Go Back   Cockos Incorporated Forums > REAPER Forums > REAPER General Discussion Forum

Reply
 
Thread Tools Display Modes
Old 03-03-2011, 02:28 AM   #1
PAPT
Human being with feelings
 
Join Date: Jun 2006
Posts: 1,082
Default Does MIDI tranfer better through a MIDI in than through USB?

My midi keyboard has both midi and usb outs.

I have been using the usb because I currently do not have a soundcard with midi input.

Would there be any benefit to having a midi in on a soundcard to use instead of usb?
Does data transmit better over a pure midi path, or is usb just as good.

I am looking to get a soundcard and would like to know if having a midi input should be a consideration.
PAPT is offline   Reply With Quote
Old 03-03-2011, 03:30 AM   #2
EvilDragon
Human being with feelings
 
EvilDragon's Avatar
 
Join Date: Jun 2009
Location: Croatia
Posts: 24,812
Default

In addition to the original 31.25 kbits/sec (baud is the signalling rate and is the reciprocal of the shortest signalling element; bits/sec is the data rate) current-loop transported on 5-pin DIN, other connectors have been used for the same electrical data, and transmission of MIDI streams in different forms over USB, IEEE 1394 a.k.a FireWire, and Ethernet is now common (see below).

USB

A standard for MIDI over USB was developed in 1999 as a joint effort between IBM, Microsoft, Altec Lansing, Roland Corporation, and Phillips. To transmit MIDI over USB a Cable Number and Cable Index are added to the message, and the result is encapsulated in a USB packet. The resulting USB message can be double the size of the native MIDI message. Since USB is over 15,000 times faster than MIDI (480,000 Kbits/sec vs 31.25 Kbits/sec,) USB has the potential to be much faster. However, due to the nature of USB there is more latency and jitter introduced that is usually in the range of 2 to 10 ms, or about 2 to 10 MIDI commands. Some comparisons done in the early part of the 2000s showed USB to slightly slower with higher latency, and this is still the case today. Despite the latency and jitter disadvantages, MIDI over USB is increasingly common on musical instruments.


http://en.wikipedia.org/wiki/MIDI#Al...are_transports


Seems like it still transfers better through 5-pin DIN cables
EvilDragon is offline   Reply With Quote
Old 03-03-2011, 05:27 AM   #3
stupeT
Human being with feelings
 
stupeT's Avatar
 
Join Date: Jan 2009
Location: frankonia
Posts: 1,996
Default

Old MIDI connections go through opt-coupling. This prevents gear to create ground loop humming (when additionally audio is connected as well).

USB wont do that.
__________________
------------------------------------------
Don't read this sentence to it's end, please.
stupeT is offline   Reply With Quote
Old 03-03-2011, 05:35 AM   #4
EricM
Human being with feelings
 
EricM's Avatar
 
Join Date: Jul 2009
Location: Ljubljana, Slovenia
Posts: 3,801
Default

I've experienced all sorts of issues when using the midi cable connection,
trying different synths, soundcards & cables (u never know). With USB
or FireWire all works as it should.

just some extra info on the subject.

e
__________________
Shoelace 4 Theme | SoundCloud/erXon
EricM is offline   Reply With Quote
Old 03-03-2011, 06:23 AM   #5
Evan
Human being with feelings
 
Join Date: Oct 2006
Location: Greece
Posts: 3,554
Default

I don't buy that... I think a native USB connection (and not a USB-MIDI convertor) will be faster and better behaved.

My Edirol PCR controller has two driver modes, a normal mode which merely transfers the MIDI data, and an enhanced mode (default) which takes greater advantage of the USB interface. Surely there must be something to it.
Evan is offline   Reply With Quote
Old 03-03-2011, 06:32 AM   #6
gkurtenbach
Human being with feelings
 
Join Date: Dec 2010
Location: Toronto Canada
Posts: 347
Default

Yes, there can be advantages of using midi versus usb. Surprisingly, usb may have higher latency and jitter. In my experience in getting low latency/jitter midi I found my M-audio 2X2 midi usb interface, was the poorest, much better was Cakewalk UG usb interface. Best was the midi input on an EMU 0404 sound card.

