Just a quick addendum or example or derivate or something to this. How is audio used in Left 4 Dead?

The video game Left 4 Dead is begging to be mentioned when one talks about creative ways of using diegetic and extra-diegetic sound in games. The survivor side of the game plays with high reliance to soundscapes, sound cues and music. So much in fact that you could say beyond the very basics of survival, everything in the game boils down to listening to what’s happening in the game.

L4D uses both character level and player level audio to support the game play. The very basic soundscape of L4D is made of player level music (well, more like long synth chords, most prominent at start of levels, to open up the level) and character level ambiance sounds of the environment like crows and sounds of thunder. This creates a very strong base mood for the game.

The second way sound happens is through the sounds of the player characters. First of all, they produce the sounds that you’d expect – when you shoot a shotgun, you hear a shotgun sound. When you move around muck, you hear squishy footstep sounds. But beyond these, the characters comment on the environment actively, even without the player telling them to. When they get attacked, they scream for help. When they notice one of them is low on health, they tell the other character to heal up, etc. This is fully happening on the character level. The player can intrude on the character level and command his character to say some of the pre-scripted things, for example to call out that there is a boomer ahead, around the corner even if the character is unaware of it. There is a player level audio bit to some of the character actions as well – getting attacked by a special infected will cause a dramatic song (again, mostly synth chords) to boom out from the loudspeakers. Also, worth mentioning that the characters’ dialogue sometimes dips into the metadiegetic level of story-within-a-story when they reveal things of their past or when their comments tell the story of the infection, as well as telling the player what the character-level story is (in the “We have to follow these tracks to an abandoned military base” style). (thanks Kham for pointing that last one out)

One of the ways people grief when playing Left 4 Dead is to make their character repeat some stupid line of their repertoire like “PILLS HERE” every second or so, causing the other players lose their ability to keep the character level separate from the player level, falling into disbelief of the whole character level of the game in the progress. Not to mention it’s annoying as hell when someone keeps shouting “PILLS HERE” in your ear.

The third way, of course, is the other actors of the game besides the player characters. The helicopter flies above you and use a megaphone to tell you to head to Mercy Hospital. A gas station explodes with a thundering roar. Infected people whimper and moan. And then there are the special infected. Each of them having a distinct sound following them. The smoker coughs, the boomer gurgles, the hunter snarls and then screams when he’s moving. The witch cries and sobs, and the tank roars. And when you hear the multitude of gnarls a horde of normal infected generate, you know there’s a lot of nasty heading your way. These are all character level sounds. When you turn your head in the first person mode, you notice the sounds move left to right and right to left, allowing you to pinpoint a boomer hiding behind the corner a long time before you see it, or hear the witch getting closer as her sobs grow louder. And spotting the specials before they are upon you is really the only way to survive on the higher difficulty levels, when things get tough in the game. On top of those, the witches, tanks and hordes also have a player-level musical theme, that emphasizes the mood of the situation – with the witch it’s a surreal melody that gets piano hits when the witch is almost agitated and turns into a panicing cacophony when she’s out to get you. With the horde, you get a couple of creepy chords a moment before the screeching starts and the masses start flowing. And with the tank, it’s a very doom-inspiring track that makes the player scared enough when the big bad things is coming.

There is a fourth level of audio in the game, if you happen to get into a team that is doing it. The game supports voice communication between the players via microphones. As any conversation between real people, this sometimes is just silly chatter to lighten up the mood. Sometimes it’s completely unrelated to the game, and sometimes it’s a repetition or extrapolation of something the characters could have communicated on their level. For example, a player might say “a hunter got me” to the microphone, even if the audio cues of the character level in the game could have given the same information to the other players (the characters shouting “A hunter got Francis” and Francis screaming “Get it off me”). Thus, there is a player level communication going on about the game as well as the communication that happens via the detour of the character level. And then there is singing. There always needs to be singing to the microphone. I have no idea why.

Oh Boy. Talk about gauntlets getting thrown. So, sound-effects, soundscapes. Music versus noise. Carefully arranged notes in a constant form with pattern, lyric, meaning, sometimes good, sometimes plain awful. Themes, especially the ones that repeat, grow and mutate. And then just silence. Followed by a soft echo of footsteps. A noise here, another there. Sound of the ship’s bulkhead bending in the storm. And as you step out, the roars of wind, rain and lightning hit you in the face with the full force of nature. You hear it roar in the distance, and know it will soon be upon you.

There are soundtracks of wonderful music that guide our thoughts and there are sounds that do that as well, forming a soundscape of the environment we’re in. The big difference between a soundtrack and a soundscape when talking of things like movies or RPGs is that the soundscape, the world of sound effects, is something that exists on the diegetic level instead of the extradiegetic world of the typical soundtrack. These fancy words mean that you can assume that the characters of the story are hearing the soundscape, whereas usually a soundtrack is there only for the audience of the story, not audible to the characters inside the story (exceptions exist, of course).

Now, a well-designed soundscape is something that’s very common (one could say, a necessity) in television and movies, somewhat common in video games but pretty rare in RPGs. At least any I’ve participated in. The two things that have come in the way, to be honest, are the amount of preparation of using sound effects compared to music, and the technology that has been evading the typical game master.

