Pages

Wednesday, 23 October 2013

Interactive Music


From the early days of orchestral accompaniments to silent films, music has played an important role in enhancing media. In early cinema, music was played as a method of masking unwanted noise from the street outside and from the projector [1]. As film became more popular however, practitioners in the field soon noticed the impact of including music as a way of enhancing the psychological drama of a film [2] and thus music became a much more integral part of the performance.

In the context of computer games music is a slightly more complex subject. Due to their interactive nature, games present the unique challenge that the actions of the player in the game cannot be exactly anticipated. This creates interesting problems when considering musical accompaniment which, if used in a similar way to film, must reflect the emotional state of the player and the game environment. In a game situation a linear score can quickly become repetitive. If a player is stuck in a certain place in the game for example, the monotonous looping of music will become increasingly maddening and will only serve to infuriate the player as they struggle to progress. 

The answer to this problem is therefore to create a system which allows the music to flow and change with the game dynamics, reflecting the mood of the game in real time. There have been many attempts to create interactive music systems in games and many have been unsuccessful!


This is a good example of an interactive music score failing to react to a change in emotional state. As you can see, the high-energy action music, which is perfectly suited to fighting gorillas does not work so well when all the gorillas are dead and you’re standing still. This demonstrates some of the problems that need to be overcome when considering an interactive music score.

One of the key elements with interactive music is to address the balance of ‘Musical Structure vs Player Autonomy’. The music must be structured enough that it can complement the emotional state of the player but also allow and be able to adapt to the unpredictable nature of the game environment. This presents a constant challenge for the audio development team and especially in recent years have led to many novel applications of technology to address this critical balance.

A great example of modern day interactive music is demonstrated by the highly stylised, side-scroller puzzle game Limbo by Playdead studios [3]. Limbo takes a refreshingly different attitude to in-game music, merging the boundary between the diegetic sounds of the game and the otherworldly textures that form the music. Arnt Jensen, the creator and game director said that he wanted to give prominence to the boy’s Foley sounds, to emphasize silence and subtlety in the ambiences, and to avoid music that would manipulate the emotions of the player.

The music below is taken from the title screen of Limbo.



Immediately the tone of the game becomes apparent. Bleak sonic textures and unidentifiable acousmatic sound combine to create an eerie soundscape which is strangely organic. A subtle drone is layered with what sounds like a combination of waves and bee and traffic or possibly rain, a waterfall, maybe wind. This highlights the beauty of the sound design for this game. All the elements are suggestive of real-world sounds but are expertly manipulated and combined to become one flowing sonic journey that remains strangely ambiguous. The sound designer, Martin Stig Anderson, has a background in ‘electroacoustic’ music and has developed techniques that perfectly complement the artistic style of the game.

This next video shows an example of how the music responds to mirror and emphasise in-game events. At the start of the clip a high pitch drone can be heard, layered with what sounds like birdsong and forest ambience. This continues until the boulder crashes into the ground whereby a different tonality can be observed. Notice how after the sounds of debris from the impact settle, the music adopts a more granular character almost mimicking and continuing the sounds heard from the previous event.



This second gameplay video shows another example of the adaptive, almost narrative nature of the music. Before the encounter with the infamous spider, the music takes the form of what appears to be a wind-like texture. The spider comes into view and unfurls its legs, smashing the ground as the boy approaches. As before, in synchronicity to the impact of the spider’s leg to the ground the music reacts, evolving into much more ominous and menacing tone.



As you can see, this demonstrates an extremely effective method of implementing an interactive music score into a game which perfectly fits the aesthetic and tone the developer is trying to achieve. The music enhances and complements the emotional state of the player, seamlessly adapting to the changing game world without breaking immersion.

Achieving such an immersive example of interactive music was a long process and many paradigms about existing techniques had to be rethought. Usually with game sounds all audio artefacts will be presented in mono however, in an interview with Martin Stig Anderson he explains about how when he put everything into mono he couldn’t engage with the sounds, stating that it was just not immersive enough [4].

Anderson continues to explain that ambiences and textures were broken down into individual ‘grains’. Sound effects such as rain and wind might only consist of a two second sample which was then cut into ten to fifteen different slices and stitched back together at run time. This allowed huge varied textures to be created from very small chicks of audio. This technique also ensures that the probability of hearing any repetition within the music is incredibly low.

It must be stated that the music in Limbo does not directly confront many of the issues presented with interactive music, it merely sidesteps them. None of the music has a notable pulse or rhythm meaning that sections can easily transitioned between without breaking the immersion of the game. Nevertheless it is still a great example of how effective sound design and clever use of technology can be used to create a powerful sense of identity and style within a game.

