2019/01/17 by Constantinos Papayiannis, Papayiannis, Constantinos, Christine Evers +3
Computer Science · Engineering · Health Professions · Neuroscience · #Acoustic Wave Phenomena Research #Audio and Speech Processing (eess.AS) #FOS: Computer and information sciences #FOS: Electrical engineering #Hearing Loss and Rehabilitation #Music and Audio Processing #Noise Effects and Management #Sound (cs.SD) #electronic engineering #information engineering
paper · pdf · doi:10.48550/arxiv.1901.05852
openalex publication_date 2019/01/17 · openalex created_date 2022/09/14 · openalex updated_date 2026/07/28
The materials of surfaces in a room play an important room in shaping the\nauditory experience within them. Different materials absorb energy at different\nlevels. The level of absorption also varies across frequencies. This paper\ninvestigates how cues from a measured impulse response in the room can be\nexploited by machines to detect the materials present. With this motivation,\nthis paper proposes a method for estimating the probability of presence of 10\nmaterial categories, based on their frequency-dependent absorption\ncharacteristics. The method is based on a CNN-RNN, trained as a multi-task\nclassifier. The network is trained using a priori knowledge about the\nabsorption characteristics of materials from the literature. In the experiments\nshown, the network is tested on over 5,00 impulse responses and 167 materials.\nThe F1 score of the detections was 98%, with an even precision and recall. The\nmethod finds direct applications in architectural acoustics and in creating\nmore parsimonious models for acoustic reflections.\n