Your favorite old digital tech.

[Notes_Norton]

As long as it stays compatible with 5pinDIN, I'm happy.

 

While I look forward to embracing improved products, I am heavily invested in the old.

 

On stage, I use a Yamaha Wind MIDI controller to run 2 synth modules (I carry 3 - just in case).

 

Mrs. Notes uses a Buchla Tactile MIDI controller to run one synth module.

 

When I make my backing tracks, I use a MOTU MIDI express to mix the output of an array of synths (all 5pin) into a mixer. Since the latency of all the synths is for all practical purposes the same, I can choose whatever voice I want from whatever synth I have for what I feel is the best sound.

 

I've never found the response to be noticeably slow. Of course, it's one-way, but for these purposes, that's OK.

 

Notes ♫

[Anderton]

Getting back to gear...Peavey DPM-3 synthesizer. People didn't take it seriously at the time, which was too bad - the "digital phase modulation" (whatever the heck that was!) had a unique kind of clarity. I was able to get sounds with it I haven't gotten from anything else. Also, their PC-1600 fader box was a mainstay of my live performance act (and it still works, so if I ever get to play live again, it will be part of the setup).

[JohnG11]

19 hours ago, Anderton said:

As I understand it, MIDI 2.0 sends a data packet that can be understood by both MIDI 1.0 and 2.0 devices. If the MIDI 1.0 device can't identify itself as a MIDI 2.0 device to a Capability Inquiry, then the MIDI 2.0 device speaks to it using the MIDI 1.0 language. The MIDI 2.0 specification is available for download. 

 

Sadly, not entirely true. (according to my interpretation of the spec.)

Reading the MIDI CI specification "M2-101-UM_v1-1_MIDI-CI_Specification.pdf"  (which I downloaded and read when it first became available many months ago) it says, first of all, that the interface used has to be bi-directional.

(BTW, Craig, thanks for the repetition of the link.)

MIDI 1.0 via the DIN plug & socket is, for the most part, unidirectional. I.e. one end (out) sends commands to the other end (in).

There is no required response to these commands e.g. ACK or NAK from the receiving device.

Thus a single cable is all that is required to connect e.g. a wind controller to a hardware synth e.g. a VL70m.

The exception to this is when using a bidirectional link (two DIN cables, in and out) to implement MIDI "dumps" via exchange of SysEx commands.

(See the "MIDI Sample Dump Standard" starting on page 35 of the MIDI 1 spec.)

 

The MIDI CI command, as specified in section 5.3.1 of the MIDI 2 document above, defines a new Universal Non-Real Time SysEx message with a Sub ID#1 "0DH" not contained in the original MIDI 1.0 specification. (See Table VIIa in the MIDI 1 spec.)

If it's not defined as part of the specification to which a device was originally built, then clearly (IMV) an old MIDI 1 device is unable to respond.

The MIDI 2 specification goes on to say that if there is no response to a MIDI 2 CI message then just use MIDI 1. (my paraphrasing.)  i.e. it falls back to MIDI 1 (over USB).

 

So, in summary, for a current device to be able to work with MIDI 2 it must have, at least, a USB interface (or any future bi-directional interface e.g. Wi-Fi, Ethernet, Bluetooth(?), etc.).

DIN will not do.

Here's the start of section 1.1 of the above MIDI CI document:

"MIDI-CI defines an architecture that allows Devices with bidirectional communication to agree to use
extended MIDI capabilities beyond those defined in MIDI 1.0, while carefully protecting backward
compatibility. MIDI-CI features “fall back” mechanisms so that if a Device does not support new features
MIDI continues to work as defined by MIDI 1.0."

So MIDI 2 "falls back" to MIDI 1. (over USB)

 

But what about if my synth only has DIN, like my AN1x or my VL70m, or my CHeetah MS6, or ... ?

Reading the specification leads me to believe that none of these older, non-USB devices can even connect to a MIDI 2 device unless, that is, that device also sports DIN sockets.

 

19 hours ago, Anderton said:

As far as zero product, that will take a while. It's not just about the pandemic and no trade shows, but also "gating elements." For example, Apple only recently added native MIDI 2.0 drivers to their OS. I don't think Microsoft has yet. So, there's not much incentive to show products that won't run on current platforms   Also, development tools that will expedite the design process are just starting to hit the world.

