AI & ML
Reproduction of Simulated Acoustics Scenes Using Acoustics Room Transformer for speech intelligibility tests in Clinics
arXiv eess.AS0 views
arXiv:2611.18103v1 Announce Type: new Abstract: An Acoustic Room Transformation (ART) method is proposed for clinical applications to render simulated acoustic environments using six loudspeakers in a reverberant room. In typical clinical settings, testing rooms are not acoustically treated and only a limited number of loudspeakers are available. The ART approach enables control of the playback-room reverberation, allowing speech intelligibility tests to be performed under realistic and perceptually compensated conditions. The method focuses on the two most common Speech Reception Threshold (SRT) configurations: S0N90 and S0N90. One front and one side loudspeaker reproduce the direct sound components, while the remaining four loudspeakers (two in front and two behind the listener) generate the reverberant sound field. The direct sound is optimized in terms of energy, and the reverberant field is adjusted based on both energy and interaural coherence. An additional direct-sound gain is applied to match the clarity index (C50) of the reproduced signals to that of the reference target. Results from speech intelligibility tests demonstrate that, with ART processing, the reproduced scenarios achieve SRT values very close to those of the reference signal in both S0N0 and S0N90 conditions and therefore allow for comparable speech test results independent of the particular test room acoustics.
Source:https://arxiv.org/abs/2611.18103PreviousAnomalous Sound Detection Using Separation Features and Paired-Mixture Representation LearningNextMultiple spontaneous otoacoustic emissions as a single standing wave unstable to modulation
Related posts
Your Manager's Manager Has Booked a Skip Level Meeting With You
0dev.to

How to Make a Character Reference Sheet with AI (Prompts Included)
0dev.to Design
How Virtual Assistants Help Businesses Manage Their Work More Efficiently
0dev.to Productivity
DiffusionSSD: Achieving Extreme Visual Compression with Dual-stage Diffusion and Cross-level Near-Data Processing
0arXiv eess.IV
D2ToF: One-Step Diffusion for Depth Super-Resolution from Lightweight Single-Photon LiDAR Waveforms
0arXiv eess.IV
Few-Shot Open-Set Hyperspectral Pixel-Wise Classification with Spatial Aware Label Propagation
0arXiv eess.IV