Jump to content


Please note: You can easily log in to MPN using your Facebook account!

MIDI is so last century!


ITGITC

Recommended Posts

MIDI is so last century! (MIDI protocol was defined in 1982.)

 

And USB 3.0 is already looking pretty slow (and it's not even commonplace).

 

Here is the take-away: Intel's Light Peak technology.

 

Latency? Pshaw!

 

MIDI's transfer rate is 0.00003125 Gbps. (Gigabits per second) And it takes TWO cables; one up, one down!

 

USB 2.0 is 0.48 Gbps (theoretically).

 

Firewire 400 (IEEE 1394) is 0.4 Gbps.

 

Firewire 800 (IEEE 1394b) is 3.2 Gbps.

 

USB 3.0 is up to 4.8 Gbps.

 

Intel's LightPeak starts at 10Gbps and will scale to 100Gbps. And it can do this speed up and down the line simultaneously.

 

Yeah.

So what does this mean for us electronic keyboard musos, you ask? :snax:

 

It means that in the next ten years (by the time the keyboard manufacturers begin to consider implementing the technology),

 

DRUMMERS WILL STILL BE STEALING OUR GIRLFRIENDS (while we're stuck in our home studio trying to make it all work out). :rolleyes:

 

[video:youtube]http://www.youtube.com/watch?v=izNoF1SWtSg

 

I just thought you guys would want to know.

 

Tom :cool:

 

 

"Music expresses that which cannot be put into words and that which cannot remain silent." - Victor Hugo
Link to comment
Share on other sites



  • Replies 35
  • Created
  • Last Reply

Speed?

Speed is a measurement of distance / time.

 

Communication protocols are measured in amount of data / time. That is not speed, but a Transfer Rate.

 

If you are talking about the actual Transfer Rate of MIDI info, it is indeed small, compared to whatever is available nowadays.

 

However, to compare them is also incorrect: the physical MIDI port could have had a small bandwidth, but MIDI data is now traveling through those modern communication ports, such like USB 1.1, USB 2.0, Firewire, etc. The same specification from 1983 (plus all the add-ons) as a protocol. MIDI data is so small it makes no sense to think about a much bigger bandwidth port to move it.

 

 

Músico, Productor, Ingeniero, Tecnólogo

Senior Product Manager, América Latina y Caribe - PreSonus

at Fender Musical Instruments Company

 

Instagram: guslozada

Facebook: Lozada - Música y Tecnología

 

www.guslozada.com

Link to comment
Share on other sites

Speed?

Speed is a measurement of distance / time.

 

Communication protocols are measured in amount of data / time. That is not speed, but a Transfer Rate.

 

If you are talking about the actual Transfer Rate of MIDI info, it is indeed small, compared to whatever is available nowadays.

 

However, to compare them is also incorrect: the physical MIDI port could have had a small bandwidth, but MIDI data is now traveling through those modern communication ports, such like USB 1.1, USB 2.0, Firewire, etc. The same specification from 1983 (plus all the add-ons) as a protocol. MIDI data is so small it makes no sense to think about a much bigger bandwidth port to move it.

 

 

 

E A S Y there big fella. :laugh:

 

This issue relates to latencies throughout the system and how they affect performance. Faster transfer rates can improve many things, from the feel and user experience of a certain keyboard, to the length of time a sample takes to load in your Receptor. Newer technologies are reducing these latencies, thus opening up new possibilities that make for better electronic instruments.

 

So give a fellow forumite a break, Gus. A fresh discussion centered around technological gains trumps yet another thread about how to guarantee that you get the recognition you deserve when performing showtunes in mono in front your stepmother's family after the big Thanksgiving dinner. :cool:

 

Tom

 

"Music expresses that which cannot be put into words and that which cannot remain silent." - Victor Hugo
Link to comment
Share on other sites

... still, saying "speed" is incorrect :P and my stepmother enjoys STEREO performances. Mktnksbye.

Músico, Productor, Ingeniero, Tecnólogo

Senior Product Manager, América Latina y Caribe - PreSonus

at Fender Musical Instruments Company

 

Instagram: guslozada

Facebook: Lozada - Música y Tecnología

 

www.guslozada.com

Link to comment
Share on other sites

I would like to see better implementation of continuous controllers and "per note" data in midi ...if the spec is improved. 128 steps do step, and using pitch bend or NRPN's is a workaround.

