Groovesetter

The Routing Matrix Window.

Groovesetter’s developers claim that it offers low‑latency, multitrack collaboration between multiple people in different locations. Does it live up to its promise?

Across the world, the Covid‑19 pandemic has accelerated the rise of working from home or another remote location. It would be easy to assume that the creative industries would be more resistant to this trend, after all, there’s a world of difference between sitting in a studio co‑writing a song and dialling into the latest analysis of a company annual report. But a scan through a few Inside Track articles would show how the recording and production side of the music business has been working remotely for quite a while, albeit in a very back and forth way. Files, tracks and whole projects are regularly passed across countries or continents, between mixers, producers and artists to be updated, re‑uploaded, and shared again.

If all of that sounds like a process that might sap the creative juices, well, you’re probably right. So as we were all thrown into the world of lockdowns and travel restrictions it was no surprise to see tools like Jamulus (see our guide in SOS Sept 2020: https://sosm.ag/UsingJamulus) and Jamkazam pop up to facilitate an online jamming/practice environment. Back in May 2020 we also wrote about running a remote recording session (Remote Music Production: https://sosm.ag/remotemusicproduction) and how you could combine a few existing tools to coach a great performance from an online collaborator.

The next logical step that we’re seeing is platforms start to integrate two‑way interaction with your DAW and your remote performers. And this brings us to the subject of this review: Groovesetter.

Groove Machine

Groovesetter promises the ability to mix, record and integrate multichannel, uncompressed audio from multiple remote locations, with no added latency. That’s right, we’re going beyond just remote stereo inputs and a two‑track output to multichannel inputs and recording direct to your DAW. This means you can create a guide track in your DAW and have multiple remote participants playing along at the same time whilst recording their individual feeds directly to separate tracks in the same DAW project. It also means you don’t have to lock the singer in the booth — you can keep them in a different city entirely.

But moving beyond ‘Zoom but for music’ adds a degree of complexity, to both system requirements and setup and installation. For Windows users in particular, it’s a multistep process, with a few potential pitfalls along the way. Having said that, during the course of this review the developers have taken on board my feedback: a new version corrected some of the issues I encountered, and further enhancements, addressing a few more hitches described below, are in the pipeline too.

I’m going to make no bones about this; setting up was, for me, a complete pain in the posterior! For Mac users the process should be a little simpler but, as a Windows user, this product is still a long way from being plug‑and‑play. Whilst I’m now confident that I could walk anyone through the setup process in around half an hour, it took two sessions with the developers for me to reach this point, and it’s not something to be approached without a fairly experienced user at both ends and the addition of a screen‑sharing/meeting tool like Zoom.

Before installing Groovesetter itself you have to install the (free) Jack Audio API, but not any version — you need the version from the Groovesetter site — and you’ll also need to remember to change the install path (I’ve suggested that a more obvious prompt be added to the instructions at this point).

Once you’ve installed Jack and Groovesetter, to create or join a remote session you’ll need to configure your interface in the settings for that session. Groovesetter recommend using an ASIO device as your interface in Windows, but there’s still a bit of work needed on the app to automatically find and display the relevant drivers in the dropdown list. There is a manual process for this, but at present it’s neither intuitive nor covered in the support tutorials (see box).

So far then, the process hasn’t reached the one‑app‑to‑rule‑them‑all promised land. Fortunately it does get better from here...

Into The Groove

Once you’ve got your co-conspirators up and running, you can add them as friends, or Groovers as they’re called here (strangely, this doesn’t allow you to see all the extra detail that they may have added to their profile), and get on with setting up your session.

Creating a session, like completing your profile, gives you another hint that some real thought has gone into a professional’s requirements. As well as fields like session name and project name (all sessions with the same project name are automatically filed in the same folder), there are additional fields covering what kind of session it is, what type of output you’re recording, parental warnings, the type of recording, the type of composition, record label details, and there’s a free‑format box for general session details. It’s also here that you can add your Groovers so that they’ll receive an automatic invitation in their session list. You can copy and edit previous sessions — this greatly speeds up the process — which is a good thing, as you’re likely to do this a few times as you refine your settings.

