Loading...

How a Soundset Gets Made

Home / Sound Design / How a Soundset Gets Made

I thought that perhaps, either correctly or incorrectly, it might be interesting to know a little more about how a The Unfinished soundset is made. Not the ins and outs of synth programming as such, more the process for deciding what soundset to create, with which synth and a little about the decision making process and inspiration along the way.

Let’s start with the origins of a soundset. There are a few reasons behind this initial decision. Some soundsets come about because of bespoke work I have done. Obvious examples of this are Diva Ex Machina which features the sounds I designed for Nathan Furst to use in his Need For Speed film soundtrack, and Zebra Gravastar which primarily consists of the sounds I created for Jason Graves to use on his Far Cry: Primal score. But there are others. A handful of soundsets are mostly made up of bespoke sounds created for some very talented composers, but where the agreement was that the project they were made for would not be mentioned. So, yes, you ARE using sounds that have been designed for some very high profile projects!

Other outside influences can include suggestions made by users of my sounds. This is particularly persuasive when I’m already thinking of the idea they suggest. But, certainly, if there is enough consistent clamour for a soundset based on a particular composer or soundtrack’s style, then I’m all ears. The upcoming Omnisphere Ferox, which is designed around the sound of Trent Reznor and Atticus Ross, is a very good example of this. I occasionally get more leftfield suggestions, such as deep house basses or somesuch… these are kinda unlikely!

And then there’s simply the case of what do I want to make myself?! One of the great aspects of being a sound designer/composer is that listening to music is such a major part of the job. I love listening to new music. Whether it’s online, something I’ve bought, or simply whatever’s going on in the background of a film, TV show or game, I’m always listening for ideas and inspiration. I have favourites, of course. You’ll see some names coming up again and again in the blurb for new soundsets. But I’m constantly finding new stuff to listen to, in a variety of genres. Anyone who says there’s no decent music being made anymore is simply not listening.

The next big question, after choosing the tone/inspiration for a soundset, is “Which synth?” In the case of soundsets inspired by bespoke work, this decision is already made for me. However, for ideas made from scratch, sometimes this can be a really simple question, when there’s a particular synth that lends itself perfectly to a soundset idea. Again, Omnisphere Ferox is a good example of this. It’s difficult to think of a way of achieving so much of the hybrid acoustic nature of the Reznor/Ross sound with any other synth really.

Where there are a few synths that could work for an idea, things get a little trickier. Considerations include whether I’ve already covered that type of sound in one synth enough already, how recently I’ve previously released a soundset for a particular synth (have them too close together and it can seem a bit predictable) and, of course, there are the “business” considerations as well, such as how popular are the synths in question? A good idea might be wasted on a more obscure synth that not many people own. But, equally, there needs to be some variety in my output. I can’t always focus on just the bestsellers! I’d get just as bored as my customers.

When I have a good idea of what synth I’m using and what the influences are for the soundset, I go into a stage of deep listening. Even when I’m very familiar with the material that will inspire a new soundset, I still give it a thorough listen. At first I’ll just absorb, not focusing on anything specific, just letting the music run through my mind; taking it all on subliminally. There’s usually a lot of tracks to listen to, so this can take some time initially. I’ll then follow this up with a more detailed and focused listen to a number of tracks, often making notes as I go along about types of sounds that feature heavily, their tones, timbre, how they develop and respond; and I’ll begin thinking about how I can make similar sounds with my chosen synth.

And “similar” is the key word here. I’m not looking to emulate or copy, I’m trying to capture the vibe, the spirit, but allowing room for doing something a little bit different with it. I’ve been asked a number of times to do bespoke jobs to recreate sounds people like, and I almost always turn them down. I find that an extremely boring way to work.

Then it’s on to sitting down in front of Cubase, with a few instances of the synth open, so I can move between ideas if need be. I’ve reached a stage in my career where I don’t necessarily start from initialised patches. There’s often no need to. I have a collection of what I call “crossroads” patches. These are mostly relatively simple patches that capture a basic tone or function (maybe a bass or pad with a particular combination of waveforms, an arp/sequence that has a specific dynamic that I like to use, or simply some sounds with filter controls that I like). These sounds are too vanilla to make it into a commercial soundset usually, but they have enough useful detail built into them that means there are wide range of different directions I can take them in to create the sounds I want to make. They’re building blocks that help me move quickly, rather than start from scratch all the time.

