MIDI Manufacturers Association Says New HD Protocol ‘Has Reached A Milestone’

midi-manufacturers-associationThe MIDI Manufacturers Association (MMA) will be hosting an open discussion at the 2015 NAMM Show about a new ‘HD’ standard, designed to be compatible with MIDI 1.0.

The session, scheduled for Sunday, January 25, will cover the trademark, licensing, and product certification options for the MMA’s proposed new industry standard, “HD Protocol”. 

midi-cablesThe MIDI standard – originally proposed by Dave Smith, right – was adopted over 30 years ago, and is now supported by over a billion devices.

For the last several years, the MMA has been working on a new protocol, designed to co-exist with the 30+ year old MIDI standard, but offering some significant new features.

The MIDI Manufacturers Association ‘HD Protocol’

Note that the proposed standard is not intended to replace MIDI, but be a new standard, designed for new applications.

We asked MMA President Tom White to explain the role of the new protocol and how it fits in with the existing MIDI standard:

Technically, HD Protocol is not MIDI because it does not follow the MIDI Specification. While it performs the same functions as MIDI does, it is a different technology. Just as USB and FireWire perform the same functions but are different.

We are aware that musicians may assume our HD Protocol is a replacement for MIDI, but that’s an unrealistic expectation which we do not want to encourage.

MIDI is extremely cost effective and well suited for very low-cost products, while HD is not; and MIDI already handles most applications that people have today, so the added cost of HD can’t be justified in most products.

We believe that HD Protocol does make sense for certain products, but what those products are will be up to individual manufacturers to decide and promote.

We expect HD-equipped products to roll out slowly over time, while MIDI continues to be the primary technology for digital musical instruments for many more years. Hence, we need to promote each technology as a separate (yet fully interoperable) solution.

White also says that the name ‘HD Protocol’ is a working name for the proposed standard and may change.

‘HD Protocol’ Features:

  • Backward compatibility with MIDI 1.0 to ensure wide-spread adoption
  • Plug and Play network connectivity over USB and Ethernet for easy integration
  • Thousands of Channels for handling large complex systems
  • A massive number of high resolution Controllers and Parameters for channels, groups and even individual notes, for unrivalled precision and adaptability to new controller technologies
  • Precise Pitch Control and Articulation messages for expanded expressivity
  • Time Stamped messaging for accurate timing and tight real-time performance
  • Plenty of room for future expansion to eliminate obsolescence

According to the MMA, ‘the HD Protocol Specification has reached a milestone that warrants industry-wide consideration.’ Music technology hardware and software developers attending the NAMM session will learn about the HD Protocol feature set, and participate in deciding the environment needed to successfully develop and launch products that support the proposed standard.

For more information on the meeting, see the MMA site.

Note: This post was updated to clarify how the MMA is positioning the proposed ‘HD Protocol’ and to make clear that, while it is being developed by the MIDI Manufacturers Association and is compatible with MIDI 1.0, they do not consider it a MIDI standard.

via KVR, Sonic State

71 thoughts on “MIDI Manufacturers Association Says New HD Protocol ‘Has Reached A Milestone’

  1. all respect to Dave Smith, who would have thought a protocol this old would last this long, and it’s about time for 2.0!!! They have been talking about it for around 20 years now right? Met the gentleman at NAMM a few years ago, and he couldn’t be nicer….

  2. This should have happened 20yrs ago but finally it’s happening yipee! Kudos to Elektron for pioneering turbo MIDI just a shame no one else was interested.

    1. It would be very helpful if this midi standard somehow worked with the Turbo stuff from Elektron, I’ve no idea how possible this is though.

      But yeah – great to hear this – the future looks good!

      1. You wrote: the future looks good! – MIDI HD protocol is announced by MMA from 2005, no yet specifications, no consencus between major manufacturers. From 2 years ago, I use Copperlan, an amazing protocol!

        1. Yes indeed, CopperLan already exists for years, it is an effective solution, not a promise. You can use it ***today*** as a a cost-free solution for MIDI patching over network. But beyond this, it is a complete command & control framework, cross-platform, really plug & play (working in every condition, even without IP address configured), high resolution, time effective, and much more.

