[_kb]

In many cases, that's intentional and required.

Decoding, distributing, processing, and rendering across all the involved components can take in the order of 100's of ms. HDMI 1.3 and up has had a mechanism for equipment to communicate internal delays so audio remains time-aligned with the rendered image.

Some devices will also have manual overrides for this. If you are experiencing significant drift something is likely borked in the setup.