I use midi input from edrums to play VST drums so latency and jitter are critical.

There's free app called MidiTest which you can use to measure latency and jitter.
gkurtenbach is offline   Reply With Quote
Old 03-03-2011, 06:53 AM   #7
PeWe
Human being with feelings
 
Join Date: Dec 2007
Location: Hamburg/germany
Posts: 54
Default

Quote:
Originally Posted by gkurtenbach View Post
Yes, there can be advantages of using midi versus usb. Surprisingly, usb may have higher latency and jitter. In my experience in getting low latency/jitter midi I found my M-audio 2X2 midi usb interface, was the poorest, much better was Cakewalk UG usb interface. Best was the midi input on an EMU 0404 sound card.

I use midi input from edrums to play VST drums so latency and jitter are critical.

There's free app called MidiTest which you can use to measure latency and jitter.

http://miditest.earthvegaconnection.com/

PeWe
__________________
Intel DC 3.86GHz, 4GB PC800 Mushkin Redline, WinXP Pro 32Bit SP3, NVidia 9500, 3x WD Caviar Black 640GB, Sonic Core XITE-1 & SCOPE 5.1, Nuendo8 I/O, Nuendo TimelockPro, MQ8portSE, Tannoy DTM-8, Bryston3B Pro, Alesis M1 mkII, Alesis RA150.
PeWe is offline   Reply With Quote
Old 03-03-2011, 08:26 AM   #8
Evan
Human being with feelings
 
Join Date: Oct 2006
Location: Greece
Posts: 3,554
Default

Quote:
Originally Posted by gkurtenbach View Post
Yes, there can be advantages of using midi versus usb. Surprisingly, usb may have higher latency and jitter. In my experience in getting low latency/jitter midi I found my M-audio 2X2 midi usb interface, was the poorest, much better was Cakewalk UG usb interface. Best was the midi input on an EMU 0404 sound card.

I use midi input from edrums to play VST drums so latency and jitter are critical.

There's free app called MidiTest which you can use to measure latency and jitter.
Thanks for the information. The problem may be that edrums has MIDI OUT only, and no straight USB-to-host connection, is that the case? If so, it would make sense that the MIDI interface in your PCI/PCIe card is superior to the USB MIDI interfaces. But the question is, what if you bypass conversion and only ever route the MIDI data through USB?
Evan is offline   Reply With Quote
Old 03-03-2011, 03:19 PM   #9
gkurtenbach
Human being with feelings
 
Join Date: Dec 2010
Location: Toronto Canada
Posts: 347
Default

Its a good theory that usb to usb would be better than midi-usb and maybe midi to midi. I thougth so too and but found this not to be the case.

The only data point I have on this is my casio piano which has both midi and usb. I was shocked when I measured the latency and usb to usb was still slower than midi to midi *and* once in while usb to usb was really slow--around 20ms of latency!

I've also asked the creator of Megadrum drum trigger if his usb to usb is faster than midi out and was surprised when he sad the regular midi out was bit faster!

My theory is that usb is like a fast train that can carry lots of passengers but leaves every 15 minutes. Whereas serial midi is slow, carries 4 passengers but leaves whenever a passenger shows up. Or something like that:^) Sounds like Physic text book problem....

cheers!
gkurtenbach is offline   Reply With Quote
Old 03-03-2011, 05:01 PM   #10
Cableaddict
Human being with feelings
 
Join Date: Apr 2008
Posts: 1,910
Default

I don't think the problem is with USB per se. The problem is that most (all?) devices still use the standard midi protocols internally, then have to use a translation to turn it into USB.

If they were "pure USB" internally, I can't see how such problems would exists. Even with time-outs, USB is ridiculously faster than the midi spec. So, I'm sure it's fine with many modern devices.

