----------------------
USB midi is one of the worst thing that happened to MIDI. If I have a synth with USB midi and a controller with USB midi, I can't connect them to each other, which is idiotic and frustrating.
The original MIDI spec had its drawbacks, but the biggest have been addressed by MPE.
As for latency, jitter, granularity — as a gigging musician and producer, I've never had any issue with it. It's good enough. And most of the music you hear today is a reflection of that.
What's not good, it's that some manufacturers are ditching the 5-pin MIDI jack for USB or Bluetooth. Thanks, nobody asked. Same for losing the THRU port.
Because what musicians love is setting up their computer on stage while everyone is waiting, right?
Bidirectionality with MIDI can be achived by replacing the dual 5-pin MIDI with the mini-DIN (yes, the same connector the old PS/2 mouse used). Yamaha Tenori-On and and Reface series do that (though again, nobody asked).
And the entire NRPN space of MIDI is unused because it's unstandardized. So simply agreeing on how MIDI 1.0 should be used would address most of the problems that MIDI 2.0 claims to solve.
That's what MPE does, by the way, and that's why it has been silently becoming the de-facto standard.
The examples I see in promotional videos show a computer talking to MIDI 2.0 devices. We've had enough of that. What made MIDI 1.0 work is the simplicity of device-to-device connectivity, regardless of what the devices are (and whether one of them is a PC).
Think of this magic: if I have just ONE type of cable, I can connect ANY device to ANY OTHER device.
That means my Yamaha MX-88 I bought new last year can talk to Korg Poly-800 made before I was born.
I could have bought the cable in the 80s or yesterday, it's the same cable.
What MIDI 1.0 gave people is less thinking about how to connect devices, and more about making music.
USB midi did the same thing for devices connected to a computer, at the expense of being usable in a setup that does not have a computer. That implicit assumption — that everything will involve a computer — was flawed.
WTF, I want to use my ROLI to play my Yamaha — and I can't, because ROLI is too "modern" to talk to any hardware synth made the same year.
I hope that MIDI 2.0 avoids this fuckery, but I can't tell yet. Bidirectionality looks sus.
My litmus test is: if a device doesn't come with a 5-pin MIDI jack while having space for it, it's a gimmick (Teenage Engineering being an exception, because they make up for it with analog control: my pocket operators sync with my Korg Monologue via pulse, and Korg syncs with Yamaha over 5-pin MIDI).
And if Korg Volca could do it, so can pretty much everything else (hint: don't put it on the side, put it on the front panel if you want that slim look).
MIDI 2.0 can be a huge success if you wouldn't be able to tell a MIDI 2.0 device from a MIDI 1.0 at a glance; i.e. same connectors, same interoperability. Then it would be truly an upgrade, a 2.0, and not "thing that works like MIDI in some cases, but have fun buying dongles to connect old equipment to new gear".
MPE works on that model; and connectors aside, any multitimbral device is MPE-compatible.