Microsoft opens MIDI 2.0 repo

love the 5 pin connectors as well. Its like the 80s are still here.

Its the same feeling working with old test equipment. They have these massive Centronics-style GPIB ports that you can either connect to a converter or an Ethernet adapter. The great part is that the commands from an old unit almost always work on a brand new unit with similar features. I had to use an old spectrum analyzer from the time before floppy diskette drives were common and moved up to a brand new one that ran Windows 10. I could directly connect to the Ethernet port and send the exact same commands to get my measurements.

Can you share what precisely you were pushing for. How did you phrase it to others at the time?

To avoid sounding like you're trying to take some credit you might have phrased it:

"This is something I've been wanting for 22 years."

It's great. I have a 8x8 motu midi rack that i've had for probably 25 years, might be one of the only consistent pieces of rack gear I've had.

yep. MIDI is great for a keyboard/piano interface with velocity, pitch bends and a sustain pedal, maybe some filter modulation if you're feeling frisky. But using it for guitars and orchestras? Use the right keyswitch note to change the articulation, use a different keyswitch to pick the "right" note (since many instruments e.g. guitars have multiple notes that play a C3), maybe the actual notes themselves don't make a sound but instead shape a chord that can be strummed by a separate strum trigger note. Oh and the dynamics of the sound might be controlled by a CC, which shows up in Kontakt if you're lucky and requires looking at the manual if you're not. Oh yeah and none of this is standardized, so a Spitfire orchestra library's keyswitch controls are entirely different from EastWest. Fun.

Sounds like MIDI 2.0 will help with some of this, at least.

Yes. MIDI is the sequence data protocol, not the transport specification. There are multiple alternatives to the basic, eg Elektron had a proprietary system that gives ~10x better speed/timing accuracy.

MIDI is very similar to RS232 https://www.compuphase.com/electronics/midi_rs232.htm

But RS232 dates back to the 60s https://en.wikipedia.org/wiki/RS-232

> In telecommunications, RS-232 or Recommended Standard 232[1] is a standard originally introduced in 1960[2] for serial communication transmission of data. It formally defines signals connecting between a DTE (data terminal equipment) such as a computer terminal, and a DCE (data circuit-terminating equipment or data communication equipment), such as a modem.

Wikipedia says TCP is from 1974 (https://en.wikipedia.org/wiki/Transmission_Control_Protocol). That would mean it turns 50 next year.

There likely are way older protocols in telephone networks.

Depends on what you consider a "protocol". Morse code dates back to 1844 and is still in use. Hellschreiber is from the 1920s, still in use by some hams. You can downlink slow scan TV (1957) images from the ISS sometimes. Except for Morse code, a lot of the older protocols used by hams are mostly for novelty, legacy, and price reasons as they can be cheaply implemented into just about any computer ever made. Morse code still has use in the military as a backup system for long range communications.

> What are the major features of MIDI 2 over MIDI 1?

"What Musicians & Artists need to know about MIDI 2.0": https://www.midi.org/midi-articles/what-musicians-and-artist...

MIDI 2 has huge vendor buy-in. MS isn't leading the charge, Google and apple are supporting it too. I posted this because I thought the repo might be more immediately useful than the pure specifications, which were released a few weeks ago.

There's some boring/obvious QoL stuff like expanding the notion of channels into groups + channels (16 groups of 16 channels in MIDI 2 instead of the implicit single group of 16 channels on MIDI 1), better synchronization messages, higher resolution control changes/velocity/aftertouch, polyphonic control changes, and so on. Basically the limitations of MIDI 1's 8 bit protocol are being lifted.

But the big difference is that MIDI 2 is duplex (devices can send and receive messages from each other, rather than just broadcasting them into the void). This allows for discovery (devices can find each other on the same connection/network), capability inquiry (devices can get metadata about each other) and the super power: property exchange (get/set properties as JSON).

The reason the last bit is so big is because it fixes some giant issues with MIDI, like being unable to control the tuning system used by all the connected devices. With some conventions and the new protocol we should see MIDI 2 open up a lot of room for non-western and microtonalities without resorting to tedium or hacks.

But that's just an example. It's kind of like a binary protocol for hardware and software music devices that goes beyond basic performance controls.

Yes and yes, it isn't microsoft that started pushing it.

Biggest promises on top of better data types is actual device discovery, so for example after connecting control keyboard to a synth you can have the knob names automatically populated. And no hacks needed to get MPE

And how does it compare to OSC (Open Sound Control)?

It has broad vendor buy-in from the vendors that matter much more than OSes: Instrument manufacturers. Yamaha, Korg, and Roland all demo'ed MIDI 2.0 prototypes at NAMM recently.

They own this one website that hosts a few open source projects for free…

https://github.com/

Maybe you’ve heard of it.

In all seriousness, I would say that today Microsoft does a better job supporting open source projects than most of their peers.

I’ve noticed a lot of corporate sponsored open source projects favoring the MIT license these days.

Yes, but canyon.mid and passport.mid will always have a place in our hearts <3

It's mainly getting more traction recently thanks to the push from MPE gaining more popularity. MPE improves on MIDI in several crucial ways, mainly allowing for much better polyphonic parameter controls. However, it's also kind of a big hack internally, hijacking things like MIDI channels to work properly. MIDI 2.0 is largely a response to the proven need, to clean all that up and back the effort up to full industry collaboration (where MPE was a proprietary push from the likes of Roli).

> Man, I'm afraid to get banned by opening a GitHub issue but the usb driver looks bad.

I think the reason why you might get banned is because your tone comes across as harsh and non-constructive.

If instead of saying "The code formatting is awful", you said "I noticed the code formatting is a bit inconsistent and I'm not sure what (if any) code style it is following. Is there any way I can help improve this? Do you have a standard code formatting you follow, or any tools you've used to enforce code formatting? I've used X in the past, and I could contribute a PR to set you up with that, if that's something you are interested in" – nobody is going to ban you for putting it that way

You can fit enough of the MIDI 1.0 spec for a basically working implementation on a single sheet of paper.

Summary of article: The MIDI 2.0 standard has been under development. The standards body which develops it has the rule that all standards development happens behind closed doors, under NDA, but the NDA is lifted when the final version of the standard is formally published. Microsoft has been working on a new version of the Windows MIDI subsystem to support MIDI 2.0. Microsoft wanted to open source it, but until now couldn't due to the NDA. Now the standard is published, Microsoft is legally allowed to change the repo from private to public – https://github.com/microsoft/midi – and this blog post is to announce they've done that.

They are saying they've opened repo for the new version of the Windows MIDI subsystem – https://github.com/microsoft/midi

They couldn't legally open it until now due to the MIDI Association's NDA.