You are here

Scoring I Survived A Zombie Apocalypse

Andrew Barnabas & Paul Arnold
Published July 2015

Game Theory

How do you write music for a TV show you haven’t seen yet? It helps if you can draw on years of experience composing for video games...

Dear reader, please let me introduce myself. My name is Andrew Barnabas and I’m one half of composing duo Andrew Barnabas and Paul Arnold. Some folks know us from our earlier incarnation as Bob & Barn.

I started writing music for computer games back in 1990 on the Commodore Amiga A500, with a game called SWIV. I was 17 at the time, studying for my ‘A’ levels and spending all of my free time writing music under the auspicious moniker NightShade of Crusaders for a Norwegian Amiga demo scene group. Many games composers were still using MIDI at this time, but the ‘demo scene composers’, as we were known, were using ‘trackers’: the Soundtracker format, invented by Karsten Obarski, was a four–channel, eight–bit sample sequencer using a piano roll to enter note and parameter data, and gave us a bit of an advantage in the realism front. I’d already begun to do the rounds at trade shows meeting developers, so when Sales Curve rang in late 1990 and asked me to come in and play some music, they gave me my first gig on the spot!

Andrew Barnabas (left) and Paul Arnold in the control room during sessions with the City Of Prague Philharmonic.Andrew Barnabas (left) and Paul Arnold in the control room during sessions with the City Of Prague Philharmonic.I carried on writing scores for games whilst at university and moved to Cambridge in 1995 to become the first in–house composer for a developer called Millennium Interactive, largely because they’d built the best recording studio I’d ever seen — they had a 32–track Pro Tools rig in 1995! I met Paul ‘Bob’ Arnold in 1996 when he was employed as the second in–house composer. Sony bought the studio in 1997, and between us we ran the sound and music department up until 2001, when we left and started an independent company, called Bob & Barn Ltd after our amusing nicknames.

We focused on being games composers for most of the noughties, but often we’d come into contact with composers who suggested we look into scoring traditional media. At the same time, kids who’d grown up listening to our game scores in the ’90s got in touch in the late ’00s and asked whether we’d be interested in scoring their TV and film projects. Surprised? Definitely! Flattered, absolutely.

Until a few years ago, we actually found it to be a stigma to mention games to those in film and TV who weren’t familiar with the medium. However, Bob attended a networking event for TV folks where he met a successful producer who loved games, and knew a lot about scores that we’d written. He got us in as the curveball option to pitch on a TV show called The King Is Dead, and we got the gig. We stayed in touch, and by last Summer he was a senior exec at independent TV production company Tiger Aspect. Bob and I trundled down to London and sat in a meeting room as both he and another exec told us all about an idea they were developing for BBC3...

Made on a big budget, I Survived A Zombie Apocalypse was to be a unique blend of two very different genres: horror and reality TV. Ordinary people would try to survive in a post–apocalyptic world plagued by the undead.

Back To Front

There’s a certain irony in the way composers are chosen in a pitching environment which is, in many ways, backwards. In order to win the pitch to score a TV show, you’re asked to provide demos of what the title tune would sound like. Considering that the titles have to encapsulate the entire ethos of a show in a small amount of music, they’re the hardest thing to get right. When working in house, we’d invariably write the title tune last. This was partly due to the way game development worked, but was ideal for us as composers. Whilst composing and developing the score, you begin to establish themes, instrumentation, motifs and so forth, and you get to know the project inside and out — so if you can write the title theme last, you’re already armed with all of this knowledge and the tune pretty much writes itself.

In this case, there was absolutely nothing to go on, and what made things even harder was the fact that there had never been a show like this made before. So, we did our homework. We listened to a lot of reality and horror music, and documented our thoughts, trying to identify common factors. Research showed that reality–show music is generally contemporary, upbeat, propulsive, light and occasionally a little fluffy. Horror music, on the other hand, is mostly dark, serious, atonal and textural. Meaning that the ideal title track for this particular show would be light, contemporary and a little fluffy, yet also dark, atonal and serious. Hmmm. Add to that a request from the producers that it needs to appeal to a younger audience — though, the show being fairly gory, it’s still 15+.