Another problem might be that some testers were using an iffy USB hub, or even (horrors!) the USB input on a keyboard. (which on a Mac, is VASTLY slower than the chassis inputs)

For instance, I've measured both outputs from my Roland RD-700GX. the midi out gets to Protools (my main DAW, sorry!) via a top-quality interface, and the usb goes to a dedicated 2.0 input on my Mac. Timing-wise, they are virtually identical.
Cableaddict is offline   Reply With Quote
Old 03-03-2011, 05:03 PM   #11
Cableaddict
Human being with feelings
 
Join Date: Apr 2008
Posts: 1,910
Default

Quote:
Originally Posted by gkurtenbach View Post
My theory is that usb is like a fast train that can carry lots of passengers but leaves every 15 minutes. Whereas serial midi is slow, carries 4 passengers but leaves whenever a passenger shows up.
My theory is that you are using a Casio! (seriously.)
Cableaddict is offline   Reply With Quote
Old 03-03-2011, 05:09 PM   #12
Cableaddict
Human being with feelings
 
Join Date: Apr 2008
Posts: 1,910
Default

Quote:
Originally Posted by gkurtenbach View Post
There's free app called MidiTest which you can use to measure latency and jitter.
Interesting. Does this test USB-midi as well?

What is the danger with midi jitter? I'm not familiar with this at all.
-----------------

Quote:
Originally Posted by gkurtenbach View Post
I use midi input from edrums to play VST drums so latency and jitter are critical.
FWIW, if you want REALLY low latency with drum pads, you have to use direct-audio triggering, via Drumagog, Slate, etc. Midi conversion is inherently slow, especially for low-frequency signals like a kick pad or (the worst) a trigger mounted on a real kick.

I'm still using a D-Drum II pure- analog brain, triggered from Hart Pro pads, because I can get sub 3 ms total latency with decent dynamics. Even Drumagog can't touch that. (and the Roland / yamaha brains are much, much worse.) I use drum-replacer software afterwards. It sucks that we can't hear the "correct" drums during recording, but the band ALWAYS grooves better this way, and the sounds are still quite decent.
Cableaddict is offline   Reply With Quote
Old 03-03-2011, 06:11 PM   #13
Tod
Human being with feelings
 
Tod's Avatar
 
Join Date: Jan 2010
Location: Kalispell
Posts: 14,779
Default

Quote:
FWIW, if you want REALLY low latency with drum pads, you have to use direct-audio triggering, via Drumagog, Slate, etc. Midi conversion is inherently slow, especially for low-frequency signals like a kick pad or (the worst) a trigger mounted on a real kick.
Humm, what does frequency have to do with actual true midi pads?
Tod is offline   Reply With Quote
Old 03-03-2011, 06:24 PM   #14
TimOBrien
Human being with feelings
 
Join Date: Jun 2007
Posts: 2,301
Default

midi is a relatively SLOW protocol.
remember that it was designed on 8mhz pcs using a 1200baud serial port.

usb and modern card interfaces are FAR faster than that
so you dont have to worry about how it gets in the computer.
__________________
iMac i5quad/12gbRAM/1Tb Glyph drive/1Tb samples drive > Motu828mkII ---- Still run my GatewayM520 2.8ghz XP laptop. Video editing, 3D work and audio with zero problems.

Last edited by TimOBrien; 03-04-2011 at 08:13 AM.
TimOBrien is offline   Reply With Quote
Old 03-03-2011, 06:52 PM   #15
audioguytodd
Human being with feelings
 
audioguytodd's Avatar
 
Join Date: Mar 2008
Location: Atlanta, GA
Posts: 1,057
Default

Interesting that this question comes up now. I am currently building a USB/MIDI (both) controller. I have the USB side of things working now and I can say that the latency seems fine. I say seems because there is no real scientific way for me to test it right now. Once I get some MIDI jacks and setup the MIDI side of things I intend to try and test both at the same time to see if I can find any difference.

