Hybrid ISM+BRIR convolution reverberation acoustic model with tuning of absorption coefficients based on single-position measurements

Spatial audio for Virtual and Augmented Reality requires perceptual fidelity and computational efficiency to deliver interactive, immersive experiences across dynamic environments. This paper presents a hybrid reverberation model to render real-time and perceptually convincing binaural room acoustics simulation by combining the Image Source Method (ISM) for early reflections with convolution of a given Binaural Room Impulse Response (BRIR) for late reverberation. The approach uses a single measurement at a fixed room position to derive frequency-dependent wall absorption coefficients by means of an iterative optimization based on a selected acoustic parameter. These coefficients are used for ISM rendering early reflections up to the perceptual mixing time, while the measured BRIR tail preserves the characteristics of the late reverberation, enabling the synthesis of different source-listener configurations. Objective evaluation and subjective tests were performed comparing simulations of the hybrid model against real room measurements. The results highlight the importance of parameter tuning and the early-late transition time for achieving perceptually plausible simulations. The proposed method provides a practical solution for generating spatially varying binaural responses from a limited set of measurements.