Quad, Dolby surround, 5.1, 7.1, etc., etc., etc.. All are rare to extinct in my world at this point.... Atmos is the lone holdout, but having a legit atmos mixing room is a significant commitment. Especially as it significantly compromises other aspects of my work environment.

Super agree with that based on hearing it at one of Apple's theaters in Cali.


seawell the Working Class Audio podcast guy Matt has been on the journey and has talked a ton about it on his episodes. Might be helpful as well as he's deep in the community and such.

Big, big difference.  I'll agree that quad was pretty bad, since it relied on a simple matrix to extract that rear stuff from a stereo pair. But the big problem was that it was channel-based (and analog).  You had to have 2 channels or 4 channels. Object-based like Atmos is a completely different thing.  There's a single deliverable that also contains stereo and 5.1 and can easily be read by older systems. The object-based part of the deliverable does not contain a predetermined channel assignment--only spatial replacement.  The actual render takes place in the repro system. That might be your A/V receiver, a soundbar (some are excellent and some are bad), your computer, your theater. If you've got stereo at your end, it will fall back to the included stereo mix. If you've got a big home theater, the renderer will give you discreet speaker output that matches up to your room. The metadata that's part of the Atmos deliverable is interpreted by your repro system to match up to what you've got.  In quad, you had to have a very specific system layout.  Atmos will tailor itself to what you're listening with.

One of the big problems with quad (and later on with 5.1) was the inventory issue.  Whoever you were buying your music from (I do remember record stores) would have to provide the same album in any of the competing formats that were in the market.  Of course that never happened, so all of those formats failed. Atmos is well tailored to digital delivery and solves the inventory problem. There's one deliverable.

Think of it another way. Atmos targeted for home delivery may have something like 16 channels of mono audio in the file, each with panning metadata. Each of those channels is a sort of submix.  Your repro system will look at the metadata and pan it appropriately to the speakers you have.  But there's some magic in the 'submix'.  The person mixing might have as many as 128 objects with 118 of those objects dynamically pawnable anywhere in the field. When the Atmos deliverable is prepared, these 16 (or more) channels are created from all of the objects.  A particular object may find itself in different submixes at various times, depending on where it currently is in the field.  All this happens without clicks or pops.  There's some pretty slick stuff going on.  The result is well-tailored to the complete range of listening equipment you may have, whether it's a phone or a 100K home theater.  This reduction doesn't happen for theatrical release, since an Atmos theater will typically have many more channels. But even there, the render is done at the theater. Panning and channel assignment can then best match the room, whether it's 16 channels or 64 (or anything in between).

I do appreciate stereo. It's been with us a long time now and it's a good format.  99% of what I listen to is mainly historical and is unlikely ever to be anything but stereo. I'd love it to be in Atmos, but it's not of much interest to the people with the money. But going forward, we have new tools that are getting easier to use. A well-done Atmos mix (for any sort of music) opens up great opportunities both for mixer and for listener.

Quad was a discrete format in the early 70's.  It pre-dates the Dolby 2:4 format - which was horrible - by a few years.

trakworxmastering - I think you are spot on.  Atmos is being led by the people who make the hardware, the music folks who want to do atmos mixes, and by the audio mags - not by the general public.  Until that reverses (which I don't see), I'm out.  I can't tell you how many clients I previewed surround mixes for - both commercial mixes and their own music - who LOVED it.  But no one was willing to spend an extra dollar to make their own music surround.  (5.1 / 7.1 / +)

great question.  In "theory" that should work.  In practice, IME, it rarely works.  Never drives the compression in the same fashion. The same reason that stems - when recombined for film, etc. - rarely have the "punch" of the original stereo mix.  Whoever is recombining stems - the remix engineers on films, or the mastering engineer for records needs to recreate that again on the recombined stem tracks.  It's tricky.

I guess I would wonder why that might be? It's the same exact signal feeding the side chain in both scenarios. Provided you're keeping everyone else exactly the same, it ought to work, at least in theory.

Now one possible gotcha that comes to mind, is how clean or colored your mix bus compression is. I would think you would want it be pretty clean, as doing each stem this way, through a colored compressor, would cause that color to build up on all the stems in a way that it wouldn't be doing if it was just applied to the stereo mix.

But outside of a scenario like that, it still seems like it ought to work. My curiosity is piqued.

This is exactly what I'm planning on trying Quint!  Some of the the videos I've checked out where people are just saving a new version of the session, enabling the dolby renderer and then totally getting rid of whatever 2 buss processing they had going on seems crazy to me. I don't rely too heavily on 2 buss compression at all but I really do like the tone of the Red 3 > AML ezP-1a pair.  I mix into it from the start so I just can't imagine losing all that to start the immersive mix. 

I can already tell it's like the Wild West here so at least I'm not worried about doing something "wrong" haha.  I want to take advantage of the fun possibilities of immersive, but I really do not want the immersive version to be wildly different from the stereo mix.  The thought of getting away from all of these crazy loud master requests is most appealing to me!  

We'll see how this all goes, thank you to everyone for your input so far!  I will add that when spatial audio first hit Music a year or so ago, I checked it out on AirPods and Beats Studio 3 headphones and was really underwhelmed.  Every spatial mix sounded worse to me than the stereo mixes.  Giving it a new try now(with AirPod Max Headphones) has been an entirely different experience.  I figured if I was going to make a decision on it, I should at least try with their top of the line headphones.  

Speaking of headphones...I can't mix to save my life on a laptop/headphones set up(I've tried Slate VSX, Waves emulations, etc..)  My hope with this though, is that I can have a stereo mix that I know the low end is right, etc.. and just move things around and have fun in the immersive space(on the AirPods Max) without screwing anything up.  I'd love the full speaker set up but I do have a lot of the same reservations that others here do about the longevity of it(for music) so I'm just barely dipping my toes in the water here. Having fun so far though!

Summing??  

Wait..NOW summing does something?  haha...I kid..I kid.

Personally, I love consoles, but I've found that Pro Tools is MUCH more adaptable to surround mixing than 99.9% of consoles.

This is an interesting option for a "master buss" processor for atmos mixes: www.hornetplugins.com/plugins/hornet-samp/