Does anyone even make a PCI based MIDI card anymore??? For modern operating systems????
__________________
Comp Specs: WIN XPSP3, Q6600, ASUSP5K, 3 GB PC6400 RAM, Focusrite Saffire, AlphaTrack, lots of plugins and hardware.
audioguytodd is offline   Reply With Quote
Old 03-03-2011, 07:37 PM   #16
EvilDragon
Human being with feelings
 
EvilDragon's Avatar
 
Join Date: Jun 2009
Location: Croatia
Posts: 24,812
Default

Quote:
Originally Posted by TimOBrien View Post
midi is a relatively SLOW protocol.
remember that it was designed on 8mhz pcs using a 1200baud serial port.

usb and modern card interfaces are FAR slower than that
so you dont have to worry about how it gets in the computer.
First of all, MIDI is 31.25 bauds. And USB is much faster than that. But it behaves differently, so a pure MIDI-MIDI transfer still wins.
EvilDragon is offline   Reply With Quote
Old 03-03-2011, 07:39 PM   #17
ObiK
Human being with feelings
 
ObiK's Avatar
 
Join Date: Feb 2011
Location: Salt Lake City, UT
Posts: 1,095
Default

Use what you have that works. If it isn't broken don't fix it.
__________________
IKMultimedia.com
Musicians first. Follow us on Facebook, Twitter & Tumblr!
ObiK is offline   Reply With Quote
Old 03-03-2011, 07:39 PM   #18
gkurtenbach
Human being with feelings
 
Join Date: Dec 2010
Location: Toronto Canada
Posts: 347
Default

Quote:
Originally Posted by Cableaddict View Post
Interesting. Does this test USB-midi as well?

What is the danger with midi jitter? I'm not familiar with this at all.
-----------------



FWIW, if you want REALLY low latency with drum pads, you have to use direct-audio triggering, via Drumagog, Slate, etc. Midi conversion is inherently slow, especially for low-frequency signals like a kick pad or (the worst) a trigger mounted on a real kick.

I'm still using a D-Drum II pure- analog brain, triggered from Hart Pro pads, because I can get sub 3 ms total latency with decent dynamics. Even Drumagog can't touch that. (and the Roland / yamaha brains are much, much worse.) I use drum-replacer software afterwards. It sucks that we can't hear the "correct" drums during recording, but the band ALWAYS grooves better this way, and the sounds are still quite decent.
The story behind jitter is that our brains are good at adjusting for consistent latency by adapting to delay--you just play a bit earlier. With jitter, you can't adapt like that because the delay is always changing, that is, "jittering".

I didn't believe this until I switched from my old midi interface which had 5ms of jitter (measure by MidiTest) to my new one which has < 1ms. I noticed that my playing sounded tighter and bit more in the groove.

Studies on human hearing show that when 2 beats are separated by >15 ms you start to hear them a 2 distinct beats. So if you are playing a steady beats and one beat is jittered 5ms earlier and next one is jittered 5ms late the brain starts to hear the last beat as if you were slowing down tempo because its 10ms from where it should be--hence your timing sounds loose.

Like I said I didn't really think that this would matter all that much when I started but it seems to be true.

BTW, my roland td4 drums take about 3ms from the time the stick hitting on pad to sound coming out of the brain. It takes the same amount of time from stick hit to the midi event appearing in Reaper. The vst plays out the sound for the midi event in about 2.5 ms so total latency is < 6 ms.

Cheers!
gkurtenbach is offline   Reply With Quote
Old 03-03-2011, 10:59 PM   #19
Cableaddict
Human being with feelings
 
Join Date: Apr 2008
Posts: 1,910
Default

Quote:
Originally Posted by Tod View Post
Humm, what does frequency have to do with actual true midi pads?
In order to track velocity (ow hard you hit) you have to read both the initial hit AND the crest. This takes a lot longer on a large pad / drum than on a small one. Do the math.
Cableaddict is offline   Reply With Quote
Old 03-03-2011, 11:05 PM   #20
Cableaddict
Human being with feelings
 
Join Date: Apr 2008
Posts: 1,910
Default

Quote:
Originally Posted by gkurtenbach View Post
BTW, my roland td4 drums take about 3ms from the time the stick hitting on pad to sound coming out of the brain.
That's amazing if true, though you may have measured incorrectly. - OK, maybe not. The cheaper brains tend to be faster because they do less.

