Blog:Lets Talk About SysEX Baby

Blogger Lagrange Audio explains      10/02/14

MIDI 30

Lagrange Audio (aka Jason Durbin) writes: 1983 probably rates as one of the most significant years in all of music technology history. It heralded the introduction of MIDI or Musical Instrument Digital Interface. This protocol was originally promoted by Dave Smith (with some inspiration from Tom Oberhiem) who recognised that as electronic musical instruments began to move into the digital realm it opened the door to solve one of the enduring challenges of the day. This was entirely to do with the fact that although instruments could be connected together in a very limited performance fashion these analog connections made cross-platform and cross-manufacturer interactions extremely difficult and unreliable, if not commercially impossible.

Dave Smith and Ikutaro Kakehashi from Roland demonstrated the way forward at the Winter NAMM show in 1983 using a Prophet-600 and Roland JP-6 and the rest, as they say, is history. While this demonstration only extended to essentially one machine being 'played' or triggered by the other the impact of what was introduced cannot be overstated.

Personally I rate MIDI as the most significant development in music technology after digital audio. Both gentlemen would be awarded a Technical Grammy in 2012 in recognition of their efforts. This article is deliberately not highly technical so for the down and dirty specification I recommend you go to the Midi Manufacturers Association (MMA) who control the entire MIDI spec at http://www.midi.org/techspecs/ if you want to find out more.

The premise of MIDI initially was very simple. In terms of fundamental principles the scenarios being driven were these:

1) A keyboard from one manufacturer should be able to play, or trigger notes on a keyboard from another manufacturer, including basic performance controls;

2) By definition a sequencer should be able to trigger a series of notes on another sound source; and

3) Any tempo driven machine should be able to synchronise the clock, or be synchronised by the clock of another machine, including transport control.

The common element of course is that the principles above promote 'manufacturer independence'. This meant that while a much expanded range of performance control information could be exchanged between devices such as keyboards, sequencers, drum machines and later computers it did not immediately address individual manufacturers that might want to exchange specific information between their own machines. In particular when you consider the challenges associated with getting a bunch of highly competitive manufacturers to even consider adopting a standard the last thing they would entertain were restrictions on their own innovation.

From this environment SysEx or System Exclusive, an extension to the MIDI specification, came into being. It could be argued that SysEx is the one element of the overall spec that has contributed to the ongoing longevity of MIDI because of the flexibility and possibilities it offered. Chief amongst these was the fundamental way that patches on a synthesizer for example could be managed as a library and edited externally. To really understand the significance of this it's worth noting two restrictions that were quickly beginning to emerge.




More From: LAGRANGE AUDIO
Even more news...

10 Comments...  Post a comment    original story
Saxifraga    Said...

I think MIDI was a great achievement for musical machines. But it is now stuck in 80ties tec and has become a bottleneck. The protocol is unable to cope with modern volumes of data. Esp. if you have samplers to manage. This has lead to 'solutions' that are inferior. That´s why there is no workstation with a consisten way to write librarian software and manage big sample libraries. We need all instruments to be able to plug into high speed wlan or ethernet cable/ thunderbolt/USB3 and an API that breaks thetransmission limits of MIDI.

11-Feb-14 04:58 AM


to-pse    Said...

Actually MIDI is still able to cope with high volumes as long as one uses other transports than the DIN serial ports like USB, Firewire or Ethernet.

And with the CC #88 high resolution velocity prefix, one is not even limited to 127 different velocity values...

11-Feb-14 05:55 AM


Superlative    Said...

Mon dieu, Sysex... if there is a good software to transfer Data it is no Problem, but....i want not to write a Sysexfile by myselfe... remember anyone NRP and NRPN??? It was a nightmare to get access to the Filtercutoff and other stuff on the AWE64.... I don't understand the critics to midi. I never had Problems with it. If youn have Midiproblems you have A Midiinterface that is not as good as it can be.. to test a good midipipline they use allways the Nordmodular... i had a Nordmodular and never Problems, cause i had a good Midiconnection but there are Interfaces out there which can not handle Mididata very good... Midi works. You can easlie integrate it in every modern workstation... i never had Problems all works fine for me..., but please no Sysex... Sex yes, but without sy....

11-Feb-14 06:04 AM


S R Dhain    Said...

A great post and understandably a tech-flavoured one. It's remarkable how much data is spat out via midi and even more remarkable that for over 30 years ( just about), that hardware and software sequencers can mostly cope with the inflow/outflow. I used to use a phillip rees midi merge and a V3 midi thru box in my old studio setup and i NEVER had any crashes, even with loads of knob and slider movements being recorded in from multiple devices and obviously being played back. Now everything has the processing power, so there really isnt any excuse for clog, especially when a lot of it is in the box. The same rule still applies; you're only as fast as your slowest component, which ironically i don't think is midi itself, but the pc's and macs which are sometimes overloaded with stacks of plugs. To my knowledge there is no realtime way to monitor and auto adjust buffer sizes for midi throughput, which is necessary for the more complex stuff that has to be nested within a more layered software environment. It's akin to VM'ing within a VM, which is really heavy duty stuff. Kudos to a lot of the softsynth and DAW manufacturers for having to plough through as many combinations & permutations of stuff to get it ship shape from out of the box and in the studio, as Sysex and indeed software programming is a real game of multi level chess.