  3. I Wanna say congrad’s to MIDI for all are sound’s of musical Instrument’s that we use, its coming a mighty long way, and happy aniversary for midi for 30 years, since 1985, My first keyboard that i use for midi controlling, is the YAMAHA PSS-480, THAT SMALL SIZE keyboard’s are in the exact same age as MIDI WHEN IT FIRST CAME OUT, AND THAT VERSION IS STILL 00.1 AND WE’RE STILL USING MIDI FOR ARE COMPUTER FRIEND’S RIGHT NOW AND TODAY.

  4. Pretty sure this actually has nothing to do with Dave Smith as he has said many times that the existing MIDI spec is good enough for 99% of applications, and I tend to agree with him.

    1. It’s the MMA, like the article says!

      They are the manufactures that have to decide if they can’t agree to standardize on this, or not!

    2. So let me get this straight, everyone complains about stebbing in VA Digital synths is fucked up, every test they make on VA, they talk about stepping, but in midi it´s ok ? Sure you can do tricks to get a better resolution, but, why use tricks if you can make a new standard- But yeah lets go back to SD television, DVD, cause they all did the job, sure it didn’t look that realistic, but who cares, we could see a picture and hear the sound.

      1. First, let me state that I am 100% all for a new MIDI standard as long as it is backwards compatible.

        When people say that MIDI stepping is a hardware problem, it is at least partially correct. Lets look at this example:

        A VA synth uses a 32bit value to change its filter cutoff frequency. A 7bit MIDI CC is mapped to also control the cutoff. That means that every step in the MIDI CC change is going to change the cutoff by the same as changing the 32bit by 33554432 (if I did my math correctly). That will be heard as a big step. However, there is no reason for the VA synth to step immediately to that new value. The impact of the step can be minimized by smoothly changing the value – having the synth count up between the steps. This can be done in a VA with a simple math equation. In a hardware synth, a capacitor (RC filter) can be used to smooth out the CV changes.

        While this solution can smooth out the steps, the obvious problem is that the value from a MIDI CC change can never be anything between the 33554432 steps on a 32bit filter. That is where improved MIDI resolution will help.

        1. But what “simple” math to use? A linear curve or an exponential curve? How much time do you give it? What if the knob is moving quickly? What if the knob is moving so slowly that the MIDI rarely changes? Do you try to match the speed of the knob even though you don’t know the next time it will change? Do you add latency so that the values can be smoothed with a higher-order curve? It’s actually startlingly easy to find yourself in a tangle trying to do this well.

          1. z = x^2 + y^2 (iirc – my memory is bad anymore)
            It is a 6db lowpass filter. The same thing is used for poratmento.

            But I agree with you! I would like to see a new MIDI standard. It would make things easier in a lot of ways. I was just explaining Mike B’s comment about Dave Smith, and responding to Piv’s stepping example. I saw a video where Dave Smith mentioned the issue in my example. It is a fix that resolves the stepping problem, but it isn’t ideal.

            1. A lowpass filter is indeed a solution. But you’re just smoothing the steps–guessing at the missing data in realtime. You don’t hear the zippering if it’s done well, but it’s just a coverup.

              Sometimes it doesn’t matter and sometimes it does.

    3. Mike B – as noted in the article, Dave Smith introduced the MIDI standard and was instrumental in its adoption. And we’ve seen him express skepticism over the need for updates to MIDI.

      The MMA is the industry group that was formed to promote and manage the MIDI standard, and they are behind the proposed standard.

  5. please god.. can we actually use MIDI to get MOTHERFUCKING ACCURATE TIMING?!?!


    1. I Don’t think any technology can make this possible.

      If you don’t send too many events, they will be tighter, under any protocol.

      You can try sending the clock/sync info through a different cable, that helps a lot.

