• Simon Rostami Mosen
4. Term, Sound and Music Computing (Master Programme)
This thesis presents a hybrid algorithm for efficient dynamic rendering of room acoustics. To find the optimal perceptual trade-off between resolution and computational efficiency, perceptual listening experiments on the significance of different orders of early reflections were conducted. Based on the findings, a hybrid tool was designed and implemented in JUCE, combining dynamically rendered first-order reflections with convoluted late reverberation. A usability evaluation and a heuristic interview, proved the performance of the designed algorithm to be very effective. Moreover, a signal comparison and a perceptual comparison against a measured room impulse response, proved that while the proposed algorithm shares many attributes with the real room, some differences (possibly from the measurement of the room impulse response) are present.
LanguageEnglish
Publication date25 May 2022
Number of pages84
External collaboratorAudio Modeling
CTO Emanuele Parravicini emanuele@audiomodeling.com
Place of Internship
ID: 471369514