I’m trying to set up an Aggregate Device on my new Mac Studio to use my two primary interfaces as well as a couple other I/O devices (adat expansion) as well as Roland Boutiques… basically my channel count was around 110 I/O. I’ve been trawling the forums and find some older threads mentioning Aggregate Devices are not officially supported, but many of these threads are several years old and related to Cubase 11 or earlier… Some I did find which said there should be support for 256 channels of I/O, which I’m nowhere even close to.
Anyway, the issue I’m having is it simply will not work. No playback, no I/O. But if I remove the second interface from the Aggregate Device, reducing the overall channel count to around 66 I/O, suddenly everything works fine. But basically might as well not even bother with the agg device and just use the main interface on its own.
FYI I know someone will mention something about latency, but I do have the primary devices and ADAT expanders all clocked together via high quality word clock cables, and set the first and primary interface as the master clock. So I think I should be fine there.
Other tech specs - Running Cubase Pro 12 on a MacStudio M2 Ultra (Ventura 13.4.1)
Aggregate devices have always been kind of hit or miss. I dont think there is much in the way of support for it as that would be the OS itself handling the device and ‘translating’ the information from DAW to devices to split it properly and whatnot…
Your DAW should just see a ‘system driver’ for your audio interface, so beyond that its how well the system handles merging them together, bad drivers=bad performance. It was a great idea, but the few times over the last 20 or so years I’ve tried them, there were always some hiccups or quirks with them (even in Apple’s darling Logic Pro).
Hmm, interesting, I only discovered Aggregate Devices were even possible about a year ago after adding a second interface for more outboard expansion, and when using Logic which had been my primary DAW, it’s been flawless so far. And hoping it will work even better now on the new M2 I just got
I’ve started using Cubase more for scoring work because I really enjoy how it works with piano roll and virtual instruments, and therefore using a ton of outboard isn’t a dealbreaker in my particular use-case scenario. But I’d like to start consolidating a lot of my work into Cubase and was hoping it would play well with the aggregate device so I can keep all my synths plugged in (what I use the second interface for) without needing to do any additional patching. Not the end of the world, but just a little bummer it doesn’t work.
Yeah aggregate devices were one of the big things about the ‘new’ CoreAudio layer when OS X came out. I think it was a great idea, but my guess is there is still some functionality missing or its just got too much to deal with.
It needs to combine the two units at the system level, so it appears to your DAW as one single ‘system device’ instead of two separate units. The DAW only sees the Core Audio ‘system driver’ and has no clue what is connected on the other end. It just sends what it needs there, so its up to the MacOS to actually handle what comes from the DAW and ‘translate’ it to the devices, send stuff back, and so forth.
Its probably one of those things that will never be perfect as its supposed to deal with any number of different devices and make them all work as one cohesive unit. Its kind of like Steinberg actually beta testing Cubase with all of the possible combinations of Windows hardware out there. Its just too much for anyone to really do so you get the best compromise.
Interesting. It seems like hardware has come a long way in the meantime. The computer I was using just before this where it worked no problem in Logic was a MacPro 2013 trash can (6-core with 64GB of ram etc). I recently got the MacStudio with M2 Ultra and 128GB of ram. My point being I feel like the hardware has likely by now caught up with any limitations with MacOs having to handle and translate between the devices, so whatever buggy limitations from 10-20+ years ago may not be quite the same issue, I’d imagine anyway.
In any case, worse comes to worse I have additional ADAT ports on my first interface, and MADI as well should I ever go down that road of expansion, so I could in theory use just the one interface to expand my channel count without the need for an aggregate device. The main reason I didn’t opt to do this is I will lose out the console and unison-enabled preamps from my second interface (UA apollo). But at least I could still have those synths plugged in and it should reliably work with Cubase I believe.
Indeed, this has worked reasonably well at some point(Snow Leopard era) with some devices, only to stop working when macOS got an update, becoming more reliable again, failing after an update etc.