The Session window allows you to specify various details about the project, as well as to search for and join existing sessions.

When you’re ready to go, you start your groove from the sessions panel (or join the session if you’re not the host). This opens the session settings window where you select input and output device, sample rate, project buffer size, input and output channels (up to the limit of your subscription plan and interface), and virtual I/O (up to your subscription limit) — these are your DAW inputs and outputs. The session host determines the sample rate and word length and everyone joining is automatically set to that rate irrespective of how they might have their own interfaces configured, which does cut out another potential argument...

The session window is split in two, with the left‑hand window showing the basic details and Groovers, and the right‑hand panel displaying the routing matrix — I’ll come back to that in a moment.

The left frame will show you who’s connected into the session, along with the latency (in ms) of their connection to the host, which will be driven largely by geography. In my tests, other computers on the same local network reported latency of 2‑3 ms, a connection between York and Manchester (about 70 miles) around 14ms, and a connection from York to Norwich (170 miles) registered 18ms. Along with these latency reports there is the option to refine each connection by reducing the jitter buffer. This starts at a generous 74 samples but, in conjunction with the session buffer, can be brought down to as few as eight samples before the audio quality is affected.

I need to point out that “In conjunction” is doing some heavy lifting in that sentence. Although the jitter buffer can be easily refined within a session (with only a short disconnection of the other Groover), changing the session buffer means ending the session, creating a new one (I told you that copy function was useful), and starting again with a new buffer size. You might need to do this two or three times, balancing the jitter buffer each time, to get to the point of minimum latency without compromising sound quality.

Once you’re all connected and ‘minimally latenced’ (which I appreciate may not be good grammar) you move on to the routing matrix in main window. As you’d expect, this defines what audio signals are routed where with a simple pin matrix: inputs along the top, outputs along the side. As a session host, you see the ins and outs of everyone in the session, but as an attendee you see only your interactions with the session host — a factor of the peer‑to‑peer nature of the system.

The routing outputs default to a left‑right pair, with the inputs being whatever is plugged into your audio interface. A ‘Connect everyone’ button speeds up the basic process. Deselecting this to customise the routing wipes the matrix, though, so in reality it isn’t as useful as it would be if it left these default connections in place to be edited.

It’s a simple process but as this is configurable by both the host and the attendees it actually gives a very flexible setup. For example, the host could route a stereo DAW track to L+R as usual, along with a dual‑mono track from contributor A; contributor B could then choose to route the DAW track to their left ear only and contributor A’s audio to their right ear. Or they could take a single mono track sent by the host to their left channel and route it to both ears. Inputs can be renamed and colour‑coded for ease of navigation and consistency with any conventions you might use in your studio.

Once you’ve got your buffer down, your jitter optimised, and your routing configured, you’re ready to go. It’s worth reminding ourselves here that Groovesetter is a recording tool as well as a connection tool. Each channel coming into the host can be recorded and the session window has a built‑in visual metronome. Hitting record means every primed channel is saved to your session storage as a separate, uncompressed WAV file (up to 32‑bit, 192kHz) in a logical folder setup (see 'Serious Stuff' box).

You can also route any inputs back into your DAW to whatever channels you have set up and armed for recording. I mention the ‘armed for recording’ bit here because the only way I could get Reaper or Cakewalk to show up in the routing matrix was to have a track permanently armed for recording. I got into the habit of just tagging an unconnected track onto the end of the project and leaving it primed the whole time.

The author, in an online session with his bandmates.

Groove Is In The Heart

Promises are easy to make but, in the world of low latency, notoriously hard to keep. So once you’ve actually hit the red button, how does it work?

Unlike Zoom, for example, where everyone connects to a central hub, Groovesetter works on a peer‑to‑peer basis, with remote contributors connecting directly to the session host. This approach has pros and cons. The main pro is that it allows separate, optimised connections to each contributor. So if I’m hosting a session with two people, let’s call them Bob and Kate, Bob’s connection doesn’t get slowed down if Kate’s isn’t as slick. And Kate won’t be affected if someone accidentally unplugs Bob’s router.