    2. MIDI 1.0 did let you put in timestamps, but it’s definitely a detail that many MIDI synths don’t bother with. Even though MIDI is defined over the standard wiring, you know that it will get transported across carrier pigeons (or worse… WIFI) if that’s what is available. In general, the design of MIDI 1.0 let synths cut an immense number of corners (ie: 0x90 message is implemented, and it passes a basic plugin kb controller and play a tune test).

      I kind of suspect that they got a lot right this time around, but it’s irritating that it’s all done in secrecy as if there is a well defined set of “manufacturers”. Now a “manufacturer” is any app developer who managed to make a popular synth app.

      1. Corrections, Rob:
        Despite having “Manufacturers” in the name, anyone that is developing a commercial MIDI product can be an MMA member, including app developers.
        And MMA’s “secrets” are no different than any other company’s: we don’t publish the details of products in development (Specs and RPs) until they are ready for release… just like you, I expect. But anyone in the business can join and see what we are doing, no closed doors here.

  6. This is incredibly exciting. The MIDI protocol is clever, and works damn well, but there definitely is room for improvement. I also think it is kind of surprising that the original MIDI protocol has lasted so long before this happened.

    1. Connectivity with USB-MIDI is problematic though. You need a computer to act as a hub to connect everything. There is literally 1 (one!) device, the iConnectMIDI4+, which can actually perform that job as a standalone hardware device, and it’s not even the main focus of the product. And still it’s some extra piece of gear that you need to acquire and bring with you. That’s the sorry state it’s in.

      Hopefully with USB 3.1 maybe things will be better, I believe the new type C connectors are proper bidirectional. I also hope the MMA does things right, I have to agree with Rob Fielding that making a standard behind closed doors is bad practice.

      1. Also, while I’m on the topic, since early 2014 the MMA has on their website the promise to *finally* publish the MIDI specification online in a free digital format, however it’s been saying “coming soon” ever since. Very frustrating, and it does not speak well of the MMA.

  7. I oppose any change to the MIDI standard. People have been ignorantly complaining about MIDI being old or lo-res for decades and proposing alternatives. We don’t need some high bandwidth bloated new standard to struggle with.

    MIDI is not broken. It’s not low-res. You can implement it in ways that eliminate the limitations via dual messages, internal slew, etc. People conflate MIDI with 0-127-step controls – there’s nothing about MIDI that says you have to have 127 audibly discrete steps of filter cutoff. That’s bad hardware, not a bad digital interface standard.

    Everyone that uses and depends on MIDI should be demanding that manufacturers make better and more clever hardware. We don’t need a standard war even if it’s “backwards compatible.”

    MIDI forever.

    1. I think dual MIDI 1.0 and MIDI 2.x support is probably a good idea. But MIDI only seems not broken if you are using a keyboard controller. In every band in the world there are 2 guitarists and 1 bass player – so assume that the MIDI standard is quite underutilized at the moment. The standard has been totally inadequate for creating string controllers that really work correctly (nothing to do with wiring standards or latency so much…but the design of the packets themselves are full of corner cases that bite everybody except simple keyboard controllers and sequencers).

      The design of MIDI is one of a virtual keyboard controller, where it’s simply assumed that this is adequate for emulating other controllers. What it needs to look like… more like controlling additive synthesis.

      The MMA *would* sit on it forever, but if they do, the iOS world is going to give up on them and ignore them. That’s how AudioBus started. Imagine if the iOS came up with a crazy OSC-like protocol that practically forced the hardware manufacturers to support it. That’s what’s happening on iOS today.

    2. you can’t slew a value that’s wrong. If it needs to be 2.35 and your options are 2 and 3, then you can’t meet the spec. this is the main issue with bend width because you have 14bits to represent the bend range. On a string instrument you must be able to bend up from the lowest note to the highest note and be able to stop on any pitch within 1c; now do that with two voices…one going up, and one down.

      It sounds like a dumb corner case, but all guitar controllers are broken because they assume that there is only one A2 note (when you could be on a 12-string guitar controller when the pitches are a few cents apart – a real thing I did to get chorusing in Thumbjam). And you most definitely can dive bomb from a high E on 24th fret down to lowest note (whammy bar); and this is where MIDI 1.0 got seen as a joke.

