Designing the sound of DARQ – an indie game audio deep-dive

 

H i, this is Bjørn Jacobsen and I’d like to tell you all about DARQ, my latest game project which was released in August on Steam and GoG.

I’ve been working on this game for about two years, on and off. Some periods were more hectic than others and sometimes I was waiting in between levels being finished, so it’s not two years of full-time, full-on production. I had other full time projects to work on at the time. So in the beginning it was strictly after hours and weekends I was putting into this.

This game is quite unusual in all aspects. As developer Wlad Marhulets and I discussed the soundscape of the project it quickly became clear that we were on the same page about what we wanted.

Wlad is a composer by heart and wanted to make a video game that was not only his visual and game design idea, but was something that he could score. This made the project quite unique; it’s different from many other projects I have worked on where there is a strict audio direction or production pipeline.

I made several videos and recordings of the sounds during the development of the game, so there is lots to share about the process.

The core game is created in Unity3D and we used Fmod as audio middleware which made implementation easy. As for the actual audio production tools, I’m primarily a Cubase and Wavelab user, but will use whatever is around so I’ve been running things through Transformizer, Reformer, Granulab, Glitch 2, and tons of plug-ins from various sources — I’ll go into more detail on those later.

 

Audio direction

 

The basics

It was both a relief and scary to be able to mold a project this much from the very start. I always like very over-produced sounds and pumpy soundscapes, deep and aggressive. When I found out that Wlad agreed upon this direction, I was really happy to continue my work here.

One thing on my wishlist was to set a few dogmas to follow, and stick to them to the very end. I was quite inspired by the dogmas set by Lars von Trier and Kristian Eidnes Andersen, who in an interview explained how setting strict rules for production helped create unique soundscapes and experiences. And on Kristian’s work on films such as Antichrist, Melancholia and Nymphomaniac , they had musical composition dogmas forcing them to only create sounds out of organic materials, like blowing grass straws, meat, flowers, etc.

That interview had quite an impact on me. I also heard Kristian talk about these kinds of rule sets over lunch one time while he was working on Melancholia and I was working on a short film called Karim. It was truly inspiring to overhear such ideas coming directly from the source.

With this in mind, I made a few rules of my own. Without spoiling the whole plot I will say that all sounds on each level have a reference to either the real world or take place in a very specific location, such as an opera or hospital, so all ambient sounds should be made from recordings or library material recorded at such specific locations, and as many assets as possible were created from such material.

For the Hospital references in the game, I went to Herlev Hospital in Denmark (where my second daughter had just been born around that time) with my trusty Tascam HD-P2 recorder and a Sennheiser 416 microphone. Not an ideal microphone for ambiences, but I used it specifically to record close up sounds of various things, such as parking lot vents, elevators at the hospital, and doors opening and closing. They had these huge sliding doors for driving bedlinen to and from the cleaning facility in the basement.

Since my daughter was born there I also knew where certain items were stored. While recording, I made sure I wasn’t disturbing anyone and I wasn’t going to any restricted areas of the hospital.

We also went to an abandoned school bound for destruction in Allerød which had some nice things left in the building still and an almost fully equipped gym where we could really throw things around

Before signing up with DARQ I also had been on a recording trip with some former colleagues of mine at IO Interactive. They weren’t sound people, but during one of our field trips I had arranged a recording trip to an abandoned “hospital” treatment facility up near Gilleleje in Denmark. We also went to an abandoned school bound for destruction in Allerød which had some nice things left in the building still and an almost fully equipped gym where we could really throw things around, like cinder blocks and lots of leftover metal. A reverberant giant gym is also nice to play around in.
 

The levels

Wlad and I started out by testing our combined skills on the so-called opera level.

This was the level Wlad sent me to test out both my skills and direction before we got seriously started on the project.
Visually, it was slightly different than the level that made it into the game. There were items on the stage that could move with the rotation of the level. You can see a tiny bit of the original Opera level setup in the level video I made about it here:

These sounds are mainly constructed by taking my daughter’s toy kitchen and moving it around on the floor at home. Some sounds I recorded with my Tascam, others with my phone, depending on what kind of fidelity I was looking for. I had a big recording session at my old apartment, which had really bad acoustics with lots of reverb and flutter echo because of the concrete walls. I took advantage of that and placed some of the objects around the apartment directly onto the wooden floor, making use of the special reverberation of the walls here and there.

I had an old guitar that I could throw around, slide on the floor, some chairs, and so on.

