Dance Dance Gradation: A Generation of Fine-Tuned Dance Charts

Abstract

This paper proposes a system to automatically generate dance charts with fine-tuned difficulty levels: Dance Dance Gradation (DDG). The system learns the relationships between difficult and easy charts based on the deep neural network using a dataset of dance charts with different difficulty levels as the training data. The difficulty chart automatically would be adapted to easier charts through the learned model. As mixing multiple difficulty levels for the training data, the generated charts should have each characteristic of difficulty level. The user can obtain the charts with intermediate difficulty level between two different levels. Through the objective evaluation and the discussions for the output results, it was suggested that the proposed system generated the charts with each characteristic of the difficulty level in the training dataset.

Curve-Based Level Generation for a Full-Body VR Rhythm Game

“…Level generation strategies for two-dimensional platformers often generate levels based upon specific elements of gameplay, such as the rhythm of a player's movement through a level (Smith et al 2009) or play styles (Summerville et al 2016). Approaches to rhythm game level generation have leveraged techniques including the use of neural networks (Donahue, Lipton, and McAuley 2017;Tsujino and Yamanishi 2018) and selection from a set of prerecorded moves (Martin et al 2019). These approaches typically limit designer input to a "difficulty" parameter.…”

Deep Learning for Procedural Content Generation

“…Here the player is automatically moved along the level, and has to carry out certain actions in time with the music, as prompted by level features. Some interesting work has been done on learning to create such music game levels from existing music (e.g., [21,139]).…”

“…Tsujino and Yamanishi [139] represented rhythmbased video game levels by charts and implemented Dance Dance Gradation (DDG), a system with LSTMs trained on levels of different degrees of difficulty to generate new levels. DDG can tune the difficulty degree of generated charts by changing the fractions of easy or hard charts used to compose the training dataset [139].…”

“…Tsujino and Yamanishi [139] represented rhythmbased video game levels by charts and implemented Dance Dance Gradation (DDG), a system with LSTMs trained on levels of different degrees of difficulty to generate new levels. DDG can tune the difficulty degree of generated charts by changing the fractions of easy or hard charts used to compose the training dataset [139]. [105] to generate levels of rhythm games, represented by actions and corresponding timing, of different difficulties, trained on the beatmaps collected from OSU!, a famous rhythm game.…”

Ordinal Regression for Difficulty Estimation of StepMania Levels

“…StepMania-level creators and players. So far, only level generation [Donahue et al, 2017;Tsujino and Yamanishi, 2018], and difficulty prediction [Tsujino et al, 2019;Caronongan and Marcos, 2021] have been applied to this data. However, other tasks or subtasks of the former utilizing this data may also be interesting.…”