It’s not just that it plain doesn’t work, but there will be random weird errors and glitches, especially in elaborate setups with external clocking.
Combining interfaces from different manufacturers is the biggest challenge.
It’s quite possible that Apple ignores bugs and/or has unsufficient documentation leaving it to the device manufacturers to find workarounds.
So Steinberg decided not to support this. It might work fine, though. Test extensively when in doubt.
About a year ago out of idle curiosity I had a play about with an aggregate device on an intel macbook pro. I had access to a Zoom H6 and a Zoom H8 equipped with the Zoom add-on 4 x xlr unit. In other words, a six track and a ten track recorder that if aggregated would be a handy portable setup offering compact 16 track mobile recording. It worked perfectly straight away. I’ve recently got an M2 macbook pro that I hope can also access this same H6 + H8 setup. I haven’t explored aggregate devices any further because I haven’t had any reason to do so but the experience with my very modest portable Zoom setup has left me hugely impressed with Apple’s aggregate device. (Edit: running Cubase 11 and 12)
This is part of what I was getting at in my last post… These devices are class compliant so everything they need to run is part of the USB spec. Plug them in and they just go. Class compliant devices should be perfect for aggregate devices… Its when you run into devices using proprietary drivers that trouble can start, and then when you try to mix and match different vendors drivers together, well… Sometimes things just weren’t meant to work together…
I’m not an expert in the area, I just wanted to mention that I have a master clock device so that is one less problem in the pile, in case it hasn’t any impact on performance (should devices be all clocked internally and independently).
For what it’s worth, if anyone is curious I tapped my second interface (UA Apollo) via ADAT and it seems to be working fine this way, so I can route it to my primary interface (Orion 32) and Cubase sees it all as one.
One concern I did have was not being able to use UA’s Console for unison-enabled preamps and effects, but turns out I am wrong about that. It’s still useable, I just then route those inputs to ADAT out.
There is a latency in doing this which does not surprise me, and will be mostly fine for mixing. When tracking off the Apollo, I can switch over to monitor the Apollo directly without any latency (or minimal latency – rather than monitor it thru the Orion and thru Cubase, which would be intolerable while playing an instrument).
There are some minor trade-offs here and there but all things considered I am able to achieve the setup I was after for my hybrid home studio and don’t have to deal with an aggregate device.
Hello! I’m the originator of the now 2-year-old thread that @zip referenced above. macOS Aggregate Devices are Core Audio-compliant devices. Anything that supports Core Audio should work fine with Aggregate Devices and other than with Steinberg software, this has been my experience. Through all of the issues that (for me) have come up with Cubase, Nuendo, Wavelab, etc., since macOS 11.3 Big Sur, all other applications worked fine, e.g. Pro Tools, Native Instruments, Spectrasonics and several others.
While I definitely appreciate the post by @Monotremata on USB Class Compliant devices and how that simplifies things from a driver perspective, I use Thunderbolt 3 (with optical cables in a couple of instances) for the benefits that this type of connection offers over and above USB-C. Core Audio was designed to provide a layer of abstraction from the hardware device driver(s). Something in the way that Steinberg handles Core Audio broke Aggregate Devices in the macOS version I mentioned above.
I’m now past the issue because of the great folks at Lynx. They gave me pre-beta drivers in Apple’s new Auto Driver Kit (DEXT) format to test knowing that I was having this issue. I found a couple of issues that they fixed immediately and they then made the beta drivers available to all of their customers. ALL audio interface manufacturers will be going through this effort because Apple will eventually sunset kernel extension (KEXT) drivers. Hoping that the move to DEXT drivers solves this Steinberg-specific issue for good.
I just wanted to summarize my own experiences. I hope that others find this useful.
Man I really hope MOTU gets on that soon. The drivers for my 828x have been popping up that “The dev needs to update their drivers” warning for like 3 years now!