      If MIDI HD/2.0 fixes this (and it probably will actually), it won’t be the existing keyboards that benefit. It’s the 4x as many instruments being sold that can’t be sold now because even state of the art midi guitar controllers are kind of useless even ignoring latency issues.

    3. “MIDI is not broken. It’s not low-res. You can implement it in ways that eliminate the limitations via dual messages, internal slew, etc.”

      Its not broken, if you limit yourself to keyboard style controllers and don’t do any complex sequencing.

      When you get into advanced controllers, guitar controllers, wind and string controllers, microtonal music, polyphonic aftertouch and all the other places where MIDI has not been widely adopted, though, you end up running into some serious MIDI limitations.

      Another issue with traditional MIDI devices is just shear bandwidth. MIDI, by definition, will never be ‘tight’ without a lot of jumping through hoops. If you try to play multiple notes on a beat, you are just about guaranteed some flanging.

      There’s definitely a need for an updated standard, but I’m not convinced that the MIDI Manufacturers Association will be able to get fiddly-squat done on this. At this point, they’re holding the MIDI standard back as much as they are moving it forward.

      I think it will take some industry giants throwing their weight behind a new standard, like it took Sequential Circuit’s Dave Smith & Roland’s Ikutaro Kakehashi to make MIDI have some momentum.

  8. I agree that the original MIDI is not as bad as people seem to think – it is just not often used correctly. However, I am all for updating and improving it.

    My only concern is the part where they mention licensing fees. As a DIY enthusiast who hopes to sell some some creations one day, I hope that it remains free and open for everyone to use.

    1. i bought a copy of the MIDI 1.0 spec, but licensing fees that must be Apple paying out for it. I just started learning to read the packets on the internet back in the Win95 days…. it’s free for all practical purposes unless you do it in hardware. The only reason there is any benefit to doing it in hardware is that iOS isn’t a hard real-time OS. Putting this standard deep into hardware is one reason it has been stuck in a time warp for 20 years.

  9. Seems like it has some great improvements. Hopefully this makes higher resolution encoders an everyday thing. Very few hardware and software supports 14bit midi. Also infinite channels will be nice and the possibility of having more throughput to avoid bottlenecks in the midi pipe.

    The one thing I think everyone hopes for is rock solid timing without the need for a expert sleepers device. Hopefully this comes close.

  10. If Dave is involved and its backward-compatible, great, but with DAWs as mature as they are, I’m a bit dubious about the real need. As DAW-like as workstations & iPads have both have become, what can a more amped-up spec do in the face of that big a shift since its inception? Also, what happens when the USB spec hits 5.0 or 6.0? I play just one remaining piece of MIDI-only hardware from a controller. The rest is now all MIDI *events*, with even my hardware workstation running MIDI over USB. The concept is still ideal, but the real leap ahead is that the hardware version has been partly superseded by the natural extension into software. I’ll need a bit of convincing to see it as a grand expansion rather than a multiplexing update.

  11. Who needs MIDI2.0? i think nobody. New synths could sync with PC VIA old and popular USB or FW or even a Thunderbolt. Older synths cant even use new features implemented in 2.0 standard so why we should like to use MIDI2.0? its useless.

    We dont need separate standard… we could use universal one but sumbody decided to make money on “revolution” but in reality its kind of regress.

    1. Good old Dave already thought about that back in 1983 – there IS velocity in the note off messages.

      As some commenters have said, MIDI isn’t fully utilized, and can cover for 99% of cases.

  12. It’s sad how many people haven’t got a clue about the MiDi specification and its capabilities….after reading all the comments here that is obvious.

    1. I´m asking for myself why not adopt a non propietary protocol like OSC, that already have those improvements pointed. Just money? What do you think?

  13. MIDI changed my life. I had a Commodore 64, Dr T’s sequencer, a DrumTracks, and a Juno 60 with a DCB-MIDI converter. I got a job at a video prod house and started a life long career. No one EVER complained about any stepping in my music.