Obviously, I have less of these “crossroads” patches when working with a synth I haven’t used before. So, if you’re ever wondering why it can take me so long to create a soundset for a newly released synth…

Throughout the design process, I’ll refer back to the inspiration music regularly. There can often be something I’ve missed, or a new variation on a sound will strike me. It’s a dangerous business though. With Omnisphere Ferox, I had initially planned to make only 150 patches. But returning to the source material of Gone Girl and The Girl With the Dragon Tattoo again and again kept inspiring me! And soon enough, I had 300 patches created. Ah well… it was all good fun.

I think I’ve got more ruthless over the last couple of years. Before, I would occasionally let through sounds that I really liked, but weren’t 100% “on message” for the soundset. Now, I don’t do that. With Omnisphere Ferox, for example, I would review sounds at regular intervals to make sure I hadn’t strayed too far from the path. It’s not so disastrous even when I do. The handful of sounds that didn’t make the cut for Ferox found a nice home in two other Omnisphere projects I’m working on. Perhaps they were subconsciously at the back of my mind the whole time?

There are few other interesting little aspects to synth programming that are worth a brief mention. One of them is “happy accidents”. Yes, I’m prepared to admit that sometimes the cool sounds I come up with are more by mistake than design. Such is the nature often with synths with great modulation options, that accidentally clicking on the wrong parameter or hooking up the wrong option to be controlled by an LFO or envelope, can result in something really cool and interesting. I am not ashamed to roll with that, it can often open up brand new avenues for ideas. There are some synths that even have randomisation functions built in, Absynth and Massive for example, that can sometimes be fun to play with when you want to throw something leftfield into the equation. More often than not though, they can be way too leftfield and you’re left with a noise nobody wants to hear. Or in the case of Absynth’s randomisation function, a noise that can split a skull at fifteen paces!

Another intriguing element of programming presets is a particular bête noire of mine: naming the damned things. I’ve created several thousand synth sounds and samples over the years and coming up with fresh names for them is… very trying. A lot of commercial soundsets adopt pretty conventional, user-friendly, but somewhat repetitive naming procedures. The time when such a luxury was open to me is long gone. As such, I’m sure some of the more eagle-eyed amongst you have noticed repeated patch names across my soundsets, though I do think I’ve managed to keep them to a bare minimum… somehow! One that sticks out in particular is the patch name ‘Hotorget’, that I believe I used in both a Massive and an Omnisphere soundset. Hotorget is a stop on the Stockholm metro system that I like the pronunciation of (Her-tory-et).

The current process for naming patches most often involves having a Wikipedia webpage open and hitting “random article” until something interesting crops up. Yes, that is how I currently do it. Given how random the process is, it always amuses me when customers congratulate me for naming my patches so helpfully! Ah well, perhaps there is some latent relevance in the choices I end up making? I’d like to think so. Although occasionally inspiration just appears from the ether and a patch name strikes me. Sometimes, I’m clearly tired when this happens though, as the unusual Omnisphere Ferox patch ‘Derek Sportspants’ will attest.

If anyone wants to create a useful random name generator, I’d be very happy to beta test it. There are a few online, of course, but they are very niche and very limited in their vocabulary.

And that, more or less, is the process of developing a soundset. I’m sure I’ve missed some aspects out, but I guess that leaves some wiggle room for a follow up article, perhaps? Alternatively, feel free to ask any questions below in the comments and I’ll follow them up as best I can.

Comments(4)

  • 26/05/2016, 3:56 pm  Reply

    Do you use your own original samples much in your soundsets, or are they mostly derived from manipulating the synth you are working with? If you use samples, what are some examples of raw sources you have recorded to create your sounds?

    • Matt
      26/05/2016, 4:47 pm

      Great question. I haven’t done much of that so far. I think it’s a matter of what is necessity for me. Take Omnisphere 2 for example… So far I’ve been so absorbed using the wealth of samples/soundsources that come with it (especially the new stuff for Omni 2), that there’s been no need to use my own samples.

      Having said that, I plan to in the future. I have a collection of hardware synthesizers and ethnic instruments to have fun with. And I also want to dos ome more experimentation with field recordings and found sounds.

  • 28/05/2016, 3:20 am  Reply

    How much time it usually takes you to make 150 synth patches Matt ?

    • Matt
      29/05/2016, 11:24 am

      How long is a piece of string? Haha, no, seriously… It depends on so many factors: which synth, what types of sounds, what else I’m working on. I’ve done 150 sounds in a couple of intense weeks and I’ve taken a few months to do it. On average though, I would carve out a good 4 to 6 weeks to do a soundset of 150 patches.

Leave a Comment