How has MIDI changed live electronic music production and performance?

Hey DJTTs,

This is the topic for an essay for my engineering degree and i would like to get opinions and brainstorming from you guys! The community which is responsible for much of the innovation and boundary pushing!

Some interesting things I've noticed, as many of you may have as well, is that MIDI is now all about customisation when it comes to live performance, and that especially with digital djing software, MIDI has been used to turn music into an expressive musical instrument in its own right.

I think we are living in a time of unrivalled creativity because of MIDI as it has made music production accessible to the general public, and MIDI customisation has made it possible for every performer to be unique in their delivery even with the exact same equipment.

There are so many things that MIDI has done for us and will continue to do in the future.

What do you guys think?

It made you be able to dj and imitate mixer and other controls with a midi device through mapping for example. You can DJ without a traditional mixer+tt, cdj or whatever jog wheel setup now.

MIDI on the whole also changed vast aspects of sound production through use of DAW's and soft-synths. MIDI is/was designed around production. Even now you're seeing controls switching over to HID or other formats, because MIDI is limited in the end to what it can do.

I think the production aspect of MIDI is much more important than anything anyone has done otherwise. It's not really MIDI that you are talking about in your post, but the ability to have unique controllers that talk back and forth with the computer. In the past it's used the MIDI protocol, but we are on the edge of moving away from it. Because MIDI wasn't designed to handle the sheer amount of data we are throwing at it these days, and can get bottlenecked (especially with platters).

Quote:

---

Originally Posted by Bassline Brine

MIDI on the whole also changed vast aspects of sound production through use of DAW's and soft-synths. MIDI is/was designed around production. Even now you're seeing controls switching over to HID or other formats, because MIDI is limited in the end to what it can do.

I think the production aspect of MIDI is much more important than anything anyone has done otherwise. It's not really MIDI that you are talking about in your post, but the ability to have unique controllers that talk back and forth with the computer. In the past it's used the MIDI protocol, but we are on the edge of moving away from it. Because MIDI wasn't designed to handle the sheer amount of data we are throwing at it these days, and can get bottlenecked (especially with platters).

---

I definitely agree that it has had much more impact on music production, but for my assignment i get to look at one fairly specific area of music production and go into depth about about it, And Im more interested in the use of MIDI in other ways than its originally intended.

I'm interested in what your saying about HID though, In your opinion, do you see a shift from MIDI to HID or other protocols? Not meaning to flame or anything, but i would imagine something would be made to handle the bulk data bottleneck would be more likely, as much of its appeal is the fact that MIDI instruments and controllers share a standard protocol which allows for communication between devices. I know that NI controllers run on HID and it does offer better resolution and integration with the software, but thats only with that specific software, and i would imagine an even bigger data bottleneck as higher resolution means more 0s and 1s. Stop me if im wrong though.... Im not trying to flame, just discuss, it seems alot of other protocols have come and go in the past and MIDI is still at version 1.0.

Quote:

---

Originally Posted by Bassline Brine

MIDI on the whole also changed vast aspects of sound production through use of DAW's and soft-synths... ...Because MIDI wasn't designed to handle the sheer amount of data we are throwing at it these days, and can get bottlenecked (especially with platters).

---

None of this is actually true. You can certainly say with confidence that the people who originally developed MIDI probably never imagined it being used to allow communication with DJ gear - but it's also true that the creators of the first computers never imagined them in the role of production studio, either.

As it stands right now, the current MIDI messaging that is accepted as standard has more than enough resolution and throughput to satisfy the needs of any DJ. There are certainly reasons for developers to adopt different protocols, but for reasons OTHER than the ones you imply.

To the OP, I'm not sure I would frame the essay in that way. You seem to imply that this is a new phenomenon, when in truth it was what MIDI was DESIGNED to do back in the early 80s. The barrier wasn't that a suitable protocol didn't exist, but rather that the cost of entry was high enough that it was out of reach for most people. Instead, a more accurate way to frame the discussion would be that as computing power becomes commoditized, more hardware and software has been developed to leverage that newly accessible power.