The most important thing for us was to identify the tone of the show. Was it going to be tongue–in–cheek, or were we taking this seriously? Music has that wonderful ability to convey the mood of a show very quickly, so it was imperative that this first impression hit the nail on the head. With this in mind, we pitched five unique title -theme demos, and Seb, the show’s executive producer, gave detailed feedback as to what he thought worked — and, just as importantly, what didn’t. This all paid off, since he then offered us the gig. Hooray!

The tone we were guided towards was serious; any humour would come from the situation itself, not from music. We were then shown a script telling the story of how the outbreak happened over the title sequence, using faux news footage they were going to film. There was a lot of exposition, meaning that a voiceover would run over the whole sequence and the music thus couldn’t be too thematic (when we read the first draft, it ran to over five minutes, so we knew there’d also be a lot of editing to be done!).

Under Score

We threw all of these requirements and feedback into a musical pot and started to devise our approach. The main hook came from an orchestral string motif, giving us the serious angle. We knew from the script that the reason for the outbreak was a little sci–fi — new 5G on mobile phones — so we got a bit clever and sampled the interference patterns that phones used to cause near speakers, made it into a rhythmic track and subsequently into a piano motif (piano is used frequently in horror music, so that ticked that box too!). To propel the narrative throughout the sequence we gently upped the tempo, which increased momentum. With the target audience in mind, we chose to use glitchy sounds for rhythm. The hardest challenge was to keep it simple and ride the narrative, but still give it momentum, with no more than four main musical elements at any point, to leave enough space for heavy sound design and a voiceover running throughout. (Fortunately, the title sequence wasn’t finalised until late in the day, so we had the opportunity to tweak the music to fit.)

We have a long relationship with the City Of Prague Philharmonic going back to 2002, so we drafted in Nic Raine to orchestrate and conduct, James Fitzpatrick to produce, Jan Holzner to engineer and Gareth Williams as music editor and mixer. Once we’d completed the title track, we got into producing the music for the show itself.The City Of Prague Philharmonic recording their contributions to the soundtrack.The City Of Prague Philharmonic recording their contributions to the soundtrack.

Sharp Shooting

This project was unusual for several reasons, but above all, for the schedule. The production team needed all of the music delivered a week after filming finished, to coincide with the start of the edit. We knew immediately that they wouldn’t have time to get any footage to us beforehand, as there was too much footage to sift through (the data wrangler told me they’d shot 36,000 hours of footage across 40 cameras, many of which running 24/7!). They were on such a tight timeframe for delivering edits to the BBC that they needed to assign a different editor to each episode, and in order to hit the ground running, they needed music right from the start. This meant that for the first time in our careers we had to produce all of the music blind without seeing a thing.

So, we needed to think laterally. First off, we asked whether we could attend the shoot. This bit isn’t unusual for us. Ever since we got invited on set a few years ago we made a pact that we’d make the effort to come and see the action unfold in front of us whenever we got the chance — and I highly recommend it. If you get the chance, do it! Not only do you learn an enormous amount about the process, but the buzz of being on set is great. As composers, most of our work comes to us in post, when the shoot has already happened. Being on set gives you a chance to chat with the actors and crew. Bob and I came off a film set once and found we’d both written themes for the characters in our heads just from chatting to the actors between takes! It helps connect you with the project in a tangible way.

