add other MIDI sequencing bytes to clock output

[catfact]

currently, beatclock sends 0xfb to connected midi devices on start, which is a Continue message.

makes sense to me, but it could also be 0xfa (Start), different behavior but maybe more commonly needed.

in addition, there are the Song Position Pointer and Song Select messages (0xf2, 0xf3). my understanding is that it should not be necessary to always transmit these with the clock, but maybe i'm wrong, or maybe SPP+clock is a common enough requirement to make it a default behavior.

[Dewb]

In my experience (which is not necessarily representative) most devices implement 0xfa Start as a reset to position 0 + start running. 0xfb Continue just starts running, which is slightly safer, as it doesn't reset already-running devices, and if you're intentionally trying to follow an ear phase rather than sync phase on all devices, 0xfa results in surprise.

An 0xfa start is certainly valuable when you do want to reset phase on all devices on start. Having an option on the clock params to control this, something like "send message on clock start: start/continue/none" would probably be a good idea, and so would the option to zero SPP (though I'm less familiar with the use cases there in practice. I suspect the goals of deeper SPP integration would be highly script-dependent, and better implemented at the script level.)

[catfact]

thanks @Dewb, that's the feedback i was looking for. i don't use midi clock enough to know what are best practices, never have.

i've noticed that e.g. ableton live has (two?) different midi clock modes that seem to involve SPP and/or continue/start, and whatever those are doing seems like a good model to emulate.

agree that it is looking like some kind of option is called for

[tehn]

this is now likely relevant to the midi output section of the clock system

see https://github.com/monome/norns/blob/master/lua/core/clock.lua#L255

[catfact]

bumping this because it has indirectly come up several more times in past months ("midi clock isn't working right with xyz")

emulating whatever ableton clock TX modes are doing still seems like a reasonable and acheivable way to avoid undue user surprise.