Hi Ben, welcome to our pro audio forum. This link explains how you can increase the number audio inputs and outputs available at the same time.

"It also lets you use multiple audio devices at the same time, like several USB microphones".

If this helps, please share your experience:

https://support.app…

Aggregation on a Mac is a very useful feature - for a relatively small number of active input channels. It does a lot of behind-the-scenes data movement, but that's expensive in processing time, so it can fairly quickly run out of bandwidth, even on higher-powered Macs.

It could be that if you aggregate all your interfaces, you could use a small number of channels in a session if they were spread around the interfaces, or maybe more if a session is either studio-sourced or control-room sourced. What aggregation is not so good at is substituting for a hardware patch panel wired to the channels on a single interface.

The second point relates to how you want to connect several 18i20 interfaces to a computer - they can't simply be "consolidated". The normal way for connecting multiple USB devices to a single computer USB port is through a USB hub. However, even at moderate data rates, the drivers in the computer have to work quite hard to pack and unpack the data from a single port. My golden rule is never to use USB hubs for audio work.

I think you are going to have to moderate your demands, tuning them to fit your actual usage. To achieve this, you could look at something like a MIDI-controlled analogue multiplexer (patch panel). This would allow you to have, for example, 64 input channels available for selection spread around studio and control room, of which you used not more than, say, up to 16 at any one time in a session, making it much more likely that your Mac could capture the data without dropping samples.

All of the above assumes that you have a studio master sample-rate clock that you distribute to all your interfaces so that they are all locked together to perform conversions at exactly the same speed.

Good post, Bos. You are a wealth of knowledge.

Thank you so much for your in-depth and helpful response.  This is really what I need. I tried looking into this MIDI-controlled analogue multiplexer idea online and had a hard time figuring out what exactly that would look like.  I researched patch bays of different sorts and how they work (I'm currently messing around with mine), but I was wondering if you perhaps don't mind giving me a brief explanation or examples of how this works?  Is there a product available or is it an arduino thing?

Thanks again,

Ben

Boswell wrote:

Aggregation on a Mac is a very useful feature - for a relatively small number of active input channels. It does a lot of behind-the-scenes data movement, but that's expensive in processing time, so it can fairly quickly run out of bandwidth, even on higher-powered Macs.

It could be that if you aggregate all your interfaces, you could use a small number of channels in a session if they

I really only scratched the surface of what you need to consider in your situation.

I mentioned MIDI-controlled multiplexing, but it's a big topic in itself. Nowadays, a lot of manufacturer offerings are aimed at DJs who want to go further than standard mixing hardware will let you do. It means that modern multiplexing products at the lower end of the market are not going to be of much use to you. Equally, the high-end of this is mainly based on Dante, and is aimed at sound designers of theatre and stadium concerts. So, somewhere in the middle, there might be one or two products that could work for you.

Many years ago I was called in to set up a MIDI-controlled switching system for a mid-sized studio. The specification for what they required had similarities to what you described, but there were some significant differences. The number one requirement was that the configurations had to be stored and recalled electronically, so that not only could each incoming band have their own set, but each music track (song) that the band laid down would have its own configuration. The configurations had to be instantly recallable, so that a quick overdub could be done without using up unnecessary studio set-up time.

I was given a budget that was ridiculously small for the equipment that would conventionally be needed to meet the specification. The situation was saved by me finding a Yamaha 01V96 mixer stored in one of their cupboards. Since I used one of those as an auxiliary mixer in my own studio, I knew them inside-out, and realised that I could press this one into service as a 24-in, 24-out ADAT multiplexer using stored configurations. The mixer would not do any mixing, only routing, carefully avoiding using the rather average-quality ADCs and DACs in the unit. It all worked well, but not without some difficulties, such as having to move all the ADC/DAC converter boxes to be within ADAT cable range of the mixer. However, it really boosted the studio throughput. I learnt later that they were surprised by how much time was saved by eliminating configuration errors through having exact configurations stored from previous sessions. I believe they were still using the setup up to the time when the studio, like many others, had to close for financial reasons.

It's not impossible that you could find an 01V96 (or one of its bigger brothers) on the second-hand market for not much money, and then connect the ADAT outputs of three of your 18i20s to it. You would use Focusrite Control to route the ADC outputs of the 8 analogue inputs directly to the ADAT outputs in each 18i20. In some senses, this pushes the problem further down the line, as you would still have to deal with the 24 possible ADAT channels coming out of the mixer. However, your 4th 18i20 could handle a chosen 8 of these plus its 8 direct analogue inputs, giving you 16 input channels, with half of them chosen (multiplexed) from 24 other inputs. This would avoid using device aggregation in the Mac by making it external.

I'm giving you a description like this to illustrate the sort of ways you may have to be thinking to get yourself a working setup. Flexibility is the keyword. There's no instant solution to what you say you need, but I hope there's some food for thought here.

I should point out something that it is possible you may already have tackled: the 18i20 does not have a BNC wordclock input, only an output, but it can lock to ADAT or S/PDIF inputs. That means you could consider running a 75 Ohm BNC cable with T-pieces to daisy-chain the BNC S/PDIF inputs of 3 of your 18i20s to the S/PDIF output on the 4th. A 75 Ohm terminator would be needed on the remote end. You would use Focusrite Control to set BNC S/PDIF clocking. Which generation of 18i20s do you have?

