Geração de conteúdo procedimental para jogos com base em música

Abstract

Procedural content generation (PCG) has been used to create content for video games since the 1980's. It allows an easier creation of levels and other game content while providing the player with a more refreshing and less boring gameplay, due to the constant game level variety. Although the process of PCG is said to be automatic, it is usually used with an interface where the level designer has to choose and define the creation parameters for each kind of level desired. The purpose of this thesis is to create a new paradigm, where the whole level is created using the existing designed game content and parameters without the need of anyone's intervention. The whole choice of the parameters and settings used would be defined by music. We can extract various kind of information from any piece of music, for which different Musical Information Retrieval(MIR) tools can be used. And then using this information we can map them in various different ways to select our parameters, creating in this way a paradigm where the focus is changed from a low-level parameters input method, to a higher one where these are altered and interpreted from a music track, which in turn would result in different kinds of levels and gameplay by simply using a new piece of music. The results was the mapping from the features extracted from Essentia to the developed PCG methods where positives results can be seen although not always perfect