Another weird corner case is that COP2 (GTE) instructions actually start executing one instruction early with respect to where the program counter is. As a result when an exception happens the kernel interrupt handler checks to see if the next instruction is COP2 and has to add 4 to the program counter in order to not execute it twice.
Also COP2 instructions were not allowed in branch delay slots (presumably for similar reasons), however some games (from memory Tekken3) do actually do this. I always wondered if it was a sneaky anti-emulation thing :) (because many emulators had issues with this, or needed special handling).
This is nonsense. The main reason behind the demise of dedicated sound cards: motherboard sound chipsets got "good enough". The value add wasn't adding enough value any more because you can get decent sound quality just by using the default sound output provided by your motherboard.
3D sound and other processing got baked into middleware for games because it became trivial to do all of the processing in software - and the processing became more advanced than anything that the sound card vendors were offering (and they didn't move quickly enough anyway).
Pro audio vastly progressed past anything that is possible to provide in fixed silicon. For input, dedicated USB (and ethernet) audio interfaces progressed to the point where it would be ridiculous to provide such functionality on a general "sound card".
It's just evolution - there just isn't a compelling enough niche for a dedicated sound card any more.
Also COP2 instructions were not allowed in branch delay slots (presumably for similar reasons), however some games (from memory Tekken3) do actually do this. I always wondered if it was a sneaky anti-emulation thing :) (because many emulators had issues with this, or needed special handling).