 

However, I am grateful for what we have at this point.

 

I wonder if a new spec needs to developed for wireless/bluetooth communication with mobile devices ... or will new transmission protocols simply transmit existing forms of midi data?

 

Link to comment
Share on other sites

I wonder if a new spec needs to developed for wireless/bluetooth communication with mobile devices ... or will new transmission protocols simply transmit existing forms of midi data?

 

Hi Tusker!

 

At this moment, MIDI is also being transmitted via WiFi and Bluetooth; at least, that's what I'm doing right now with my iPad and iPhone, controlling soft synths and Pro Tools.

 

I'd like to think it will keep being the same given I don't feel it is necessary to create yet another physical port for a new version of MIDI.

Músico, Productor, Ingeniero, Tecnólogo

Senior Product Manager, América Latina y Caribe - PreSonus

at Fender Musical Instruments Company

 

Instagram: guslozada

Facebook: Lozada - Música y Tecnología

 

www.guslozada.com

Link to comment
Share on other sites

It's funny to see this posting, because I had decided last night NOT to post anything on this forum about AVB, which has made some sudden breakthroughs and which now has major players all across several industries hopping on board (including Yamaha, who joined the AVnu Alliancce DURING the AES Convention this weekend!).

 

I haven't had time to read up on Light Peak these past few days to see whether it is a consumer version of AVB or has any relationship to it, but Intel is now one of the two major players pushing AVB forward (my company is one of the founding members of the AVnu Alliance as well).

 

With AVB on old Ethernet frameworks, probably maximum 2 ms latency with high track counts (32+). Down to 0.2 ms or less, with new Ethernet bandwidth.

 

I expect AVB to replace MADI as the protocol of choice fairly soon, even though it has taken this long for MADI to start getting more widely adopted (including by Avid, who surprised everyone by going all open standards, either at the AES Convention this past weekend, or right beforehand, via PT 9).

 

It will also filter down to consumer and prosumer space fairly quickly, due to the players involved, the extremely low cost, and scalability.

 

As for MIDI, that is a data protocol that often piggy-backs atop existing DIN cable standards. MIDI can shoot down Ethernet as well.

 

MIDI is flawed due to low resolution, more than by speed issues. It works OK for most pop/rock, but not so well for the kinds of articulations common in classical music.

 

It takes a lot of work, but you can do it. Notion Music chose to take a different approach with their notation package, giving them 1000 levels of resolution for all parameters.

Eugenio Upright, 60th Anniversary P-Bass, USA Geddy Lee J-Bass, Yamaha BBP35, D'angelico SS Bari, EXL1,

Select Strat, 70th Anniversary Esquire, LP 57, Eastman T486, T64, Ibanez PM2, Hammond XK4, Moog Voyager

Link to comment
Share on other sites

And of course there's always OSC, which we use at my company but which never got widely adopted as a replacement for MIDI protocol.

 

The Lemur uses it. It's a completely different philosophy from MIDI. I haven't personally worked with it but will soon, so will at that point probably form an opinion of its merits vs. MIDI.

Eugenio Upright, 60th Anniversary P-Bass, USA Geddy Lee J-Bass, Yamaha BBP35, D'angelico SS Bari, EXL1,

Select Strat, 70th Anniversary Esquire, LP 57, Eastman T486, T64, Ibanez PM2, Hammond XK4, Moog Voyager

Link to comment
Share on other sites

I just read this morning that the next version of iOS 4 (Mac) will support MIDI, so not everyone thinks it's dead just yet.

Will our next round of synths have an iPod/iPhone dock on them? You can bet your sweet iBippy!

Botch

In Wine there is Wisdom

In Beer there is Freedom

In Water there is bacteria

Link to comment
Share on other sites

It's funny to see this posting, because I had decided last night NOT to post anything on this forum about AVB, which has made some sudden breakthroughs and which now has major players all across several industries hopping on board (including Yamaha, who joined the AVnu Alliancce DURING the AES Convention this weekend!).

 

I haven't had time to read up on Light Peak these past few days to see whether it is a consumer version of AVB or has any relationship to it, but Intel is now one of the two major players pushing AVB forward (my company is one of the founding members of the AVnu Alliance as well).

 

With AVB on old Ethernet frameworks, probably maximum 2 ms latency with high track counts (32+). Down to 0.2 ms or less, with new Ethernet bandwidth.

