English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

We propose a mesh-based neural network (MESH2IR) to generate acoustic impulseresponses (IRs) for indoor 3D scenes represented using a mesh. The IRs are usedto create a high-quality sound experience in interactive applications and audioprocessing. Our method can handle input triangular meshes with arbitrarytopologies (2K - 3M triangles). We present a novel training technique to trainMESH2IR using energy decay relief and highlight its benefits. We also show thattraining MESH2IR on IRs preprocessed using our proposed technique significantlyimproves the accuracy of IR generation. We reduce the non-linearity in the meshspace by transforming 3D scene meshes to latent space using a graph convolutionnetwork. Our MESH2IR is more than 200 times faster than a geometric acousticalgorithm on a CPU and can generate more than 10,000 IRs per second on anNVIDIA GeForce RTX 2080 Ti GPU for a given furnished indoor 3D scene. Theacoustic metrics are used to characterize the acoustic environment. We showthat the acoustic metrics of the IRs predicted from our MESH2IR match theground truth with less than 10% error. We also highlight the benefits ofMESH2IR on audio and speech processing applications such as speechdereverberation and speech separation. To the best of our knowledge, ours isthe first neural-network-based approach to predict IRs from a given 3D scenemesh in real-time.

MESH2IR: Neural Acoustic Impulse Response Generator for Complex 3D Scenes