This is the old dilema, isn't it, of which comes first.

MS says "no point in implementing MIDI 2" there are no devices.

Device manufacturers say "no point in implementing MIDI 2, there's no OS support.

 

19 hours ago, Anderton said:

Remember that the MIDI 2.0 spec is very much about being ready for the future. Look how long it took for MPE or USB transports to become part of MIDI 1.0. This is why it was so important to allow people to use existing MIDI 1.0 gear as the spec and products evolve. Besides, not all MIDI gear needs all of MIDI 2.0's capabilities. There's nothing "wrong" with MIDI 1.0, obviously it's worked well for decades...so there's no need to just throw it out, and for companies to say "okay, everyone has to get MIDI 2.0 gear now." 

I accept your argument.

My only response would be (coming from the world of CCITT and IEEE) it's important not to start talking about something until it's almost ready to be, quite literally, "unveiled".

(My job, before retirement, was as a protocol analyst in the world of data communications both terrestrial and via satellite, working internationally. I began in the early seventies.)

 

19 hours ago, Anderton said:

The hype about the protocol itself is justified. It's a very well-thought-out spec, and there's a huge amount of potential in there. It's designed to (hopefully) be relevant for decades, just like MIDI 1.0.

I agree about the quality of the specification. It certainly looks the part and, believe me, I've read a good number of specifications in my time starting way back with IBM 2780 and 3270 protocols.

I think the IEEE Q.921 was the one I struggled most with. (Data Link Layer of the ISDN D channel signalling protocol.)

My view, gained from experience in my field, is that it's not a good idea to announce something too soon before availability.

The trouble is that positive vibes can very quickly turn negative.

 

I love MIDI 1, I visit the MMA forum several times daily, as a moderator, getting rid of spam and answering questions here and there.

(I answered a lot on the original forum which is why, I think, Athan made me a mod. Dunno!)

I use a DAW almost daily to create orchestral acompaniments for my wife, an operatic mezzo.

 

But, at the moment, I remain to be convinced that we really need the extra complexity of MIDI 2.

And just how many manufacturers will implement it when they can't even get a 14 bit pitch bend to work properly?

Well, we'll see.

 

Sounds very skeptical doesn't it?

Didn't mean to do that.

 

JohnG.

[Anderton]

I think you're missing a fundamental point. Consider this scenario:

 

MIDI 2.0 master keyboard with ONLY USB

MIDI 1.0 synthesizer with ONLY 5-pin DIN in and out

 

Regardless of whether you use MIDI 1.0 or MIDI 2.0, you need a USB-to-DIN adapter cable, of which there are several (or go through a computer, Bome Box, etc.).

 

Master keyboard sends MIDI Capability Inquiry over USB.

Synthesizer doesn't respond.

Master keyboard says "hello? anybody home?"

No response.

Master keyboard says "okay, I'll talk to you using MIDI 1.0 data." Sends out data packets over USB that end up being received by MIDI 1.0 synthesizer.

 

USB-to-MIDI adapter cables have DIN MIDI in and out, so the reverse could happen, where a DIN master keyboard sends MIDI 1.0 data to a MIDI 2.0 tone module. The tone module can understand MIDI 1.0 or MIDI 2.0 data. So it reacts to the MIDI 1.0 data.

 

As to support, there's already Apple, so I assume Microsoft and Google aren't far behind. 

 

<<But, at the moment, I remain to be convinced that we really need the extra complexity of MIDI 2.>>

 

Wouldn't you like to have ONE hardware controller that could hook up to any MIDI device, and would auto-configure any faders, rotary encoders, switches, specifically for that device? And label the controls on LCD scribble strips? And common parameters like filter cutoff control would map to the same physical control for every synth? And if you had a MIDI 2.0 hardware synth connected to your computer, your DAW would be able to see and edit its parameters onscreen, without needing an editor - just as if it was a virtual instrument? And you'd have no stair-stepping or artifacts when moving faders on MIDI-controlled mixers? I sure would! And those are just things I could use now. If new controllers and VR start to become mainstream, the spec has enough available controllers and capabilities to accommodate them. 