 

I expect AVB to replace MADI as the protocol of choice fairly soon, even though it has taken this long for MADI to start getting more widely adopted (including by Avid, who surprised everyone by going all open standards, either at the AES Convention this past weekend, or right beforehand, via PT 9).

 

It will also filter down to consumer and prosumer space fairly quickly, due to the players involved, the extremely low cost, and scalability.

 

As for MIDI, that is a data protocol that often piggy-backs atop existing DIN cable standards. MIDI can shoot down Ethernet as well.

 

MIDI is flawed due to low resolution, more than by speed issues. It works OK for most pop/rock, but not so well for the kinds of articulations common in classical music.

 

It takes a lot of work, but you can do it. Notion Music chose to take a different approach with their notation package, giving them 1000 levels of resolution for all parameters.

So, since you're not posting about AVB, you're not even going to tell us what the heck it is and let us svengle it?

Next time you're not posting something, don't. Or, if you do, do. :-P

Link to comment
Share on other sites

The problems with MIDI aren't the rates, as mentioned above, except when you're using MIDI cables.

 

And we're still using MIDI cables between our keyboards, and often between our keyboards and our computers (at least to the MIDI interface). In that case, the data rate can be an issue, since it adds a 1ms per note latency when playing a chord. I never notice it since I never play the notes all at the same time anyway, but no doubt world-class percussionists would.

 

As mentioned above, the limit of 128 or 256 steps in most parameters is a severe limitation. There are workarounds for this with extensions to the MIDI architecture, but yeah, it would be far better (theoretically) to use 16-bit values for all parameters.

 

As always, it's painful to break from the simplifications and limitations of the past and move on to new ground, due to the importance of backwards compatibility. So we usually end up with a complicated compromise which drives us all crazy with the caveats.

 

That's life in the tech fast lane. Even if we're still driving at 1980 speeds. :-))

Link to comment
Share on other sites

As mentioned above, the limit of 128 or 256 steps in most parameters is a severe limitation. There are workarounds for this with extensions to the MIDI architecture, but yeah, it would be far better (theoretically) to use 16-bit values for all parameters.

Good example would be High Resolution Velocity CC#88, which allows (how many bits are 16,000+ again?) finer control of velocity- it is implemented in the VAX77, but I have yet to find anything that responds to it. I don't think implementing it would break backward compatibility, but it would be nice to see MIDI cables go away. As it stands now, nearly my entire studio is wired via USB, and my 8x8 MIDI interface has exactly one thing plugged into it!

 

 

Link to comment
Share on other sites

great. something faster. BIG DEAL.

 

usb3 and an updated driver will do THAT trick.

will anybody notice? i doubt it.

 

If you do want to bite the bullet and improve the midi format, OH BOY are you in for some standards fun.

 

Just get every music hardware manufacturer and sequencer manufacturer to ALL AGREE.

 

Good luck with your .mid2 files...

...Steve (writing my own midi sequencer - yeah, i'm crazy like that - http://pianocheetah.com)
Link to comment
Share on other sites

Ok, if it's going to be a royal pain to upgrade, we may as well get bang for the buck. What would it take to have a live/studio rig without cables? (seriously)

 

There's got to be some part of the electro-magnetic spectrum that's friendly to musical devices right?

 

 

Link to comment
Share on other sites

This thread is both interesting, and in some ways laughable. Let me explain...

 

In my day job in Industrial Instrumentation and Automation, we have very old, very slow digital protocols that have been around since the early 90's that are still the standard and widely used. There are also newer digital protocols (including industrial ethernet) that offer greater bandwidth and more features. Thing is, it's application specific. If you're going to control your manufacturing process with the digital protocol, then you need speed and bandwidth (note, I list both because they each have different implications in a control environment :wave:). But if you're still controlling off the analog signal and just need to interrogate field devices for maintenance/configuration, or bring back limited data, the old protocol is fine and is widely available, standard, inexpensive, etc.

 

MIDI is pretty much the same thing. If you're using it as it was designed - basic control - it's fine... plus it's standard, widely available, etc. People tend to not use MIDI the way it was intended anymore. People have gone more to audio transfer and more bandwidth intensive data transfers that require different protocols. But if you're controlling modules, sending patch changes, user manipulated controller data, and note data - you don't need more than MIDI.

