Diegetic Music - Game Sound Technology and Player Interaction: Concepts and Developments - page 66

• the mathematical music theory-based ap-

proach by Mazzola, Göller, and Müller
(2002).

tions through trained neural networks (Hörnel,
2000; Hörnel & Menzel, 1999). Based on a graph
representation of possible alternative chord pro-
gressions (a Hidden Markov Model derivative
called Cadence Graph ), Stenzel (2005) describes
an approach to variations on the harmonic level.
Beyond varying musical material it is also
possible to generate ever new material. There-
fore, Hiller and Isaacsons (1959) have already
attempted this through the application of random
number generators and Markov chains. This is still
common practice today, for example, for melody
generation (Klinger & Rudolph, 2006). Next,
harmonization and counterpoint can be created for
that melody to achieve a full polyphonic setting
(Ebcioglu, 1992; Schottstaedt, 1989; Verbiest,
Cornelis, & Saeys, 2009). Further approaches to
music composition are described by Löthe (2003),
Taube (2004), and Pozzati (2009). Papadopoulos
& Wiggins (1999) and Pachet and Roy (2001)
give more detailed surveys of algorithmic music
generation techniques.
The nonlinear aspects of diegetic music as they
have been discussed up to now omitted one fact
that comes along with interactive media. Music,
as part of the diegesis, not only influences it but
can also be influenced by it, especially by the
player. Which player is not tempted to click on the
performer and see what happens? In the simplest
case a radio is just switched on and off or a song is
selected on the music box. Interaction with virtual
musicians, by contrast, is more complicated. Two
modes can be distinguished: the destructive and
the constructive mode.
Destructive interaction interferes with the
musician's performance. The player may talk to
him, jostle him, distract his attention from playing
the right notes and from synchronisation with the
ensemble. This may even force the musician to
stop playing. Destructive interaction affects the
musical quality. A simple way to introduce wrong
notes is to change the pitch of some notes by a
certain interval. Of course, not all of them have to
be changed. The number of changes depends on the

Even the expressivity of the performance itself
can be varied. This can derive from the scene
context (the musician is happy, bored, or sad) or
be affected by random deviations (just do it dif-
ferently next time). Systems to adapt performative
expression were developed by Livingstone (2008)
and Berndt and Theisel (2008).
But modifying performative expression is not
the only way to introduce diversity into music. A
further idea is to exploit the potential of sequential
order, that is, to rearrange the sequence of musical
segments. The idea derives from the classic musi-
cal dice games which were originally invented by
Kirnberger (1767) and became popular through
Mozart (1787). The concept can be extended by
so-called One Shot segments that can be inter-
posed occasionally amongst the regular sequence
of musical segments as proposed within several
research prototypes by Tobler (2004) and Berndt,
Hartmann, Röber, & Masuch (2006). These make
the musical progress appear less fixed. Musical
polyphony offers further potential for variance:
Building block music 2 allows various part settings
as not all of them have to play at once. One and
the same composition can sound very different
by changing the instrumentation (Adler, 2002;
Sevsay, 2005) or even the melodic material and
counterpoint (Aav, 2005; Berndt et al., 2006;
Berndt, 2008). Thus, each iteration seems to be
a rearrangement or a variation instead of an exact
repetition.
Generative techniques can expand the musical
variance even more. Imagine a virtual jazz band
that improvises all the time. New music is con-
stantly created without any repetition. This can
be based on a given musical material, a melody
for instance, that is varied. The GenJam system,
a genetic approach (Miranda & Biles, 2007), is a
well known representative. MeloNet and JazzNet
are two systems that create melody ornamenta-