Boswell wrote:
I really only scratched the surface of what you need to consider in your situation.

I mentioned MIDI-controlled multiplexing, but it's a big topic in itself. Nowadays, a lot of manufacturer offerings are aimed at DJs who want to go further than standard mixing hardware will let you do. It means that modern multiplexing products at the lower end of the market are not going to be of muc

I'm very impressed with your expertise and once again really appreciate the thoughtful and helpful response. I will definitely consider the route you described. I love the idea of being able to instantly recall different configurations and this could very well be the ticket for me in the near future. However, do you think I could create a more makeshift workable solution temporarily using a patchbay until I've saved up the requisite $ for the 01V96? Would be great to have a nice MIDI board functioning as you described but I'm running on limited loan funds for now. A much less sophisticated idea occurred to me and I wonder if it would work; I was picturing something like the whole 32 inputs mapped to the bottom row of a pair of patchbays (one 16-channel and one 24-channel)

32 of the 40 channels would be linked to the interfaces, with 8 slots left over for outboard gear. 1-16 being the master interface+adat channels from second interface; 17-32 being mapped to the other two interfaces, and 33-40 being available for effects and processors. Channels 17-32 *can* be used as-is if I select interfaces 3 and/or 4 as my input (could be useful in multitracking), but for band tracking I would mostly just reroute signals on the front of the patchbay, so that whatever mics I want for the sessions are being fed from the front-facing side's top row into the inputs for channels 1-16 on the front-facing side bottom row with patch cables. With a very limited grasp on audio routing, this method seems manageable and relatively convenient for the price. Because there is a semi-permanent default mapping, all it requires is simple patching to reroute for individual sessions. The thing I am concerned about is signal degradation since, in order to piece everything together, the setup would involve using more cables and hardware connections such as XLR->TRS adapters (to patch individual mics, a faceplate, or my drum snake into the back-facing top row of a patchbay), an XLR faceplate (as a means to access ins on the back of the patchbay from the front), and long-ish cable sends due to the distances in between the interfaces in the control room and the interfaces in the tracking room (~15-20ft).

Wondering if all this makes sense. I'm no studio expert so I apologize if I'm missing any critical info or understanding.

And to answer your last question, my 18i20's are two 3rd gens and two (I believe) 1st gens, (I wasn't able to access the serial numbers to verify they're not 2nd gens but I'm pretty sure they're 1st gen).

Hardware patchbays are useful for manually routing line-level signals. They are really not recommended for microphone inputs, for two reasons: (1) microphone signals levels are usually low in amplitude, meaning that the extraneous noise, clicks or hum that patchbays can introduce are likely to be much more audible in the signal being passed, (2) you must not route phantom power through either 1/4" or bantam jacks, since the act of plugging or unplugging invariably causes a brief short-circuit from tip and/or ring to sleeve (ground).

As I understand your latest post, you are thinking that you could group your 18i20s into a studio pair and a control-room pair, each pair consisting of a master (analogue channels 1-8 of the pair), and a stand-alone unit, configured to copy its analogue inputs to its ADAT output, which is in turn connected to the master as channels 8-16. That would work well as long as you did not require to vary anything dynamically in the stand-alone box. It would reduce your connected devices to two USB 16-channel units, with all 32 inputs available on composite connectors (XLR/TRS).

You may find that aggregating two 16-channel boxes involves significantly less overhead that four 8-channel boxes. However, even if you were to be careful about using channels from only one of the master units, it's possible that all the data from the other box would be captured as well and then thrown away.

Before making a decision about what to do, I think it would be worth doing a few experiments. Firstly, position all 18i20s side-by side and plug them all into the Macbook (assuming you have 4 USB inputs available). Aggregate them, and attempt to record 32-channels, with the only analogue input being the same audio signal connected to two channels in different boxes. Process the captured signals by inverting one signal channel and subtracting it from the other to see if it manages to deliver clean data. Then unplug the USB cables of two boxes and repeat, this time with only 16-channels total. Finally, try a single box (16-channels) with no aggregation.

Wondering if an important consideration is being overlooked, though there are some very good suggestions?

You wrote:

"...way too many things daisy-chained via USB (midi controlers, interfaces, hard drives) in my studio, including four focusrite 18i20s hoping to be able to get all 32 channels out of them. but I know it was wishful thinking since my 13-inch MacBook..."

First off, "daisy chained" is generally "serially-connected", out to in, out to in, etc., but that's nitpicking.

What I see is that you have at LEAST 8 (?) USB devices that you want to run simultaneously to a Macbook? How many USB ports does that Macbook have? You cannot physically install a USB expander card like you might in a desktop computer with an open slot.

A USB hub could allow you to connect a few more things, but the data throughput is shared, so the more you add that are trying to work simultaneously, they will be constrained by the maximum speed of the one (or two) native Macbook port(s), that speed divided by each additional device...such as trying to run all channels simultaneously on four interfaces (even if running one into another one according to the manual), never mind the MIDI devices and hard drives?

If this is an issue that needs addressed, maybe it will spark some ideas?

Good luck.

Kapt.Krunch