When I tested the Roland Pro system (v 5, I think?) & top-Yamaha, about 4 years ago, things were horrible. With the Roland, the pads all varied, with the FASTEST pad (snare IIRC) at around 8ms. The kick was 12ms, I remember that distinctly, and the toms were the worst, probably because the brain was tracking positional info.

I sold all that garbage & got the Ddrum brain & Hart pads. Amazing difference.

Last edited by Cableaddict; 03-05-2011 at 11:26 AM.
Cableaddict is offline   Reply With Quote
Old 03-04-2011, 04:16 PM   #21
Cableaddict
Human being with feelings
 
Join Date: Apr 2008
Posts: 1,910
Default

Another issue:

When folks test midi speed, they are probably hitting one note, yes?

Well, the HUGE problem with midi transmission is that the notes have to be sent serially, and they get backed-up while waiting. I forget the numbers, but if you hit a chord (or have a complex midi files being transmitted) the delay is something like 1/2 ms per note.

Assuming a device used USB, and did NOT have to first go through any midi protocol, these delays would not occur.
Cableaddict is offline   Reply With Quote
Old 03-05-2011, 06:42 AM   #22
e.g.:
Human being with feelings
 
Join Date: May 2009
Location: USSA
Posts: 31
Default

Quote:
Originally Posted by TimOBrien View Post
midi is a relatively SLOW protocol.
remember that it was designed on 8mhz pcs using a 1200baud serial port.
And hasn't improved since then. Nothing beats the on board MIDI of an old Atari.

That said, unless you actually hear practical differences, or suffer technical difficulties, it doesn't matter how you connect.
e.g.: is offline   Reply With Quote
Old 03-05-2011, 07:10 AM   #23
XITE-1/4LIVE
Human being with feelings
 
XITE-1/4LIVE's Avatar
 
Join Date: Nov 2008
Location: Somewhere Between 120 and 150 BPM
Posts: 7,968
Default

I buy my child USB because he'll never do more than a note or two at a time, so he'll think USB is great.
Eventually I suspect he too will know he was scammed and say " Dad, I want a real soundcard and MIDI controller."
Then I will supply the child with a mans tools.
But we were all children and suckers before, its just part of a learning experience.
Smart kids even parrot impressive benchmarks to back their decisions of staying in the consumer markets.
Its all good.
But eventually they'll learn to play chords and graduate to higher levels.
You'll only notice then that MIDI is for men, and USB for boys and toys.
__________________
.
XITE-1/4LIVE is offline   Reply With Quote
Old 03-05-2011, 08:18 AM   #24
Evan
Human being with feelings
 
Join Date: Oct 2006
Location: Greece
Posts: 3,554
Default

Quote:
Originally Posted by XITE-1/4LIVE View Post
I buy my child USB because he'll never do more than a note or two at a time, so he'll think USB is great.
I sometimes manage to do three notes at a time, and still think USB is great!
Evan is offline   Reply With Quote
Old 03-28-2013, 06:24 PM   #25
Tom Drinkwater
Human being with feelings
 
Join Date: Apr 2008
Posts: 262
Default

most midi interfaces are connected to the computer through ... usb

so if usb is inherently a problem then you haven't removed it by using a midi cable from your keyboard to your interface instead of direct to the computer.
Tom Drinkwater is offline   Reply With Quote
Old 03-28-2013, 08:37 PM   #26
Stipes Vigilo
Human being with feelings
 
Join Date: Feb 2013
Posts: 152
Default

Quote:
Originally Posted by e.g.: View Post
... unless you actually hear practical differences, or suffer technical difficulties, it doesn't matter how you connect.
I had problems that I associated with either the DAW or the system. The DAW would continually see every signal as an audio signal to double on another track when only MIDI notes were being played. Another problem was in daisy chaining more than three on one with the one USB/MIDI-I/O. The laptop didn't seem to power more than two USB devices without a Powered hub to hook it up to. I finally broke down and got a multi-MIDI box that has its own power and not getting it from the computer USB. So now the the USB sees the device as MIDI and not the USB-Audio Device.

So for me, it has proven to increase consistency and timing.
Clearly defined what each stream of info is and removed unneccessary track usage (and even freed up another USB port).