The big con, however, is that latency between Bob and Kate is the sum of the connections between Bob and me plus Kate and me. And that brings us to the crux of this review. In practice, despite extensive playing around (sorry, configuration) of our various buffers we weren’t able to get the round‑trip latency between host and contributor down to below 100ms without detriment to the sound quality. On one session we pushed it down to about 80ms but the recordings were peppered with pops and crackles. This was consistent over a series of tests with a set of fairly typical UK connection speeds of around 35Mbps down and 8Mbps up. Interestingly, one user switching between wired and Wi‑Fi connection didn’t appear to make any measurable difference to overall latency.

Of course, this being a peer‑to‑peer system meant that the latency between peers was then at least 150ms. That’s too much for any kind of multi‑participant jamming activity; there’s no way Kate can react to Bob’s part if she’s hearing it 100ms after the master track or host performance. Should Bob then attempt to listen to what Kate’s playing? With an additional 100ms delay, he’s going to be similarly challenged as to who he’s following. It’s confusing enough just writing this paragraph but Figure 1 should help clarify things.

Figure 1.

Jam Tomorrow?

So if we’re not (yet) able to jam effectively, what is this software for? Well, firstly let’s not underestimate the value of being able to record multiple parts remotely from multiple locations. If Bob and Kate don’t need to hear each other to play, then their latency to each other doesn’t matter. And if they’re playing along to a pre‑recorded piece that I’m sending from my DAW, then their latency to me doesn’t matter either. I’ll need to do a time correction on their parts after they’ve recorded but that’s simple enough, especially if we’ve taken a couple of minutes at the beginning of the session to sort out and measure the latency.

Secondly, I think this is where Covid‑19 has thrown us yet another curve ball. I started this test thinking of Groovesetter as being a lockdown solution, but actually its greatest value might be in tackling the pre‑lockdown challenges of remote working. Allow me to hypothesise... What if, instead of me (as a recording engineer) sending Bob the project files for him to download, import to his DAW, record his lead vocals, and then upload to Kate (as the producer), for her to do the same — making some suggestions and requesting a re‑record, and so on around the houses... What if, instead of all that, we were all on the same session, listening to the same thing at the same time? Kate could coach Bob to his best performance, whilst making mix suggestions that I could apply directly to the next take. And everyone would have a copy of the session on their computer immediately afterwards. In a world of deadlines, release dates, and multiple revisions, that kind of product could save vast amounts of time.

“But what about the latency?” I hear you ask. How can Kate hear properly what Bob’s doing if she’s dealing with the delay of LA to New York? Well, that’s actually really easy to deal with as soon as you stop trying to make Groovesetter a jamming tool. At the beginning of the session I, the recording engineer, can take a quick measurement of the latency times between me, Kate and Bob. I can then apply delays both to my playback and to Kate’s feed so that we’re all hearing things in time with Bob’s delivery. See Figure 2 for how that could work.

Figure 2.

The Bottom Line

When I first looked at the pricing plans I wondered who Groovesetter was aimed at. The video tutorials showed studios rather than individuals as the remote collaborators, and while I’d taken that to be artistic licence, in retrospect it’s probably a better suggestion of the potential market.

I started this test thinking of this software as being a lockdown solution, but actually its greatest value might be in tackling the pre‑lockdown challenges of remote working.

I raised a number of bugs and potential improvements with the developers, and still think a lot of tidy‑up work is needed if they’re really to get this software up to the standard it needs to be. Yet there’s clearly strong potential here. I have a final suggestion for them: as things stand, the real‑world latency I experienced undermines the current advertising pitch, so stop trying to reduce the latency, focus instead on getting a smart latency‑compensation tool built in (this is already in development), and pivot the marketing to support the workflow improvements that this tool could deliver to the modern professional recording world. Businesses may be changing post‑pandemic, but time is still money and the potential saving here is huge.