For me, it was really important to create the ambient sounds out of the mentioned materials and record as much material as possible using the correct physical objects. But if the player was to leave the controller or stand still, just the background ambience would play and be almost inaudible, with only small subtle changes occurring, or else it would be almost silent.

Yet if the player was to interact with anything again then those interaction sounds would be quite loud and aggressive.

Each level would then have its own set of rules and dogmas that I would follow and the sounds used to create each level would come from the same material.

Sounds should be loud when interacting with items or anything noticeable in the game, but standing still should result in a detailed, but low amplitude ambient soundscape. This made it possible for me to focus heavily on the production of sounds for interactions and really use every bit possible when creating those. Each level would then have its own set of rules and dogmas that I would follow and the sounds used to create each level would come from the same material.

Though some sounds were made in special ways, which I’ll explain a little later.
 

Tools used

Most of the sounds I created in Cubase, but sometimes I used Reaper when I wasn’t in my studio. I also enjoy working directly in WaveLab and have an effects chain there, but the workflow in the newer version has changed and it’s not as effective for design as it was before, for me at least.

This may sound overkill (although done with quite a bit of control), but a lot of processing in the plug-in chain would be purposely distorting EQ’s, adding limiters, then more EQ on top of that — basically using the EQ as a waveshaping tool — reverberation, compression, maybe more reverberation, more compression, maybe even side chain compression with something else, and then render it out as a big file which could then be cut into pieces that I could use for certain things.

On my Patreon and my YouTube channel I have mentioned this technique a lot and called it “lego production”. Not the same as layer cake production of layering things; that’s for later in the process, but simply just creating sounds with certain source material or end material in mind and maybe creating several versions and storing them all in a folder from that day, naming the files correctly so they can be indexed properly and then using that folder as the source material of the later layer cake production to create the actual sound I want.

I use a lot of impulse responses of quite weird sources, I also use impulse responses as a waveshaping tool, so an ultra short impulse can change the way the spectrum appears. If you know how the algorithm for convolution and deconvolving actually works then you can take advantage of that, especially if you have played with those toys a couple of times then you can sort of predict the outcome of certain source material with certain effects and impulses. This not only makes your workflow faster, but it also helps you spice up your creativity and craftsmanship.

A lot of the sounds in the game were made like this — sending source material through various effects to create small libraries and using those for the final product.

I also took advantage of how video game audio actually works and how triggering assets can be split up because you have the randomization factor to play with, which is quite different from working with linear material. Even if you have a linear video representing the visuals while creating the sounds, you can create layers and variations of the layers individually. So when putting the layers together, you don’t just have two or three versions to randomize between; you can easily have four layers each with five variations for a total of 625 variations.
Lots of the sounds are split into variations like this, so even if you play the game twice or do the same action several times, the sound you hear will be just that tiny bit different from the last time you heard it.

For the impulse responses, I used a very easy-to-use, free VST plug-in called keFIR along with a lot of SSL channels or SSL bus compressor plug-ins. And Soundtoys Decapitator almost always playing a tiny part in it. I also used the Krotos Reformer and Reformer Pro and Transformizer to morph between sounds. And I used the plug-in that everybody loves to hate: Glitch 2, by illformed.org.

You can check out some of the workflow using Glitch 2 here:

And last but not least, let’s not forget every JACK DARK plug-in out there just causing absolute mayhem.

They are kind of hard to come by these days, but you should really try and find them. They’re only 32-bits I think, which is one reason why I haven’t updated beyond Cubase 8 actually.
 

Recordings:

I moved to a new house, which had a nice old piano placed in the living room that I specifically asked the previous owner to not throw out…so I could destroy it

During the creation of DARQ I moved to a new house, which had a nice old piano placed in the living room that I specifically asked the previous owner to not throw out…so I could destroy it. I have to say that this was a very old, out of tune, wrecked piano with a big burn mark on it.

It was most definitely not a well taken care of piano and it would have ended in the junkyard anyway, so I decided to keep it and record it as I smashed it to pieces before we took it to the junkyard.

Here, any big pieces of metal left on the sidewalk get “stolen” by looters ultra fast. So after taking the piano apart, we only had to drive the wooden parts to the junkyard. We left the heavy metal frame on the sidewalk and when we came back it was gone; we didn’t even have to carry it that far.

The smashed piano recordings made a lot of assets that I used in the game, particularly the sound of an actual piano breaking in the Train level. You can check out the three YouTube videos I have about recording the piano smashing right here:

Destroying a piano part I:

Destroying a piano preparing gear:

Destroying a piano part II:

Here you can also see quickly how to solder your own little contact microphone and a couple of other small nifty tricks.