The video production industry experienced the same phenomenon in the early 90s. It used to be VERY expensive to do what would be professional video production, with A/B roll edits, transition effects, keys, text generation, etc. A beta deck alone would cost you close to $30K. But a company called NewTek developed a system called the Video Toaster, which was a relatively cheap bit of hardware, and some software that ran on an Amiga. All of a sudden, production quality exploded and that single product opened up professional video production to anyone who could cough up $10-15K.

And there wasn't a perception that this new system produced inferior quality product. Indeed because it was digital in nature, it allowed people to be more creative and productive while also producing product that would have better video quality than if it was edited in a traditional A/B roll edit suite.

The original MIDI messaging protocol was never properly designed for high resolution controllers like smooth channel volume controllers. Not that it's been hacked in now with HID. Anyway me thinks it was a good example of a design that was good and is still good for note information but the designers didn't think of scalability -- no wonder as they were not network engineers by trade. Or 127 channel limitation.

Quote:

---

Originally Posted by ksandvik

The original MIDI messaging protocol was never properly designed for high resolution controllers like smooth channel volume controllers. Not that it's been hacked in now with HID. Anyway me thinks it was a good example of a design that was good and is still good for note information but the designers didn't think of scalability -- no wonder as they were not network engineers by trade. Or 127 channel limitation.

---

Again, totally untrue. In fact, one of the first MIDI messages defined was a 14bit message (the pitch bend message, with WAY MORE messages than what would be needed to provide a 100mm fader with say 4 messages per millimeter of travel.). In the General MIDI spec, more messages were added that fit the bill.

And HID isn't a "hack" to MIDI. Not even close. HID is another protocol entirely.

Quote:

---

Originally Posted by nem0nic

To the OP, I'm not sure I would frame the essay in that way. You seem to imply that this is a new phenomenon, when in truth it was what MIDI was DESIGNED to do back in the early 80s. The barrier wasn't that a suitable protocol didn't exist, but rather that the cost of entry was high enough that it was out of reach for most people. Instead, a more accurate way to frame the discussion would be that as computing power becomes commoditized, more hardware and software has been developed to leverage that newly accessible power.

---

Although I completely agree, The topic for the assignment is actually how electronics have changed music production, and with that we are expected to choose something electronic related to music production and discuss how that piece of equipment has changed music production, be it in the studio or the live arena, I've decided to focus on MIDI controllers for live electronic music performance to be specific about, Mostly because of my own personal interest. I would have loved to do a paper on how the personal computer has changed music production but you could write a book on that and I've only got 2000 words allowed.

Also history of MIDI will be a part of the essay so I definitely wont be implying that its by any means a new development. MIDI definitely had its biggest impact upon inception and standardisation in the music industry, and once computer speeds picked up more of the MIDI protocol became more usable in real applications, unlike the days when the piano roll on your MIDI sequencer used to max out your computers processing before graphics cards improved dramatically in personal computers. And although it was designed to interface physical controls with software and hardware controls with software controls, i think were missing the point, how has it changed how people produce and perform live be it djing or whatever.

Quote:

---

Originally Posted by nem0nic

Again, totally untrue. In fact, one of the first MIDI messages defined was a 14bit message (the pitch bend message, with WAY MORE messages than what would be needed to provide a 100mm fader with say 4 messages per millimeter of travel.). In the General MIDI spec, more messages were added that fit the bill.

And HID isn't a "hack" to MIDI. Not even close. HID is another protocol entirely.

---

I'm sure you like the latency with longer messages due to the serial nature of MIDI messages.... Fortunately inside computers this is not an issue. However with MIDI running over USB and multiple channels operating at the same time...

Usually faders are using control change messages but MIDI is being very much hacked and stretched so maybe some devices are using pitch midi messages to get better resolution. To notice Traktor S4 just bypassed it all and uses their own high-speed low-latency control to get good performance numbers.

I assumed HID was running inside MIDI to handle the traffic via USB but I could be wrong. Anyway, if the MIDI spec engineers would have done a little bit more thinking they would have had HID like properties as part of their spec.

