Open Access

Felix Dennerlein

• A controlled listening experiment investigates how dynamic range compression interacts with distance attenuation functions to influence auditory distance perception in game-like virtual environments. Speech stimuli from a single talker are processed with three compression settings (uncompressed, moderate compression ≈6 dB gain reduction, strong compression ≈12 dB gain reduction) and loudness-normalised to −23 LUFS. For each compression condition, stimuli are rendered at six source distances (2, 4, 8, 12, 16, and 28 m) using two Unity-based spatial audio pipelines with different distance attenuation models (Steam Audio and Wwise). To isolate level- and spectrum-related distance cues, the virtual scene approximates a free field without reverberation or reflections and is presented over headphones with a fixed listener orientation. Participants estimate perceived source distance and may replay a 1 m reference stimulus as an anchor. Output levels are matched between pipelines using a KEMAR head-and-torso simulator to ensure comparable earA controlled listening experiment investigates how dynamic range compression interacts with distance attenuation functions to influence auditory distance perception in game-like virtual environments. Speech stimuli from a single talker are processed with three compression settings (uncompressed, moderate compression ≈6 dB gain reduction, strong compression ≈12 dB gain reduction) and loudness-normalised to −23 LUFS. For each compression condition, stimuli are rendered at six source distances (2, 4, 8, 12, 16, and 28 m) using two Unity-based spatial audio pipelines with different distance attenuation models (Steam Audio and Wwise). To isolate level- and spectrum-related distance cues, the virtual scene approximates a free field without reverberation or reflections and is presented over headphones with a fixed listener orientation. Participants estimate perceived source distance and may replay a 1 m reference stimulus as an anchor. Output levels are matched between pipelines using a KEMAR head-and-torso simulator to ensure comparable ear signals. The analysis will quantify bias, sensitivity, and variability of distance estimates across compression levels, source distances, and rendering pipelines, providing guidance for distance rendering in game audio.…