 

FWIW timing jitter goes away with MIDI 2.0, too. There are a bunch of little tweaks like that, such as the whole NRPN thing basically being treated as conventional controllers. That alone is a nice simplification.

 

 

 

[robmok]

How could I resist a thread on old digital tech? 🤣 -   I also miss the Deltalab, was an amazing piece even after it had been around for a while.

 

The one piece I can't seem to replace is the Korg S3 Drum Machine.  An amazing little box, with a very unique take on the matching of "heads" and "shells" for drums... not to mention SMPTE sync to work with tape machines (though I don't really miss tape machines that much).   In the end, the sonic character of the S3 has yet to be really replicated, IMHO.

 

[JohnG11]

Fair point, Craig, now consider this:

I have two Yamaha MU1000-XG modules, each has 2 MIDI ins a midi out and a MIDI thru.

each has a variety of PLG cards in it from AN, DX, VL, PF and VH.

(Total 8 MIDI Din sockets)

I have a Roland SC-8850, same config. as the MU's no PLG.

(total 12 MIDI sockets)

I have a Yamaha VL70m synth MIDI in, out & thru

I have a Yamaha AN1x; inn, out, thru.

I have a Cheetah MS6 synth in, out, thru.

(total 21 ports so far)

Okay, I don't use MIDI thru, so knock off 6 ports, leaves 15.

I have an Akai EWI 4000s wind synth with a MIDI out.

I have a small controller keyboard with MIDI out.

 

BUT ... I have a MIDI patchbay with ins along the top row and outs along the bottom row. All labelled up.

I have this box alongside me and the synth kit is mounted over the other side of the room in a bookshelf. (How long can a USB cable be?)

Patching the EWI to any synth is one patch cable in a new socket, ditto the keyboard.

I can send SysEx messages or a set of patches via MIDI to any of the synths from the PC to configure it.

 

Now, to do this over USB I do ... what?

How many MIDI/USB adapters do I need to buy?

Do I need a computer in the network to control the USB protocol?

 

I replace my controller with a MIDI 2 controller. Will it give me more than ten bits of pitch bend?

 

How many times do I have to rotate my new control knobs to give me 16 bit accuracy?

(Current controllers are 25 clicks per turn, so from 0 to 127 is five full turns. So to go from 0 to 65,535 (16 bit) is more than 2500 turns. Or if not this way, how?)

Lastly, timing jitter was largely unheard of until USB was introduced.

A polled protocol, by its very nature, tends to introduce timing issues.

Did we hear about timing problems over Din in the past? Not that I recall.

 

In my view there are a number of issues that don't seem to be being addressed.

 

JohnG.

[robmok]

2 hours ago, Anderton said:

Getting back to gear...Peavey DPM-3 synthesizer. People didn't take it seriously at the time, which was too bad - the "digital phase modulation" (whatever the heck that was!) had a unique kind of clarity. I was able to get sounds with it I haven't gotten from anything else. Also, their PC-1600 fader box was a mainstay of my live performance act (and it still works, so if I ever get to play live again, it will be part of the setup).

 

I think the Peavey name hurt the DPM-3 more than anything, as nobody expected that from them.  It was an impressive module for sure.

[Anderton]

10 minutes ago, JohnG11 said:

Now, to do this over USB I do ... what?

How many MIDI/USB adapters do I need to buy?

Do I need a computer in the network to control the USB protocol?

 

I think you're overthinking this! If you don't see any advantage of MIDI 2.0, don't use it - just keep using what you're using   That's the whole point of being backward-compatible. 

 

12 minutes ago, JohnG11 said:

How long can a USB cable be?

 

16 feet for passive cables, much longer for active cables, and around 160 feet with Cat-5 extenders. The Bome Box can hook up to networks.

 

16 minutes ago, JohnG11 said:

Did we hear about timing problems over Din in the past? Not that I recall.

