FOA Tokenizer: Learning Discrete Representations of Spatial Audio with Multichannel VQ-GAN | Microsoft Research | Podwise