Re: midi looping fixes

On Tue, Sep 13, 2016 at 07:58:30PM -0500, Paul Davis wrote:
> One other clarification: if a loop *begins* in the middle of a note, that
> note will not play. If it ends in the middle of a note, the note will play,
> but will be truncated (turned off) at the loop end.
> 
> Fixing this requires some deep algorithmic design, since it requires the
> ability to replay MIDI events from the beginning of time, on demand, in an
> RT context.
> 
> On Tue, Sep 13, 2016 at 7:46 PM, Paul Davis <[email protected]>
> wrote:
> 
> > I just pushed a bunch of commits that should at least massively improve
> > MIDI looping, and at best fix it completely.
> >
> > I'd appreciate it if anyone with an interest in such things testing out a
> > nightly or self build (starting with nightly builds from September 14th).
> > The more people that play around with it, the higher the chance that we
> > catch some aspect that I failed to test.
> >
> > Please note: this only covers *playback*. It does not affect MIDI loop
> > recording (the status of which I don't know right know).

That's great, the loop of the recorded midi events seems correct.

A minor detail: I'm noticing truncated notes during an improvisation;
it makes sense if we are recording the events, but I think it's better
to avoid a forced midi noteoff at the end of the loop if the midi
track(s) is in monitor input mode ("In" button pressed, no recording),
because there aren't notes to write.