Building a proper soundscape for a roleplaying session isn’t as easy as one might think. In movies and television, you don’t need to think of the lasting aspect of your sounds. The scenes will linger in environments for a few moments so you can think very nicely what exact sounds you need and time them to the millisecond. When you’re playing games like Half-Life or Left 4 Dead, the sounds around you get generated by the computer in response to your actions, the actions and presence of other entities, and randomized from a pre-generated pool of environmental sounds to create the needed effect.

In tabletop RPGs, neither approach is fully usable. Compared to TV, there is no guarantees how long a scene lasts, as all description of action is done as speech and there is a factor of the actors having a say to what happens. You might think a warehouse scene lasts for 5 minutes tops, but your players might spend 10 minutes with their characters at the warehouse, or they might take an hour. Having the same 10 minute planned soundscape loop six times gets annoying really soon, especially if the selected sounds are there to provoke a response. They not only lose their effect, but they will turn it against you as the immersion gets broken – suddenly (gasp) the same creepy footsteps are heard for the third time. As they are present on the character level, they will get ridiculed on the character level. Or you will have to explain that they are not part of the diegetic level anymore.

Computer games give a bit better starting point – a video game’s sound director can’t go about thinking beforehand that it takes 5 minutes, 32 seconds for a player to complete the level. He has to build a generic sound base that will give the player a feel of the environment, without invoking specific action. He avoids the repetition by adding some computerized randomization to the base sound pattern and of course avoiding sounds that would require a response from the player on this base-track. Then he moves forward, having something on his side that a standard GM doesn’t – he can combine the sounds with the triggers from specific events beforehand, so that they play when someone fires a shotgun or walks on a metal floor or topples a pile of cans by accident. Or even when the player is approaching a certain situation or location on the map. Automated responses to the actors, and to be honest, the most realistic soundscapes stem from that.

For a RPG soundscape, like for a soundtrack, you’d pretty much need that base that you can play while there is nothing interesting happening. As a soundscape can be assumed to be heard by the characters as well as the players, it requires more attention than a soundtrack. This base-scape will be playing a lot in a campaign, so it would have to be extremely long to avoid the feeling of repetition. In fact, you’ll need a few extremely long, different bases when you’re building a campaign – there will be various situations that need a somewhat distinctive feel attached to them. Besides those you’ll need other sounds. Sadly, what you can’t really do, is to have sounds for the actions that the players have their characters do – you can’t have a sound-effect for every one of their actions as you have no clue what those actions might be. And even if you did, you’d be always firing the sounds of only after the action has happened on the diegetic level, so there will be unnecessary redundancy that doesn’t especially benefit anyone. And in my opinion, most of the sounds made by other actors should be left to imagination and description as well “you hear the approaching footsteps of the scientists returning from their break”, “the guard accidentally drops his soda can on the floor and curses” as you will need to provide the players with information like where the sound came from and in most cases who was making it. On the other hand, using audio samples to play when something breaks the standard ambiance, yet wouldn’t benefit much from emphasizing on it on the level of the story, would work. A strange sound, a scream, a roar, the sound of thunder approaching that would gradually turn into the base sound once the thunderstorm is upon you. Things that predate actual action or are there just as sounds, nothing more. Yet. Echos of the future.

On practical side, for campaign play, I’d probably start building the base soundscape myself from scratch instead of using something fully made beforehand by someone else. Even if getting quality sounds for free can get tedious, it becomes much more personal and the feel is less hectic if you take time to manipulate the material to fit your needs. There are good services like YLE’s Tehosto around that can get you started, but the samples you get from one source are usually too short to be used on their own, so you need to look around the internet to find enough sound bits. And even when you have enough material to fill a nice long soundscape, you’ll notice that when you start mixing from sources, the difference in quality and style can sometimes be audible. And you will end up ditching a good portion of the sounds you have, eventually needing a lot more sounds. If you want to be really personal, you can record a sound from some place by yourself. Getting a good enough microphone to be able to record the ambiance of an environment on your own might be a very costly task, unless you have friends in high places who can borrow or rent you one. But recording an hour-long thunderstorm will really pay off if you use it on the background in a game – it will sound a lot less cheesy than something you stole from a sound effect box. Once you have a large enough library of sounds to play on the base, you compile them together with some sort of a tracker program, I suppose. I haven’t done that stuff since the 90’s, to there might be better ways to do it these days.

For the specific sounds that get played, you’ll need some form of a mixing software that allows you to play multiple tracks at the same time. – a somewhat affordable solution is Virtual DJ Studio, that offers a very nice, fully working trial version for you to toy around with while you wonder if you want to use it. Then just line up the effects you want to play next to your base track and slide them in when it is their time. And there you go, a soundscape. I can imagine it working very well in horror scenarios, where the players are already on the edge of their seats because of the story.

As a counterpoint to all of this, I must admit that I love using a non-story level soundtrack on the background of RPG sessions – partly because I takes a part of the pressure off from me if I try to get the game going toward a certain mood, and mostly because I love toying between the story-level and the storytelling-level of the game. You can create a lot more contrasts with music than you can do with stuff that exists on all levels of play. Using music on the other hand allows you to stop using music, which on its own can be very very effective (as Buffy the Vampire Slayer has taught us). Doing a full campaign where the ambience was fully created with sound-effects is something I probably would never do, but using some ideas from this post might eventually creep into the games.