 

There were definitely timing problems as soon as any computer entered the picture - so much so that I wrote a article for Keyboard magazine about timing jitter and MIDI data that graphed the jitter over time. Part of this was to counter the marketing hype about "we have 256 PPQ" and then another sequencer company would say "well we have 1,024 PPQ!" as if that was better. What the article showed was that the jitter at 1,024 was four times what it was at 256 PPQ, so you ended up with the same resolution anyway   The Atari had the tightest timing because MIDI had internal hooks into the processor itself. But, none of the MIDI timing was time-stamped. It got sent along when it got sent along. With MIDI 2.0, timing data is time-stamped. Of course, there were also complaints about the scanning rate of MIDI 1.0 keyboards, and how chords ended up with timing differences between notes played at the same time.

[Anderton]

29 minutes ago, robmok said:

 

I think the Peavey name hurt the DPM-3 more than anything, as nobody expected that from them.  It was an impressive module for sure.

 

True story: I was visiting Keyboard magazine, and saw a DPM-3 sitting around. I told them I really liked mine, and wondered when they were going to cover it. They weren't really interested...c'mon, it's Peavey, they make amplifiers for country guitar players, right?

 

Fast forward six months, and another visit. I played one of my latest songs at the time, and the person I had talked to previously was floored. "Wow, those synthesizers sounds are incredible! What are you using?"

 

You know the answer

 

 

[Dave Bryce]

3 hours ago, robmok said:

 

I think the Peavey name hurt the DPM-3 more than anything, as nobody expected that from them.  It was an impressive module for sure.

 

No question.  We had those in stores I was working in.  No one would even be consider looking at them.  Their C8 controller was also a tough sell…although the fact that it featured a dot matrix pic of Hartley Peavey during the boot routine (really!) may not have be helped. 

 

I think General Music made a similar mistake in the US - should’ve stuck with the GEM name they used in Europe.

 

dB

[Notes_Norton]

I have an old Peavey SP sample player. I recorded a finger plucked jazz bass, and various percussion sounds like a left and right hand snare drum pair so rolls don't sound like machine guns, cymbals hit in various places from the bell to near the rim for the same reason, and some other percussion instruments not included in the GM specification.

 

The jazz bass has flatwound strings and sounds more like the bass you hear on "We Gotta Get Outta This Place" by The Animals. It works great for a lot of songs of that era, and none of my other synths have a bass that sounds quite like that.

 

The Peavy SP works, although I have to feed it a floppy disk every time I turn it on.

 

A way to substitute flash memory for a slow-loading floppy would be a product I'd be interested in.

 

Notes ♫

[The Real MC]

There was a similar reaction when Alesis announced the Andromeda.

Alesis?  An analog polyphonic synth?

There was a lot of disbelief and reluctance from the analog folks on Analogue Heaven.  Only when the curtain was rolled back to reveal the members of the beta test/sound design team (about a dozen) with many familiar names did people start to take it seriously.

[Dave Bryce]

1 hour ago, The Real MC said:

There was a similar reaction when Alesis announced the Andromeda.

Alesis?  An analog polyphonic synth?

There was a lot of disbelief and reluctance from the analog folks on Analogue Heaven.  Only when the curtain was rolled back to reveal the members of the beta test/sound design team (about a dozen) with many familiar names did people start to take it seriously.

Tell me about it. 

 

As you know, I was the one who got to shepherd it through sales and marketing development and figure out how to sell it to the dealers and customers.  There were actually a bunch of folks at Alesis who thought we'd be lucky to sell 50 total.

 

They were wrong. 😁

 

dB

[The Real MC]

7 minutes ago, Dave Bryce said:

Tell me about it. 

 

As you know, I was the one who got to shepherd it through sales and marketing development and figure out hiw to sell it to the dealers and customers.  There were actually a bunch of folks at Alesis who thought we'd be lucky to sell 50 total.

 

They were wrong. 😁

 

dB

 

Back in 2000 when the Andromeda was first shown at NAMM the only other new analog polyphonic was the Omega-8 from Studio Electronics.  By the time the A6 went into production, only Alesis and SE had new analog polyphonics with analog VCOs/VCFs/VCAs in production.  The asian makers were very late to the game.  Alesis sales/marketing failed to recognize that Alesis was in a position to fill the hole with a product feature set that would put them miles ahead of the competition.  The current market value of used Andromedas speaks volumes of that.