In an article on Kotaku [5], Stephen Totilo suggests that video game music is neither essential nor necessary for complete functioning of a computer game, implying that the same experience could be had if the music is replaced by a podcast or other such media. Limbo is a shining example of why (properly implemented) music is essential in modern games not only as an accompaniment to the graphic but also as a key game mechanic. The way the music functions in Limbo to highlight certain objects in the landscape or emphasise certain actions is just as integral to the experience as the graphics. The game is presented as a whole, with the graphics and sound so tightly integrated and essential to the feel of the game that they cannot be separated. Simply put, to play Limbo without music would be to play half a game.


References:

1.      Calvalcanti, A. (1985). Sounds in films. In Film Sound:Theory and practive (eds. E. Weis & J. Belton), pp 98-111. New York: Columbia University Press.
2.      Lipscomb, S.D. & Tolchinsky, D.E. (in press). The role of music communication in cinema. In Music Communication (Eds., D. Miell, R. MacDonald, & D. Hargreaves), Oxford, UK: Oxford University Press.
3.      http://limbogame.org/
5.      http://kotaku.com/5730637/the-year-i-gained-the-courage-to-ignore- video-game-music      

Wednesday, 2 October 2013

Non repetitive sound design

One of the biggest challenges faced by the Sound Designer and Audio Programmer when implementing sounds into a game is the issue of sound repetition. In the real world (excluding artificial systems like a fire alarm which convey information) you would not expect to hear the exact same sound twice. For example hearing the exact same dog bark three times in a row would probably be most unsettling and you would start to question whether you were actually in real life or in-fact part of some artificially created world that was experiencing a minor glitch.

Provided we are talking about non-symbolic sounds (which carry meaning or convey information), the same is true for computer games. If a player experiences something that is particularly unnatural or breaks the ‘flow’ 1 of the game then they will pulled from the game environment as their attention is drawn to this event, breaking any immersion built by the game so far 2. Immersion is seen as the Holy Grail of the gaming experience and therefore it is the challenge of the game audio team to overcome these obstacles within the technical limitations posed by the game platform.

With the increase in system resources that have come from improvements in technology, various systems have been developed to deal with the issue of sound repetition in games, two of the most commonly used systems are explored below:

·         Pitch Shifting and volume modulation can be used to increase variation in one-shot sounds and layered textures. When applied subtly and randomly, using modest values, a large range of variation can be achieved with only a few sound files giving savings on both time and memory, which is always a good thing! This technique typically works best on repetitive sounds like footsteps or gunshots but can equally be applied to layers of atmospheric textures such as bird song.

·         Horizontal Concatenation, which involves combining several small sound files to create a longer sound, is frequently used to provide variation in ‘one-shot’ sounds. When this process is randomised, this allows for a huge number of variations from only a few original sounds (the actual number of permutations is '(2x) - 1' where x is the number of sound files used.). A simple concatenation system from UDK is shown below: 



The boxes to the right side of the of the system are the sound files which will be combined to create the overall longer sound. These are combined at random using the 'Random' object and then passed to the 'Concatenator' object where they will be 'stitched' back together in a top down order.

The video below demonstrates an example of non-repetitive sound design in the latest of the ‘Elder Scrolls’ series of games, Skyrim.



As you can hear, each time the bow is fired, a different set of sounds is heard. This is relatively close to what you would expect to happen in real life and thus does not draw attention to the action and break the 'flow' of the game. Obviously I can only speculate about the methods actually used to implement this system but for the purposes of this analysis I will assume that they have used the methods I have covered above. Sonically it appears that the firing mechanic of the bow is split into four distinct stages.








Once these stages are established the various recorded or premade bow sounds can be edited and divided among the four groups. A system much like the basic concatenation system shown above could then be implemented to randomly splice the samples together in the correct order. Pitch and volume modulation could be applied to each of the stages to provide further variation.

References:

  1. Chen, Jenova. “Flow in Games (and Everything Else).” Communications of the ACM 50, no. 4 (April 2007): 31–34. - Great paper on the 'flow' theory in games.
  2. Stevens, Richard, and Dave Raybould. The Game Audio Tutorial: A Practical Guide to Sound and Music for Interactive Games / Richard Stevens, Dave Raybould. Amsterdam ; Boston : Focal Press/Elsevier, c2011., 2011. - Excellent book on all things game sound related
For some further reading on non-repetitive design check out this website: http://designingsound.org/2010/03/audio-implementation-greats-5-ambient-the-hills-are-alive/