Another recording I captured — apart from running around hospitals, abandoned schools and other places I shouldn’t actually go — was the sound of my kids screaming. The idea for screams started with my good friend and industry colleague Joshua Smith of Escalation / Bethesda Game studios in Dallas, who requested some child screams for use in Elder Scrolls Online. My daughter and I did screams for him and ESO. You can hear the results here:

I also have a super short video of the process here:

At the apartment I used to live in — before the piano — I did a brief demonstration on recording things around the house, without much acoustic treatment, but it still provided lots and lots of material to use in a project like this. You can view that here:

I created small contact mics and b-field microphones that I used around the house. Here is an example of creating soundscapes out of literally nothing:

A b-field microphone can be easily soldered using simple tools and you can create some really nice electric noises with it. Add some effects and volume automation to this and you pretty much have most of the electric sounds used in DARQ.
 

Process on the fly

Many sounds for the game were initially prototyped either on my laptop, in airports, on buses and trains, etc. — some of them in whatever studio I was in at the time or at my apartment in Krakow because I was working at CD PROJEKT RED at the time. I didn’t have any recording facilities or access to much material of my own, so I would quite often rely on pre-recorded material that I would prototype to make basic ideas and put it together. Then I’d render those sounds out either fully or in layers and go home to my actual Cujo Sound studio in Copenhagen and try to recreate those sounds. This spawned a lot of creative work; since I was using sounds that I didn’t do myself it became a massive quest to try and figure out how those sounds were made or how parts of those sounds were made, and why they are the way they are. This led to hours of trying out various combinations of compressors, reverb effects (both plug-ins and outboard gear), limiters, combinations of synths, controlling synths through DC inputs and so on before ending with what eventually made it into the game.

It was quite the learning experience trying to sound like someone else’s work that I had already put together and massively processed through my own plug-ins and setups. I wanted to recreate that sound again with my own recordings and tools.

This is where my electronic music and techno background would come in handy because I wanted most sounds to have that compressed power sound, pushing the sounds to a pretty high limit in terms of both amplitude and compression. Basically, I was thinking of each level as its own little composition where all sounds are to be part of a more overall piece. So the background noises, ambience, the sounds of interactions, and whatever would fit in should all fit together like a song. One of my great musical inspirations once said that certain pop music had tap-friendly rhythms and melodies that could be hummed, and they took out those parts. (The interview was with Autechre, who are big heroes of mine.)

I’d like for DARQ to be considered like that, as though there’s no red-line between any of the levels and sounds — apart from the way they are made and not the way they work in the environment. None of the sounds should have any form of repetitiveness to it and even if they do, they don’t.

Autechre was a great inspiration throughout my entire career really. I don’t like everything they put out, but I like their overall ideas of having said in rare interviews that they liked to work without monitors or strictly with outboard gear or simply only max/MSP — basically creating creative rulesets or creative dogmas very much along the lines that I set when starting on DARQ.

Using Transformizer and Reformer

Sounds can be created in hundreds of different ways. You can probably go somewhere and record sounds like this in nature if you look for the right location, but I like to work with tools and try to make it feel natural even though, oddly enough, it’s anything but natural.

I used Transformizer and Reformer quite a bit during the production on DARQ. Sometimes I even stacked instances of Reformer with other plug-ins to play around with, creating weird noises that could be later loaded into Transformizer again or used as parts in my LEGO library that I could then build other sounds from later.

---

Popular on A Sound Effect right now - article continues below:

---

The individual levels

The apartment

The apartment sounds are probably the most subtle and impersonal of the whole game. There isn’t much going on here, but we did go through quite a few iterations before landing on the ambient sounds that we wanted.

Our first versions had muffled sounds of neighbors, which were from really nice apartment recordings I captured during my holiday in New York City some years back. There were lots of background hums of AC units, backyards, and other typical city soundscapes recorded there and put into this as static noise here and there.

Here is a video explaining the apartment and the footsteps setup of the game:

 

The gears level

There are lots of moving parts on this level and creating certain mechanical sounds from scratch proved very difficult. For some of the sounds, I had to dig into quite specific mechanical libraries to find things I could cut into pieces and use in combination with my own sounds to get the power that I was looking for.