[JohnG11]

20 hours ago, Anderton said:

I think you're overthinking this! If you don't see any advantage of MIDI 2.0, don't use it - just keep using what you're using   That's the whole point of being backward-compatible. 

 

How is that overthinking?

MIDI 2 simply does not take into account anything that is MIDI 1 and DIN. It's not backward compatible, it just pretends DIN plugs don't exist.

Any MIDI 2 device will NOT communicate over DIN ... unless, that is, it incorpaorates DIN sockets and reverts totally to MIDI 1. But that's not in the specification.

DIN Backward Compatible ... NOT.

 

How do I create a MIDI 2 patchbay similar to the DIN one I described above using USB, assuming all my kit has USB sockets?

 

You didn't answer the question about rotations of a Contoller to achieve 16 bit resolution.

I've asked this question several times and never received an answer.

I have a solution, (but I don't like it) which allows single bit accuracy if required.

Have a three way switch above or below the controller, marked "course \ medium \ fine" this allows the switching of the resolution between 65,536 / 25 then 65,536 / 25 / 25 lastly 65,536 / 25 / 25 / 25. the fine tuning gives approximately four turns end to end.

Not one I'd like to use live tho'.

 

However, take a long throw fader length of 104mm, how do I use that to achieve single bit accuracy at 16bits?

You'd be challenged to get single bit accuracy using just 7 bits. i.e. 127 different positions.

(Oh! MIDI 1. Now why do I like that so much?) ;-)

 

A chord arriving with each note at a different time has little to do with PPQN OR scanning rate, providind it's fast enough.

MIDI is a serial data transfer not a parallel one. Notes in a chord will always arrive one after the other it's the nature of a serial interface.

It's made even worse by using a polled protocol such as that underlying the USB interface. All notes in a chord may not arrive within the same packet, so there may be even worse separation of the notes.

Could you explain how time stamping solves the problem of jitter in a live situation please?

 

I really do look forward to seeing how some of the above issues are resolved in actual product, if I live that long. I'm 75 now so I may never see it.

 

[/diatribe]

 

Still no response from Roland, I take it, on the question I asked in the other MIDI 2 thread.

 

Sorry, I'll make this my last diatribe about MIDI 2.

As a musical electrical engineer i foresee several issues which need ironing out and others that haven't been addressed.

The focus is IMV too much on the protocol and not enough on how one makes it work.

MIDI 1 was very practical, all praise to its inventors.

 

JohnG.

[Anderton]

8 hours ago, JohnG11 said:

DIN Backward Compatible ... NOT

 

I think there's a fundamental misunderstanding about backward compatibility. MIDI 2.0 is a language, and the language itself is indeed backward compatible because it can communicate with MIDI 1.0 or MIDI 2.0 devices. Your MIDI 2.0 keyboard will be able to communicate with a MIDI 1.0 tone module.

 

MIDI's hardware transport is separate from the language. Unlike MIDI 1.0, which was tied to the 31.25 kBaud serial port, MIDI 2.0 was designed from the ground up to be transport-agnostic. The assumption was that transports would change in the decades ahead, and MIDI 2.0 had to be compatible with whatever was thrown at it. Currently, much of the work is focused on USB, but I suspect MIDI 2.0 will eventually work over AVB and other transport protocols. I don't think there's any technical reason why a MIDI 2.0 device couldn't send data over 5-pin DIN. However, since the MIDI 1.0 receiver wouldn't be able to understand the language, it doesn't make sense to prioritize this ability. As to whether a MIDI 2.0 device could send the MIDI 1.0 part of the packet over 5-pin DIN, I believe that's also technically possible. However, I think it would be a hard sell to engineers who would rather advance MIDI 2.0. I suspect the answer would be "why not just use a USB-to-DIN adapter or Bome Box?"

 

Your current setup works fine, so there's no need to even consider what MIDI 2.0 may or may not do in the future. If at some point MIDI 2.0-compatible hardware could provide a more reliable and efficient system for your needs, you would have the option to integrate or replace.

 

8 hours ago, JohnG11 said:

You didn't answer the question about rotations of a Contoller to achieve 16 bit resolution.