  14. Dave Smith rules! Despite that fact, MIDI is still at version 1.0 cos there is nothing more that needs to be added. Did you know that only 2 of the ports are used; of the 5 available. In all this time, there has been no need to update it. I can’t see how HD protocol will help or change anything. Will they let it go out for free, as Dave did all those years ago (and let himself lose out on millions of dollars). I doubt it. So….fuck them.

    1. Exactly. The HD spec has been in the works for almost 10 years or so and it hasn’t got anywhere.

      Back in the days I bet Dave Smith drew up the standard in a jiffy, and it was promptly accepted and shared by all. Unfortunately those days are long gone. These days everybody wants to own everything and earn loads on it, there’s corporate bureaucracy, various agendas, endless meetings etc.. The MMA has turned into something that flies in the face of the spirit of MIDI.

      MIDI as we know it will probably never happen again, as in a musician and small manufacturer -friendly, low overhead protocol, freely available, and being cheap and easy to implement and understand for everybody.

      Good old MIDI isn’t holding anybody back artistically either. If you can’t make music with MIDI, it’s not the protocol or the gear there’s anything wrong with. For those with special requirements there’s always OSC which has been in use for 10 years or so.

      The creators of OSC are the ones who bought innovation and progress in this field, while making it open and freely available. If anything, manufacturers should embrace OSC, not the MMA who hasn’t been able to do anything of worth for 30 years.

  15. There are a few misconceptions in the comments. The trademark, licensing, and product certification options doesn’t mean that there will be a cost for the HD protocol or that it will not be freely available. It just means that there will be trademarks so people can tell what products support the spec, a license so that only products that actually support the spec can use the trademark ( believe it or not there are ‘MIDI” microphones in China, they don’t have MIDI on them, but use the logo!) and possibly product certification so that the products are guaranteed to work as expected.

    OSC is a great control protocol, but it is not an industry standard. There is absolutely no guarantee that two OSC products will talk to each other or work together. MIDI has always been about interoperability and HD should be the same thing because it is handled by a consensus based industry standards organization. These industry organizations take more time because they have to get a wide group of diverse companies to agree on things where back when MIDI was started it was six companies who agreed to start working together.

  16. “The MIDI standard – originally proposed by Dave Smith, right – was adopted over 30 years ago, and is now supported by over a billion devices.”

    Why do I find the 1 Billion number hard to believe? Really… 1 device for about every eight peaple on the Planet? REALLY?

    1. The 1 Billion number is due to the fact that every copy of the Mac and Windows (desktop) OS since 1995 has had a MIDI synth in it, and since about 2000 most cell phones have had a MIDI synth, and every device with iOS in it for the past few years has a had a MIDI synth in it, and the Android OS for the past few years includes a MIDI synth… need I go on, or are you comfortable with “a billion” yet?

  17. It sounds like it might be reserved for high end mixing controllers with a need to perform precision eq tasks on a regular basis. Also huge applications processing lots and lots of channels and parameters. But this is not actually a lot of situations. The current midi standard seems to support Multidimensional Polyphonic Expression really well. It seems crazy that the protocol is 30 years old and when you look at how far everything else has developed around it. The original protocol still allows instrument designers and musicians to continue to be innovative. I think Dave nailed it the first time around that was probably why he laughed a lot at the idea of HD. It sounds like they have progressed this but I still have doubts as to whether it will ever come out and if it does I think it will be reserved for more specific intensive tasks. I think basically original midi protocol will continue to rule.

  18. I hope this new protocol serves more than one goal. Speeding up MIDI is great, adding more dimensionality for expression at high resolution. However, the second most important thing is userfriendliness, or at least facilitating it. I imagine a future where no-one would ever have to remember abstract midi channel numbers for receivers or transmitters, nor would have to make such settings on both instruments match. I would hope for devices to broadcast themselves as agents on a network, and automatically allow (or block) requests for a control connection. Also, I’d hope we no longer would have to set patch numbers or CC numbers (having to look up what they do in an instrument), but rather have that instrument broadcast the names of parameters that it allows control over. Or the names of the patches that can be selected. That would make using such protocols so much more userfriendly!

Leave a Reply