I found a strange workaround for this utterly frustrating issue on my new rig I’m building today - which might help on some systems. It could just be unique to my rig due to my particular setup to be fair. But it’d be interesting to know if this helps anyone out…
I’m running Ventura 13.6, Cubase Pro 12, a 2018 Mac Mini and I have two interfaces: A 2nd Gen Focusrite 18i20 and a Pre Sonus 1824c. Both have a couple of Behringer ADA8200s slaved off them for extra analogue IOs for my outboards. Was having the same issue. Either no sound or weird bitcushed nightmare when aggregated together as a device. But everything was working fine on my old rig (High Sierra and Cubase 7.5)… Pretty exasperating…
What I found was that when I opened a brand new C12 project, the main/ default Stereo Out bus that’s created in Audio Connections was completely dead. No meter displays on playback or any activity as mentioned by previous users above. Dead as can be…
But, if I created a new Stereo Output bus (Stereo Out 2) and then routed the Device Ports on that to something redundant on one or other of the interfaces (the SPDIF Outs in my case), it kind of felt like it ‘reset’ things and stuff started working again on the default Stereo Out bus that was created by the project.
This is an admittedly completely dumb workaround, but I was despairing a bit like the other folks here, thinking that I was better off with my old rig and shouldn’t have bothered upgrading my Mac or updating to C12.
Anyway, after hours of arse-ing around in Audio MIDI Setup. Updating firmware on my interfaces etc. etc. Simply creating a new/ extra Stereo Out and then routing that output bus to an arbitrary physical IO (and only after routing - not after only creating?!) - brought the system into life and everything works. Both interfaces interacting with the DAW normally and blasting audio all over the place, External FX, extra busses in and out etc.
Anyway, if this can help anyone out - then it was worth posting…
If anyone hears of anything more official from Steinberg though, it’d be great to know about it…
OP here – Thanks for sharing this workaround. I was actually pulling my hair out a few nights ago when after turning on one synth (which was part of the aggregate setup), Cubase went suddenly dead as you described! I tried other projects, switching things on and off, restarting, etc. Still dead. The only thing I found that fixed it was switching from Aggregate Device back to just my primary interface. It’s not the end of the world, I lose some I/O but I have a patch bay so I can always manually deal with it. But it’s still annoying.
I’ll give your workaround a try. Maybe C13 will have some fixes under the hood which will address this? I’m no expert in saying this but I do feel like hardware is fast enough at this point to handle aggregate devices, so it’s perhaps a matter of software catching up. In any case thanks for sharing!
i made a long time ago (im on cubase, windows) a Feature request for next Gen ASIO Driver that can work with 2 or more audio devices, or at least for steinberg/yamaha devices including yamaha’s keyboards with Audio devices.
but seems nothing is happening this regard, or its impossible, or its not profitable enough to spend recourses for this.
anyway could be cool for functionality and also for company like Steinberg that invented some of the standards of music tech, to do such a move.
steinberg become in my opinion kind of “Boring” company with no new real inventions in its sleeves.
That’s an interesting fix you’ve found, especially with interfaces from different manufacturers. Thanks for posting! So many of us have been banging our heads on this issue for a couple of years now. Until Lynx released their new DEXT drivers my workaround was similar to @wing, i.e. I had to pick a single interface (I have 3 total) instead of the aggregate and then deal with the routing mess that this created. I’m interested to see what Cubase 13 brings.
No probs at all. And I feel your pain!! Extensive physical I/O is a must for me due to the nature of my work/ clients etc. So I really was thinking I’d hit a massive obstacle with this - I can’t really pair that side of my system back…
I definitely just stumbled on this workaround to be honest, after hours of what sounds like a similar experience to you. Switching things off/ on - building different Aggregate Devices to try and troubleshoot, attempting to narrow down any logical causes to the problem. It was tedious and soul destroying!! And although I think I’ve got a reasonably predictable workaround for my system, something about it still feels shaky… I’ll be stress testing this new rig as I’m phasing my old one over the next few weeks…
It’d be good to know if it’s any use for you on your system. And as you say, fingers crossed that C13 has some more serious implementation from Steinberg’s end on this…
