Acoustic Feature Extraction and Analysis for Dynamic Difficulty Adjustment Based on Players' Voice Reactions

Abstract

In recent years, there has been a great deal of studies on how to design games that follow the emotions of the players in various games. In this study, we focused on the voice that players make during gameplay. During gameplay, players may utter speech at will. We use the acoustic features of these speech sounds as input information to infer emotions. We developed a DAAF system that dynamically controls the difficulty level of the game by taking the emotions into account. Furthermore, we created a demo game using the DAAF system and conducted experiments. As a result, we were able to adjust the difficulty level for each player.