 

So really, I'd say it's application dependant. Is there an application for a faster, wider bandwidth protocol? Sure. But that doesn't mean MIDI doesn't have its place.

 

Dan

 

Acoustic/Electric stringed instruments ranging from 4 to 230 strings, hammered, picked, fingered, slapped, and plucked. Analog and Digital Electronic instruments, reeds, and throat/mouth.

Link to comment
Share on other sites

Looking back, perhaps the title of this thread is misleading.

 

MIDI is what it is. However, as you can see from my links and the video, general latencies within the computers we use are being minimized as technology progresses. It's exciting to see - and this is motivation that led me to start this topic. Advances in storage and communications links between devices are ramping up faster than I've ever seen. It's not a criticism to say that it will take some time for new technology to be designed into our musical tools. That's a given.

 

It's interesting to watch the technological advancements in the computer industry, and it's fun to envision how these advancements will be used to make better tools for us. I truly get excited about all this, and I started this topic to share that excitement with you and encourage conversation in this regard.

 

Thanks to those who contribute to this thread.

 

Tom

 

 

 

 

"Music expresses that which cannot be put into words and that which cannot remain silent." - Victor Hugo
Link to comment
Share on other sites

Let me sort of clarify - I agree, the technology is exciting and the applications it can open up are fun to think about. My only point is that I don't even think of MIDI when considering what you can do with that bandwidth - the applications are not even in the same ballpark, that's all I was saying.

 

What would be more exciting to me, and so far I'm not aware that the technology exists, would be reliable wireless multichannel audio streaming. Imagine, no audio cables anywhere. Set each channel of the mixer to a wireless channel. Set each instrument to the corresponding wireless channel. Set the output bus of the mixer to the same channel as the input to the amp rack. Wirelessly route mixer sends to IEM's, powered monitors, or outboard effects. No snakes, no wires... imagine what that would do for setup time and ease.

Dan

 

Acoustic/Electric stringed instruments ranging from 4 to 230 strings, hammered, picked, fingered, slapped, and plucked. Analog and Digital Electronic instruments, reeds, and throat/mouth.

Link to comment
Share on other sites

Where would we be today if Dave Smith had not created midi?

...stacks of keyboards and no midi sound modules.

"It is a danger to create something and risk rejection. It is a greater danger to create nothing and allow mediocrity to rule."

"You owe it to us all to get on with what you're good at." W.H. Auden

 

Link to comment
Share on other sites

The closest I've seen is wireless speakers that use live music streaming or radio transmission. While it works for reproducing sound from a remote server. Where your not hoping to sync with other transmissions. In a recording/playing environment the latency would be huge. TCP/IP isn't designed for real time application. The time for a packet to be sent, received and decoded can be worse than a midi cable. Wireless transmission via standard Uhf or Vhf carrier works great for live performance like wireless mics but they really aren't reliable enough when your talking multiple channels. And remember that wireless mic isn't producing CD quality dynamic range and frequency response. Its closer to what your home stereo does with an FM transmission. Imagine a 16 piece band were everyone is using wireless to the PA. What's the chance one channel is going to bleed into another? Or Ham Radio suddenly comes over one channel.

If the source is digital or converted to digital immediately, then I can envision a single optic cable going from a hub into the mixer. But totally wireless transmission for everything is unrealistic in my opinion.

Boards: Kurzweil SP-6, Roland FA-08, VR-09, DeepMind 12

Modules: Korg Radias, Roland D-05, Bk7-m & Sonic Cell

Link to comment
Share on other sites

great. something faster. BIG DEAL.

 

usb3 and an updated driver will do THAT trick.

This applies only to USB keyboards connected to a computer. It does not apply to connecting keyboards to modules or MIDI (non-USB) keyboards to other gear.

If you do want to bite the bullet and improve the midi format, OH BOY are you in for some standards fun. Just get every music hardware manufacturer and sequencer manufacturer to ALL AGREE.
Yeah, there's that! Not to mention dealing with compatibility when you buy new gear.

Good luck with your .mid2 files...
Totally no problem. We don't transmit files over MIDI cables anyway: software interprets the format and sends MIDI data. Any sequencer supporting a new language should be able to handle MIDI and translate as needed. Of all the problems, this is the easiest to solve.
Link to comment
Share on other sites

I look forward to the day when I can configure my gear at home and at a gig, just plug it into power and it assembles itself. I suspect that wireless ethernet is the best bet there; not because it's the ideal technology for the job, but because it's ubiquitous, fast, and cheap. (If only pretty women were .... oops moving along now.)