Some relatives of mine have this old grandfather clock with nice moving parts. I was able to get really close up recordings using my Tascam and a good microphone. I recorded at really high sample rates and bit depth, making it possible to pitch down the sounds quite a bit yet keep their actual quality intact. Lots of sounds from the school recordings I mentioned previously were recroded this way, including a wooden ball thrown at a basketball hoop plate in a high reverberation gym; it sounded especially cool when the ball was rolling and hopping over the ground afterwards.

At that schood, we found a big concrete basement with a really long reverberation tail. I kicked this discarded beer can and it made a really nice long tail which I recorded and used for lots of the hissing sounds in the mechanical parts.

The level sound design is explained here in this video:

 

Subway level

This is where my New York vacation trip really comes to life because as much as I was annoying my travel partner by constantly recording stuff, all the waiting in the subway was really worth it, because I recorded almost an hour of trains coming, going, pass-bys, at a distant, close up, and much more.

These are the foundation of all the sounds in this level, and I had a lot of fun making them into what they are. I especially like the sound of the train stopping, which I reversed and chopped by various methods, like tremolo, automation chopping of amplitude before reverberation, and so on.

There was probably an extra compressor somewhere in the chain.

Here is a video about how some of the ambient sounds in this particular level came to life:

The whole level is played through here and explained:

 

City level

This was one of the first levels I worked on after the Opera level and it was quite interesting. As with the Opera level, the City level came before I truly established my production pipeline at the studio, so the method was a little different from what it ended up being and the thought process was also slightly different, even though the same dogmas and rulesets applied.

I reused lots of my LEGO blocks of sounds from the previous levels, like the grandfather clock for example, and gears and moving parts from various sources. The lock puzzle is basically just a padlock, almost 1:1 with the one in the scene.
This level is where we first encounter some enemies. I have a video about that, or at least the idea behind it:

This is a video about creating some of the base material for breathing and attacking.

This one is about how to put all that together.

The whole level playthrough and commentary is here:

 

Opera level

This level changed a bit during development. As with all video game development, things can change during the development depending on how things progress.

As I mentioned before, lots of the sounds for this level were made by moving furniture around at home, throwing a guitar around and making it hit the floor and resonate, which adds small sounds to the room rotation effect when moving the actual stage and the harp located there.

When entering the vertical room puzzle, where everything moves up and down depending on which direction you go, there are a lot of booms and thuds in those sounds, but lots of detail was put into the small debris noises, which were made from dried sand from my kids’ sandbox and various sized pebbles that were dropped on the floor at home. This adds a lot of life to the layers of the moving sounds.

Every one of those moving parts are actually triggering the same event, but they have so many variations and layers that it’s never the same sound. So instead of having an event for the first room, the second room and so on, there is one event for room rotation and that triggers all the so-called multi instruments in Fmod which are combined in another single multi instrument event randomizing the randomized parts.

This level is also where some of my first attempts at the breathing were made. Remember, this level was the first level that was finished for sound and yet I made some breathing sounds for earlier levels. This one is where I had to actually come up with how I was going to create those sounds.

Where that happens you can try for yourself.

A full playthrough commentary level can be seen and heard here:

 

Train level

For this level, I did have some train sounds of my own, but those are from pass-bys and are usually diesel or electric trains recorded around Denmark, between 2008 and 2011 when I was commuting from Copenhagen to Aarhus to study at the Royal Academy of Music there. It was quite a long ride, but I did manage to write almost an entire album of music on those train rides over the years and record lots of train details, like nice transitions between open and closed doors in the bathrooms of the train — both during riding and parked trains.

The details of the suitcases on the floor that slide back and forth are recordings of backpacks and suitcases at my house recorded really fast with my phone, simply to create a very low fidelity recording. I pushed the EQ in the higher ends and created a nasty hiss to make them extra audible in the soundscape.

The sounds of the destroyed piano really came in handy on this level, with lots of compression, reverberation and smashing going on.

There are cool breathing sounds here too, which I recorded myself, just shouting and screaming into the mic. I was also trying to sound like the actual end product, almost making myself vomit during the recording. I’d like to say that it adds a natural flavor to the sound and a very cool feature as it’s “real”, but trust me it wasn’t very nice and I felt stupid sitting there in my studio trying to hold it back.

A commentary playthrough video is here:

 

The key level

We are getting close to the end of the game, so we intensified some of the ambient sounds; they are a little louder in general.

There are some nice puzzles on this level which allowed for lots of odd movement sounds and loops, all put together with some basic move parameters in Fmod. It’s basically a loop playing constantly but manipulated with the parameter indicating the speed of the object.

Specifically, there is a magnet ball puzzle, which is really well done by the developer and puzzle designer.

