МАСКУВАННЯ ВИКОРИСТАННЯ ПРОЦЕДУРНОЇ ГЕНЕРАЦІЇ В ВІДЕОІГРАХ

DOI: 10.31673/2412-4338.2025.016070

Abstract

Abstract: The research aims to formalize a method for effective masking of procedural content generation in video games to ensure natural visual unity of automatically generated elements. The paper presents a formalized model for masking transitions between different types of procedurally generated content, based on a combination of several layers of noise functions with adaptive parameters. The central element of this model is a masking function that combines different types of noise with corresponding weight coefficients.To verify the effectiveness of the developed methods, specialized software was created that allows automating the process of evaluating the quality of transitions between biomes, generating test landscapes, and visualizing transition zones with the possibility of interactive parameter adjustment. Comparative analysis of the effectiveness of different masking approaches revealed that using individual noise functions creates characteristic artifacts, while the proposed combined approach provides more realistic and organic transitions between different types of biomes, effectively hiding the algorithmic origin of procedurally generated content. Optimal parameters were established for four types of noise functions: fractal Brownian noise, Perlin noise, simplex noise, and Worley noise. Combining these functions with appropriate weighting factors provides higher quality masking compared to using single noises. Visual analysis confirmed that the combined approach creates more realistic and organic transitions between different biome types, effectively hiding the algorithmic origin of the content.

Keywords: procedural content generation, noise functions, Brownian noise, Perlin noise, simplex noise, Worley noise, generation masking, graphic objects, game environments.