SpringerBriefs in Architectural Design and Technology, Markus Hudert,Asterios Agkathidis,Carlos Medel-Vera, Editör, Springer, Singapore, ss.103-117, 2026
The chapter explores the integration of AI-generated music to enhance immersive historical reconstructions. Informed by the music theory and modal systems of the period and drawing on makam traditions from later eras, it reconstructs the sound of a largely forgotten music heritage. AI-generated soundscapes, developed from historically grounded assumptions are integrated into photogrammetric models and interactive, game-based VR environments. The process involves curating a specialised dataset of makam music captions drawn from historical accounts and oral traditions, expanding it through large language model (LLM) augmentation under expert supervision, integrating contextual information from each caravanserai, generating audio samples using a fine-tuned text-to-music model, and applying acoustic simulations with measured impulse responses to reconstruct the historical soundscape. This study evaluates the impact of these reconstructions on user engagement, immersion, and perceived historical authenticity, demonstrating AI’s potential to create multisensory, historically informed experiences within virtual heritage environments.