Bob couldn’t join me this time, as he’d just become a father for the first time, so I went up on my own for five days. The first thing that struck me was the sheer scale of the show. The producers had found an abandoned shopping mall between Edinburgh and Glasgow which had opened nine years ago, shut less than a year later and had been slowly reclaimed by nature. Unlike normal British shopping malls where everything’s enclosed, this was more American in style — long, open boulevards with boutique shops both sides — and even a bandstand to boot! It was pretty much tailor–made for a zombie apocalypse. There were around 60 full–time crew, and most amusingly, a coachload of 50–75 zombies would turn up every day. There was even a sign reading ‘Do not park here. Zombie coach parking only’!

The Gaming Approach

I was set up in the gallery, which served as ‘mission control’.With no chance of viewing edited footage before the music had to be completed, Andrew chose instead to watch the action live in the gallery and make extensive notes.With no chance of viewing edited footage before the music had to be completed, Andrew chose instead to watch the action live in the gallery and make extensive notes. In front of me were three huge screens, each showing the views from four cameras. During each mission I would frantically take copious notes in real time and email them to Bob. During missions where some of the survivors were in jeopardy we’d all be huddled around the screens as it got more and more tense, on occasion shouting and swearing at the screens. We all really got into it.

I’d been sat there for days and one day an assistant suggested I take a break and go to the food tent for dinner. Being my first time, and not realising there was crew catering, I instead sat in zombie catering. Watching a large fully made–up, blinded zombie with his eyes completely covered up attempting to eat a peach cheesecake with a wooden fork is an image I’ll never forget!

The eventual title theme, in manuscript (above) and Logic arrange page (below).The eventual title theme, in manuscript (above) and Logic arrange page (below).Game TheoryGame TheoryAs the show was unscripted, even seeing the action live only gave us a framework, as we didn’t know how it would be edited. A chase sequence through a corridor filled with zombies that lasted five minutes in real life might take only 30 seconds of screen time. Watching the action gave me a guideline as to what key emotions the music needed to convey during missions, since I’d documented how I felt when they happened, but the overall look was still unknown. The producers were in the same boat as us, so ultimately, we took it upon ourselves to come up with our own brief in the week or so we had remaining after I flew back to get it all done.

Our solution was to turn to the methodologies we’d been using for years when delivering interactive scores for games, since the criteria were surprisingly similar. A game composer has no idea how long the player will take to complete a given task, be it tentatively creeping around a forest looking for something or running away from a big bad boss. When you come across that boss, an interactive score can be governed by instructions as simple as ‘start boss fight music’ and ‘stop playing boss fight music when boss is dead’, but it can be much more sophisticated: ‘play boss fight start jingle, then seamlessly crossfade to boss fight intensity 1’, ‘play random musical stabs as overdubs to punctuate action’, ‘when boss health is down to 50 percent, crossfade to boss fight intensity 2 + more musical stabs’, ‘if this lasts more than 32 bars, crossfade to boss fight intensity 3 which is quicker and modulated a minor third up’, and so on. This we’ve done many times before.

Adaptable Cues

To give the editors as much help as possible, we had to give enough control over the music to enable them to elongate or shorten the cues themselves, since we knew that they wouldn’t have time to ask us to rework a cue to add four seconds here, lose two there, and so on. Not only that, but we also wanted to enable them to layer stems of the music together to up the intensity should a scene develop from exploration leading to suspense and chase. And to add a bit of colour and variety, we gave them overdubs that they could sprinkle over the top. In sequencer timeline terms, we had to think horizontally as well as vertically.

We made a list of the main areas we’d want to underscore. These were:

  • Tension bed: drones to slowly add tension to a scene where there might be danger around the corner.
  • Rhythmic bed: a subtle heartbeat rhythm that ups the tension to the next intensity, indicating that zombies are definitely close by.
  • Tension swell: a short 10–20 second overdub that ramps up the tension and fades. These add variety and could even be used as musical false alarms.
  • Low action: a timer of sorts, adding a ticking clock sound to a scene for that Mission Impossible effect.
  • High action: full–on chase music for times where the survivors are in mortal peril.

