Qsound LLE timing issues with zn.cpp

[superctr]

Appears to be caused by the Z80 getting stuck in the busy loop (check with bp d007,1,r)

Attached .inp file, wav recordings of said .inp being played with the LLE and also ValleyBell's HLE.

sfex2p lle issue.zip

[superctr]

Putting a machine().scheduler().synchronize(); right before here seems to fix the ZN timing issue, and also the other issue with sound corruption in vsav2. dino is still a bit broken, though maybe less so (I recall it crashing before).

[rb6502]

That makes perfect sense. It's standard practice in sound cores to add a sync before returning the status, in order to ensure you aren't reading the audio's state in the past. Except we usually do it with m_stream->update(); which calculates if one audio sample's worth of time has passed and doesn't actually cause a sync if it hasn't. That may not be fine-grained enough in this case though. @cuavas you wanna try this out?