Just FYI 30ms from hardware sources is terrible, I can achieve under 2.5ms round trip with Dante at 128 buffer settings and the older MADI protocol achieves similar levels.