These all had gentle endings but were, in essence, loopable. We composed 10 beds, swells and action cues. We tested as many combinations as we could think of, to ensure that each tension bed could be layered with each rhythmic bed. The swells were written in related keys so each would fit and the low and high action cues were written at tempos either half or double that of the rhythmic bed, so again, they could be layered and still sound natural.

This was all well and good if the editors wanted to quickly jump in with a musical bang when things kicked off on screen, but as with games, there are times where you want sections to come in and out more musically and not jar. To that end we also produced ‘risers’ and ‘shock stings’. The riser would act as a crescendo which could crossfade smoothly into a loopable bed. The shock stings could be used for a sudden shock moment on screen or as a way to end a looping rhythm section with a bang; it would sound natural but actually be completely fake!

Whilst on set, I noticed that they’d hired a camera drone for establishing shots. So we produced 10 establishing stings for those beautiful overhead panning shots.

Planning music for scenes where the survivors were in base was more difficult (the food hall in the shopping mall had been turned into a makeshift camp, complete with sleeping area, sofas and cooking area, a bit like the Big Brother house). There were moments of comedy, sadness, a blossoming romance, preparing for missions — but invariably those sequences would come together in the edit, so we produced 22 additional cues to cover how we imagined they might be depicted, if they decided to use them in the show at all, bearing in mind there was a lot more time spent on base than on missions.

In total, we produced 99 underscore cues and a title track, with a total running time of 52 minutes. These broke down as follows:

  • Seven base cues (comms room, mood music, preparation for missions)
  • Seven comedy stings
  • Five romantic beds
  • 3 sadness beds
  • 10 establishing stings
  • Five Low-action loops
  • Five high-action loops
  • 11 rhythmic bed loops
  • 13 risers (crescendos)
  • 10 shock stings (decrescendos)
  • 10 tension bed loops
  • 10 tension swells
  • Three success stings

This gave them the variety they needed, thanks to the number of different permutations that were possible. To make sure it all worked, we tried lots and lots of different combinations: riser #6 into tension bed #4, followed by rhythmic bed #7, peppered with three tension swells, and ending with shock sting #8, and so on (if you want to hear what it sounded like, download a 3MB MP3 file from www.bobandbarn.com/isaza.mp3). We then put together a detailed document for the editors suggesting how this musical toolkit could be used, thus putting control in their hands. Game Theory

Bob had already been working on cues whilst I was on set, and seven days after I returned, we delivered all 99 cues, on time and on budget. Phew! By that point we were very much in the zone and could’ve produced many more. However, the mandate to appeal to a younger audience meant that they also had access to licensed contemporary tracks to lift the mood, and this helped lighten the load.

Our final task in January was to score hilarious spoof infomercials they’d made for the show, mirroring those Open University programmes from the ’70s. We’ve got an example in the news section of our web site, along with the title sequence that is in the TV section.

Doing Yourself Out Of a Job

So, what conclusions did we draw from the experience? Giving editors control over a virtual interactive toolkit of music made their job more involved, but also made it quicker, since they weren’t waiting around for composers to make changes. It was more work for them than having a composer create a bespoke score for every scene, but it meant that they controlled how each scene played out and could experiment with different edits and change the music in an instant. This system isn’t revolutionary — we’ve been using it in games for years — but applying this methodology to a bespoke library of music cues for a TV programme is unusual, if not unique.

Of course the best results are still obtained by writing bespoke cues and tailoring them to picture, but if that option isn’t available, we’d do it again in a heartbeat. Applying knowledge acquired in games to a TV show — how times are changing!

Someone asked me the other day whether I Survived A Zombie Apocalypse had been commissioned for a second series. I have no idea, but thought about it afterwards: if they do, we’ve kind of shot ourselves in the foot. Since we’ve already given them hours of musical permutations, they won’t need much in the way of new material. Oops! Maybe they should set season two in space or underwater, so that we could easily justify a whole new musical toolkit...

www.bobandbarn.com