 

The biggest problem with wireless ethernet is avoiding conflicts, like the kind you'd get if lots of people brought laptops and each group set up their own ad-hoc net. I suspect that's solvable, but it poses a limit on the number of systems that can set up at once independently. (Not sure what the limit is; how many channels are there?)

 

TCP isn't a problem, because it's not the right tool for the job. IP isn't a problem; it can handle real-time protocols.

 

The biggest issue with a packet-oriented protocol is the need to buffer audio before transmitting and after receiving. The biggest impediment to this is the typical source of latency in a computer audio system: interrupts. Regardless, I'm confident that less than 10 ms end-to-end latency is very doable. That's tolerable, given that sound travels about 1 foot per ms, and we can handle a stage where the guitarist is 15 feet away.

 

Transmission delays are only a significant issue where long distances are involved. Don't plan an internet jam with anyone more than 1000 miles away. In that case you get twice the latency, because it's round-trip latency that counts, not one-way. Of course, that's also true of those 15 feet separating musicians on the stage.

 

Actually, move that guitarist farther away!

Link to comment
Share on other sites

I'm confident that less than 10 ms end-to-end latency is very doable. That's tolerable, given that sound travels about 1 foot per ms, and we can handle a stage where the guitarist is 15 feet away.

 

I agree that a 10 ms delay is not terrible sonically, but I suspect that a 10 ms delay between when you play a key and when you hear the sound may result in a less satisfying playing experience that will impede the sense of "connection" between player and instrument.

 

OTOH, it might make better players out of those who tend to play ahead of the beat. ;-)

Maybe this is the best place for a shameless plug! Our now not-so-new new video at https://youtu.be/3ZRC3b4p4EI is a 40 minute adaptation of T. S. Eliot's "Prufrock" - check it out! And hopefully I'll have something new here this year. ;-)

Link to comment
Share on other sites

I agree that a 10 ms delay is not terrible sonically, but I suspect that a 10 ms delay between when you play a key and when you hear the sound may result in a less satisfying playing experience that will impede the sense of "connection" between player and instrument.
Does it bother you if your monitor speaker is 11 feet away from your ear?

 

Actually, I'm confident 5 ms is achievable, and sufficient for most purposes.

 

While 10 ms latency by itself isn't bad, if you add another 10 ms for a speaker 11 feet away, you get to 20 ms, which is definitely noticeable and annoying when playing any percussive sound. I can tolerate it, but then I'm a hack.

 

10 ms is easily doable even on a low-end system, which is why I used that number in the first place. Well-tuned VSTi hosts can show 3 or 2 ms latency, but that's output only -- no input buffering necessary. Add a ms or so for transmission delays, but those get lower as CPU speeds and bandwidth increase.

 

I really do believe this is doable with today's technology, but the development costs would exceed the market value. I think Yamaha might have gotten a start on it with their M-LAN, but that didn't go anywhere as far as I can tell.

Link to comment
Share on other sites

I can tolerate it, but then I'm a hack.

 

:D

 

I call BULL!

 

You're no hack, Jeff. I've heard ya play. :thu:

 

Thanks for adding your wisdom to my semi-krazi thread. (It was BAIT for ya, don'tchaknow?) :)

 

Tom

 

"Music expresses that which cannot be put into words and that which cannot remain silent." - Victor Hugo
Link to comment
Share on other sites

Hey everyone,

I'm not that up to speed on a lot of the standards here. BUT, I'm very curious as regards practical applications with computers and muse receptors running vsts.

 

Do any of the technologies affect how low the latency could be made? As I understand it, what separates dedicated keyboards, with their own custom sound chips, from computer generated sources, is the latency.

 

I assume a muse receptor is best case scenario. So assuming a SSD, is 2-3 ms the lowest latency that's possible? Does that increase dramatically as you put a heavy load on it? Would any of these faster pipeline technologies improve on that?

Numa Piano X73 /// Kawai ES920 /// Casio CT-X5000 /// Yamaha EW425

Yamaha Melodica and Alto Recorder

QSC K8.2 // JBL Eon One Compact // Soundcore Motion Boom Plus 

Win10 laptop i7 8GB // iPad Pro 9.7" 32GB

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...