I've asked this question several times and never received an answer.

 

That's probably because the answer would vary, depending on how an engineer wanted to implement it. Currently, the dominant solution seems to be acceleration - if you turn a rotary control encoder faster, it covers more range. When you slow down, it covers a smaller range. But that's not the only option. 

 

8 hours ago, JohnG11 said:

However, take a long throw fader length of 104mm, how do I use that to achieve single bit accuracy at 16bits?

You'd be challenged to get single bit accuracy using just 7 bits. i.e. 127 different positions.

 

The advantage isn't being able to park a fader at a place with an accuracy of a single bit. The advantage is not having stair-stepping as you move the fader, especially with low-frequency audio that has few harmonics. For example, the Panasonic DA7 mixer didn't stair-step because it interpolated MIDI to 1,024 steps.

 

8 hours ago, JohnG11 said:

A chord arriving with each note at a different time has little to do with PPQN OR scanning rate, providind it's fast enough.

MIDI is a serial data transfer not a parallel one. Notes in a chord will always arrive one after the other it's the nature of a serial interface.

It's made even worse by using a polled protocol such as that underlying the USB interface. All notes in a chord may not arrive within the same packet, so there may be even worse separation of the notes.

Could you explain how time stamping solves the problem of jitter in a live situation please?

 

The speed of all modern transports is vastly faster than 31.25 kBaud. Correlating Baud to bit rates isn't trivial, because it depends on several aspects of the UART itself, as well as the number of symbols used in transmission (I think MIDI used two, I don't know for sure). But USB 2 goes up to 480 Megabits per second, so there's plenty of time to slip time-stamping in the data stream. As to keyboard scanning, correct, that doesn't relate to PPQN except when sequencing. Stand-alone, the scanning delay results from slow internal processors within MIDI keyboards. When driven by other devices, like a sequencer, then it's a combination of the slow serial interface and slow internal processors.

 

8 hours ago, JohnG11 said:

As a musical electrical engineer i foresee several issues which need ironing out and others that haven't been addressed.

The focus is IMV too much on the protocol and not enough on how one makes it work.

MIDI 1 was very practical, all praise to its inventors.

 

MIDI 1.0 was brilliant, but don't forget it didn't arrive fully formed by any means. 16 channels was clearly not enough, so ports had to come into being. Not being able to transfer samples until the Standard MIDI File was a problem. Not having a standard instrument set or MIDI file format held back consumer adoption. There was no way to sync it with lighting and other stage elements. It wasn't compatible with any hardware protocol beyond 31.25k serial ports. There was no ability to work with SMPTE devices, or control machines, which stalled adoption by the film and TV industries. It was not possible to attach controller data to individual notes, which held back development of polyphonic controllers. There was no defined way of dealing with instrument articulations.

 

All of these were major issues that were resolved over time. I don't think MIDI 2.0 will be any different. MIDI 2.0 was specifically designed to accommodate needs that may not exist now, but likely will in the future, as well as being able to communicate with MIDI 1.0 devices. But again, let me emphasize - you can ignore MIDI 2.0 completely if you want to stay with MIDI 1.0 gear.

[robmok]

On 5/3/2022 at 4:46 PM, The Real MC said:

 

Back in 2000 when the Andromeda was first shown at NAMM the only other new analog polyphonic was the Omega-8 from Studio Electronics.  By the time the A6 went into production, only Alesis and SE had new analog polyphonics with analog VCOs/VCFs/VCAs in production.  The asian makers were very late to the game.  Alesis sales/marketing failed to recognize that Alesis was in a position to fill the hole with a product feature set that would put them miles ahead of the competition.  The current market value of used Andromedas speaks volumes of that.

 

Thanks for that very kind, and insightful, comment...  As @David Brycealready noted, he sheperded it through sales and marketing, including naming it Andromeda.   In engineering, and one of the two (the other being the famous Erik Norlander) that convinced Keith Barr that we really should tackle an "analog" synth, we really did think we were "on to something" with the idea.    Custom silicon of that scale has not been replicated since (that I am aware of), at the very least speaks volumes to the seriousness of engineering that went into that product.