11-Feb-14 06:42 AM


mattsynth    Said...

The great part about midi is that it works. That is why it has lasted so long.

11-Feb-14 01:41 PM


MIDIot    Said...

MIDI DOES work -- and it works VERY well -- if you are using a COMPUTER DAW, and not a stand-alone MIDI "workstation."

I used to be one of those die-hard "hardware is better" guys. I gave it up and am much happier with my PC and a bunch of software.

Anyone who complains about the speed of MIDI is using a DIN-plug system, and is not using USB 3.0, an SSD, and a core i or AMD Bulldozer processor.

Get with the times folks! Get rid of that weighty malfunctioning beast that you call a music production keyboard, and get thyself to Best Buy PRONTO.

On a related note -- I agree whole-heartedly that SysEx is for the birds. I once created an entire album's worth of music -- about 41 minutes altogether -- nine out of ten songs used RPN, NRPN, and SysEx messages to control effects and filters on my Roland MC-80.

That was not fun.

But I learned something --

I HATE SYSEX.

11-Feb-14 05:19 PM


Saxifraga    Said...

@to-pse Said... "Actually MIDI is still able to cope with high volumes as long as one uses other transports than the DIN serial ports like USB, Firewire or Ethernet."

That´s incorrect. MIDI devices only accept the data with 31250 bps. The carrier may be much faster but the data is just read with that old transfer rate.

So we now have hardware synth that have modern CPUs and hidden gigabit connections (Kronos Atom board) but get controled by the slowest possible protocol.

It´s obvious that something has to be done. But as long as all manufacturers sell us that old tech and people applaude them, we will not get good modern solutions. Just look how ridiculous Kronos and Fantom G handle sample data. You have to transfer samples and patch banks via usb device or internal ftp server (via usb to RJ45 adaptor). And the Fantom G takes ages to load samples from USB stick like the Korg M3 does. This is retro. No one accepts this other than musicians. Yamaha is even more lame but has flash ram to hide it´s incontinence.

11-Feb-14 05:20 PM


Saxifraga    Said...

@to-pse Said... "Actually MIDI is still able to cope with high volumes as long as one uses other transports than the DIN serial ports like USB, Firewire or Ethernet."

That´s incorrect. MIDI devices only accept the data with 31250 bps. The carrier may be much faster but the data is just read with that old transfer rate.

So we now have hardware synth that have modern CPUs and hidden gigabit connections (Kronos Atom board) but get controled by the slowest possible protocol.

It´s obvious that something has to be done. But as long as all manufacturers sell us that old tech and people applaude them, we will not get good modern solutions. Just look how ridiculous Kronos and Fantom G handle sample data. You have to transfer samples and patch banks via usb device or internal ftp server (via usb to RJ45 adaptor). And the Fantom G takes ages to load samples from USB stick like the Korg M3 does. This is retro. No one accepts this other than musicians. Yamaha is even more lame but has flash ram to hide it´s incontinence.

11-Feb-14 05:20 PM


Saxifraga    Said...

@to-pse said... "Actually MIDI is still able to cope with high volumes as long as one uses other transports than the DIN serial ports like USB, Firewire or Ethernet."

That´s incorrect. MIDI devices only accept the data with 31250 bps. The carrier may be much faster but the data is just read with that old transfer rate.

So we now have hardware synth that have modern CPUs and hidden gigabit connections (Kronos Atom board) but get controled by the slowest possible protocol.

It´s obvious that something has to be done. But as long as all manufacturers sell us that old tech and people applaude them, we will not get good modern solutions. Just look how ridiculous Kronos and Fantom G handle sample data. You have to transfer samples and patch banks via usb device or internal ftp server (via usb to RJ45 adaptor). And the Fantom G takes ages to load samples from USB stick like the Korg M3 does. This is retro. No one accepts this other than musicians. Yamaha is even more lame but has flash ram to hide it´s incontinence.

11-Feb-14 05:21 PM


Saxifraga    Said...

Sorry for tripple posting. Please erase two of them.

11-Feb-14 05:22 PM


Post a comment 
 


More Videos

New Korg Electribe Product Preview 

We get Video with James Pullen AKA Mista Bishi


Sonic LAB: UAD2 Thermionic Culture Vulture 

All tube studio overdrive emulation for DSP


Cubase Focus: Using The Chord Track To Drive The Arrangement 

On the fly arrangement of chords and melody


Waldorf Streichfett String Machine Review 

Sounds of an era of bygone technology