> > > * midiout_win32mci.c (midiout_win32mci_write): Change midiOutLongMsg
> > > to midiOutShortMsg to tune performance.
> >
> > This could be a bad move because I think sci_midi_command passes on some
> > SysEx messages to MCI (i.e. ones starting with 0xF). I hope I'm wrong,
but
> > you checked this throughly anyway, right?
>
> Actually, I did see the System Exclusive note in the midiOutLongMsg docs,
> but wasn't sure if they would affect this or not. Everything appeared to
> play as fine as it did before (tested LSL3), so I assumed it was fine.
>
> If it is not, please let me know. the midiOutPrepareHeader and
> UnprepareHeader were introducing a decent amount of overhead, but I can
> add them back if necessary.

After looking at the code, I'm 99% sure that the sound servers are the only
ones that will have to deal with the SysEx (0xF*) messages, and they don't
pass it on to midiout. So it should be OK.

Alex.




Reply via email to