As mentioned in the commentary video, I went to my Mom’s laboratory she had at the time and was lucky to gain access to hundreds of different kinds of vials, syringes, setups, fume hoods, suction systems, air conditioners, fridges, freezers, ventilation and so on, and a lot of the sounds are formed from all those. In particular, there is a plague doctor NPC that you can meet and have to sneak past. He is fiddling around with some laboratory stuff and it was all captured on a Sony D-100 in a lab, with me, for a couple of hours. Lots of fun there.

Most sounds on this level, like so many others, are variations of sounds, exported into stems so they can be made into multi-instruments in Fmod that randomize between the variations and then a new multi-instrument which plays all these stem randomizers at the same time.

So usually a sound would be a boom with a bunch of variations, some crunch sounds, details and so on and they will each have 3 to 5 variations. This gives us variation in general so even though a sound is similar, it’s not really the same, with enough variations for the player to not notice (hopefully).

A full playthrough commentary video can be heard and seen here:

 

The cave level

The cave is the very final part of the game and by far the most linear. So running straight is what is usually happening, meaning that we trigger lots of things, not on a timeline but as a more continuous soundscape that had to fit together. There are lots of the same types of cracks, booms and rocks dropping from the previous parts of the game, but not exactly the same; usually I would re-use material but re-design them as stems of variations for each level.

So I may use the same crack sound in lots of different end product sounds, but because it’s used with other stems or mixed differently in the design process, it’s not exactly the same sound. It might be made from the same source, but that doesn’t mean it comes through as identical in the end.

There’s a time slow down effect on this level, which is a very simple parameter that changes the pitch of the whole thing, except for the specifically designed sounds for some slow rocks that drop towards the end of the level.

A full playthrough and commentary of the sound can be heard and seen here (Massive spoiler alert that this is the very last part of the game, so if you don’t want the ending spoiled then don’t watch this, or at least not the ending of the video.):

The end of this video includes much of what I have described in this article, but I really consider myself lucky that I was able to sound design this game. It was a really long process, but I didn’t feel overworked. A single person developer of basically the rest of the game can only work as fast as one person, so there was lots of time for prototyping and thinking about what to do next. Additionally, I was running a full time job on the side as well.
 

The last bits

The final parts of the game consist of the typical process of compression, limiting, EQing and some more compression.

Everything here consists mainly of big booms, loud thuds and compression, lots of reverb and very controlled sine wave bass drops. Those have been used in many of the other levels as well, but this is the technique used to create all those boomy sounds and effects throughout the entire game.

The last bits of running also have some nice drones in it. Those are from a sound library I released called Glacier Ambience; it has nothing to do with glaciers really, but I made them all during my stay in Iceland for three years whlie working for CCP Games. Thinking of cold and harsh nature and glaciers was a pretty natural thing during my stay there. You can check out the library here.

The slow motion effects are made to sound like that; it’s not pitched down sounds. The sounds are made the way they sound. I didn’t make sounds at full speed first and later pitch them down through Fmod.

The Tuba man base sounds are actually not mine, though I did manipulate the original recordings. Those sounds are made by a musician and a good friend of the game developer. But the wheelchair that he is using and the later wheelchair occurring in the seemingly same level has a video about their creation here:

As mentioned previously, many of these sounds were made from library material that I have gathered over the years. I went to my Mom’s laboratory; I went to a junkyard, my own kitchen, my kid’s room, and a friend of mine has this really cool sounding oven which squeaks and schreeeeks every time you basically touch anything inside it.
 

Summary

Thank you for reading about DARQ and how I made all the sounds of the game. I really hope that this was inspiring in some way and perhaps you watched some of the videos I made for this.

Working on DARQ was really interesting, not only for the sound design and my own process, but also our development process — working with a small developer, our pipeline, lots of creative sparring between me and the developer and how we wanted this to sound.

It was quite a long process, but as I said before, not hectic or busy. It was just the perfect combination of creative process and speedy processing when needed. The feedback I have received is really positive, which obviously makes me extremely happy.

I learned lots of new processes while working on this, new ways of tweaking Cubase and other tools. , Even though I have been a Cubase user since 1997, there are still new things added to it and I keep finding new ways of getting things out.

All the recordings I made for the project (quite a few that were unfortunately lost after the design process) still led to a massive library of mine which I can use for years to come if I have to. Now more projects are calling and there is DLC coming for DARQ as well, which is free by the way.

 

A big thanks to Bjørn Jacobsen for sharing his creative processes behind the sound of DARQ!

 

Please share this:

 

---