← all datasets

SO-Dataset

Emerging
1papers using it
2026first seen

The SO-Dataset is a collection of 400,000 First-Order Ambisonics (FOA) spatial audio clips and 2.1 million spatial question answering pairs, used to evaluate spatial audio understanding across various subtasks such as detection, location estimation, and spatial reasoning.

🔎 Find this dataset

Papers using SO-Dataset (1)