Flow-HOA improves Ambisonics encoding with flow matching

Arxiv CS.SD (Sound/Audio ML)

Proposes a generative framework that jointly optimizes Ambisonics encoding from sparse microphone arrays. Uses flow matching to improve spatial audio quality for immersive communication and XR.

Arxiv CS.SD (Sound/Audio ML)·Yuhuan YouYufan QianTianshu QuBin WangXueyang Lv

7 days ago