[sdiy] Seq stop mid song, best silencing practice?

[grant at musictechnologiesgroup.com]

> That's funny, because it would seem to indicate that you somehow crash 
> the JX-3P.  It's not supposed to react to any system realtime 
> messages.
> 
> http://myweb.tiscali.co.uk/feline1973/boards/MIDI/JX3P.pdf

I guess in this case the manual might be wrong.  I didn't actually try
it yet, 
but I saw this in the source code:

RxMIDI_ISR:   mov    A,#0xFF           ;074h,0FFh
              cjne   A,SBUF,L0671      ;0B5h,099h,008h  MIDI reset? 
              acall  MidiReset         ;0D1h,0C9h       Yes   (06C9) 
              mov    PSW,#0            ;075h,0D0h,000h
              ljmp   Reset             ;002h,000h,000h        (0000)
L0671:        mov    A,SBUF            ;0E5h,099h

That should be a cold boot, (MidiReset is actually an RTI OMG). I will
have to try it. [And they read SBUF twice... that is such a bad idea].

GB

> If you know at least the channel and the possible list of notes to stop, 
> this would be preferrable.

Yeah. Actually I was just worried about sending note-off for notes that
are already off. Hopefully most synths won't blow up if you do that.
What I'm working on is a kind of a peculiar case where the user switches
operating modes (while playing a chord for instance). It's not a normal
operation, but someone will do it sooner or later. Doing this might
cause me to lose the list of notes to turn off (actually I will know
which ones were down, but I might not know if one or more notes offs
have already been sent). So it would only be as many notes worse case as
they had fingers on the keys (one hand, the other hand is switching
modes improperly).

> If system reset is implemented correctly (supposedly it sets 
> everything back to power-up conditions), you might not get the MIDI 
> chain in working order without manual intervention, so I'd avoid that 
> too.

For the JX3P it is indeed power-up conditions. ;-)