Listen or Die

Story: John Broomhall

WWISE CAGE MATCH

James Magee: We have several audio systems at work, designed to enable an immersive, reactive and dynamic sound experience, responsive to this emergent gameplay.

We used Audiokinetic’s Wwise sound engine and developed bespoke tools to interface with it. Scripting was done through a powerful in-house game editor developed for the project called CAGE. This gave our sound designers the same tools as the level designers, enabling complex visual scripting and access to real-time data to drive the sound.

One of our most significant audio systems was based around the concept of sound networks, a way of intelligently and automatically sub-dividing the world into different acoustic spaces. Sound networks were primarily a means of ‘cheaply’ calculating high-quality Obstruction/Occlusion, but also allowed us to attach parameters to run other environmental systems such as reverb and ambience.

Our AI systems also made use of sound networks to affect logical sound representations, meaning they ‘heard’ sound in the same way as the player. We made use of this through Stealth and Threat parameters calculated from the AI’s awareness level. For example, as the player makes noises the alien can hear it based upon the surrounding environment.

This real-time data was incredibly useful in driving many RTPCs (real-time parameter controls) within Wwise to emphasise mix elements, eg. bringing the sounds of the main character Amanda and the alien to the fore while ducking or filtering ambient sound to create focus and anticipation in certain moments.

MUSICAL LAYERS

AT: How do you integrate music soundtrack in a world where environmental sounds are so important to the gameplay?

Bullock: We constantly assessed how each sound would make the player feel. That often meant sounds would take on tonal and sometimes musical characteristics.

Early on, we made a conscious decision not to have wall-to-wall music. That’s not to say we didn’t like the music — our composers did a fantastic job! — but we wanted the soundtrack to have contrast and dynamics. You don’t get used to music playing, so it has impact when it does.

Some of the best audio moments are quiet ones without music — you walk into a dark corridor, it’s all quiet except for something moving in the dark. All you can hear is yourself, the walls creaking and some unexplained movement, it’s really tense.

Also, sound can add a sense of realism which can shock people. The sound of the alien ripping apart someone as they scream loudly from the other end of a dark corridor will instantly put you on edge. That moment doesn’t need music — it would feel less real.

Magee: Using parameters such as Stealth and Threat, plus ‘Obstructed_Distance_To_Alien’ gave us tools for a highly dynamic music system. We wrote a custom Wwise plug-in that allowed us to pack three stereo tracks into a six-channel file, which can be mixed at runtime. These tracks were composed ‘vertically’, so when a threat draws closer or becomes aware of the player, the music layering system can dial up the suspense.

Haydn Payne: During encounters with the alien we usually play one of our pieces of suspense music — mostly orchestral but with some synthesised/sound design elements. We fade each of three group layers up and down in real-time according to game events. For instance, if the player is hiding in a locker, the music will sound almost ambient with the alien in the distance, but as it approaches, the intensity of the music rises and the high strings will become loudest as the alien walks right past you.

If the music always responded directly to the alien’s proximity it could harm the game-play by giving away his position too easily, so throughout the game we change how we control the parameter that fades the levels up and down. 

One of the alternatives is the music reflecting the degree to which the player is being stealthy by using code to calculate how close unaware hostile enemies are, increasing the tension level as the player sneaks close to them.

We can also script the music to change in different ways when the player sees particular objects or characters — say playing a sting when the alien walks round a corner into view, and subtly increasing the volume of the music as the player approaches a particular part of the environment. 

A lot of this is controlled by the enemies’ AI state. As the alien charges to attack we usually trigger a short music riser which builds up to the point when the alien reaches the player. We also use the AI state of the hostile humans to control action music, which has three levels of intensity we switch between based on how aware the hostiles are and how much danger the player is in.

Bullock: We’re constantly asking ourselves, what should the player feel and how can we achieve that? For instance, layering subliminal non-literal sounds underneath real-world sounds to give them an emotional connection; like adding alien hisses into steam blasts and doors closing, or alien screeches into fire bursts and metal corridor groans. We did that sort of thing a lot. It helps to keep the player on edge. The player thinks they heard the alien — or did they? 

Sound can be a great tool for misleading the player and at the same time it’s great for doing the opposite, leading the player. Sound can direct the player to a certain area or tell you what’s happening behind a door, like someone screaming in the next room. All these off-screen sounds can tell a story in their own little way and ultimately contribute to the greater narrative.