I still think that MIDI was accepted the way it was accepted because it was the only industry standard protocol. But it was really designed for having synthesizers speak to other synthesizers, and with the electrical and protocol latencies and max 16 channels that also had limitations *). It was definitely not designed as a network protocol, for example embedding the sender and receiver info instead of hard coded max 16 channels (ugh.) Anyway, I had fun with the protocol back in 1983 as I got my first engineering job from college thanks to hacking MIDI.

Now, if the MIDI designers would have known something about the Ethernet protocol... Things would have been very different. Or, as I might say, I salute to all the second and future generation MIDI engineers who figured out ways to go around all the shortcomings of the 1.0 spec.

*) If you used a MIDI sequencer like the LinnDrum one and fed let's say 3 or more synthesizers at the same time for tape recording, it was not fun hearing all the latencies recorded to tape...

Quote:

---

I'm sure you like the latency with longer messages due to the serial nature of MIDI messages...

---

But that is an issue with the INTERFACE - not the PROTOCOL. The great thing about MIDI is that it can work at the full speed of the transport interface it's using. So if you're using MIDI over a serial interface, then you may experience latency (5 to 10ms roughly, although this can easily be optimized for better performance). If however you're using MIDI over USB, then you're measuring latency in picoseconds.

Also, MIDI messages are smaller than their typical HID counterpart. This means MIDI can actually be faster than HID over the SAME transport interface.

So no, what you're trying to imply here is incorrect.

Quote:

---

Usually faders are using control change messages but MIDI is being very much hacked and stretched so maybe some devices are using pitch midi messages to get better resolution

---

You imply a hack. Stop. NOTHING is being hacked. A pitch bend message IS A STANDARD CC MESSAGE. It is simply a 14bit message versus a 7bit message. There are other 14bit messages available to accomplish the same thing. And we're only talking about absolute messages right now.

Quote:

---

I assumed HID was running inside MIDI to handle the traffic via USB but I could be wrong. Anyway, if the MIDI spec engineers would have done a little bit more thinking they would have had HID like properties as part of their spec.

---

You ARE wrong. And as far as adding HID like properties to MIDI goes, they DID think of that. It's called SYSEX, and it's been a part of MIDI for about 20 years.

Quote:

---

...and with the electrical and protocol latencies and max 16 channels that also had limitations *)

---

Again, you fail to recognize the difference between the protocol and the antiquated transport you keep associating it with.

Quote:

---

It was definitely not designed as a network protocol...

---

Although it's completely possible to do MIDI over ethernet, I understand your point. But again we run into the overhead issue. If we (for instance) were using TCP/IP instead, there would be additional latency because of the increased overhead decoding the larger packets. No doubt it would be preferable to go to a faster protocol (like UDP), but you would lose goodies like error handling.

Besides, why re-invent the wheel when a MIDI over ethernet protocol already exists in RTP MIDI?

Quote:

---

Anyway, I had fun with the protocol back in 1983 as I got my first engineering job from college thanks to hacking MIDI.

---

It sounds like this was the last time you looked at MIDI.

Quote:

---

Now, if the MIDI designers would have known something about the Ethernet protocol... Things would have been very different. Or, as I might say, I salute to all the second and future generation MIDI engineers who figured out ways to go around all the shortcomings of the 1.0 spec.

---

There was no reason to make MIDI more complicated when it was developed in the early 80s. The initial offering was very well suited to it's application. Less than 10 years later, it was modified to add a whole new slew of features. And MIDI continues to evolve year after year.

It's not like the people who developed TCP/IP were any different (as you seem to imply). Only a 32 bit address space for literally BILLIONS of users? Whoops. Good thing the protocol continues to evolve (just like MIDI) and modifications (like IPv6) are developed to address new needs.

Quote:

---

If you used a MIDI sequencer like the LinnDrum one and fed let's say 3 or more synthesizers at the same time for tape recording, it was not fun hearing all the latencies recorded to tape...

---

Using all those 5 pin DIN connections was your problem - NOT MIDI. If they had been connected with a different transport (literally ANY more modern serial transport) you wouldn't have had the problem.