I think it has mostly to do with how many external units you run, how many USB ports you're running them from, the drivers running each USB device and whether any of the units are drawing USB power along with it.

So it's not as simple as just seeing it all as USB plug n play.
Stipes Vigilo is offline   Reply With Quote
Old 03-29-2013, 05:16 AM   #27
Tom Drinkwater
Human being with feelings
 
Join Date: Apr 2008
Posts: 262
Default

If usb midi connections are bad, (and they frequently are) it is bad software or hardware design in a specific piece of equipment, not a problem inherent to usb.

we get sample accurate audio timing over usb, the only reason we don't get sample accurate midi timing over usb is that the software to do it hasn't been written.

the midi hardware spec is limited to a 1ms clock. that means minimum 1ms jitter whatever you do. but midi over other hardware need not have such a limitation.

send a lot of cc data and you clog up hardware midi very fast, leading to much higher jitter than the 1ms minimum. That's part of why high end 80's and 90's gear has multiple hardware midi ins and outs to distribute the load.

we should be over midi by now and using a sample accurate updated version, that we aren't is mostly historical.

developers tend to think a ms here or there of extra jitter doesn't matter, but each time someone makes that decision it adds up through the system, and the end result can be horrible, i think that's why usb midi often performs worse in the real world, lazy programmers and too many layers of software. hardware midi is not immune to this either it depends where the midi has been before it ever gets to the hardware midi interface.
Tom Drinkwater is offline   Reply With Quote
Old 03-29-2013, 06:28 AM   #28
tramway11
Human being with feelings
 
Join Date: Mar 2013
Posts: 139
Default

My USB-MIDI works unstable - there is big latency and other issues. I'm finding the ways to eliminate this
tramway11 is offline   Reply With Quote
Old 03-29-2013, 10:58 AM   #29
bluzkat
Human being with feelings
 
bluzkat's Avatar
 
Join Date: Jun 2007
Location: Northern Michigan
Posts: 6,919
Default

Quote:
Originally Posted by tramway11 View Post
My USB-MIDI works unstable - there is big latency and other issues. I'm finding the ways to eliminate this
Keep looking for those "ways", USB has been working fine here... for several years.

Do a little research, it will go a long way. Just sayin'...


__________________
Peace...
bluzkat
bluzkat is offline   Reply With Quote
Old 03-29-2013, 01:51 PM   #30
tramway11
Human being with feelings
 
Join Date: Mar 2013
Posts: 139
Default

Some cables may work poorly. Not all is fine. Once I bought absolutely non-working.
Maybe the reason in software. I'll look further.
tramway11 is offline   Reply With Quote
Old 03-29-2013, 09:25 PM   #31
Stipes Vigilo
Human being with feelings
 
Join Date: Feb 2013
Posts: 152
Default

The problem with USB/MIDI is in the drivers it decides to use when no specific device driver can be found. This can put the USB into a method of grabbing everything coming through and trying to place it in a specific realm. So a MIDI connection could be seen by the computer as a standard USB Audio device. I believe the reason this clears up when you use one of the multiple MIDI boxes from M-Audio or MOTU is because it comes with a specific driver for it (and sometimes some nice patching software). The USB is not the problem by itself. It's just how it handles assigning things connected to it with generic 'universal' drivers. And if you run three or more external devices using USB/MIDI or if the drivers for those devices are setup to do both audio & MIDI simultaneouly, then getting one of these boxes will make everything simpler.
Stipes Vigilo is offline   Reply With Quote
Old 03-30-2013, 08:04 AM   #32
Gerry G
Human being with feelings
 
Gerry G's Avatar
 
Join Date: Jul 2006
Location: Antipodes
Posts: 1,190
Default Lots of concepts here

I think that there are lots of aspects to this.

Measurement
Although measurement matters it may not give us the full picture in a practical sense and it may not be relevant to some uses of midi.
By Measuring latency from say the hit of a skin or key to the sound in the ear you are measuring input via midi or usb and then audio latency that may be delayed (further) by effects processing.
If we are just recording (solo for instance) then only the input latency will matter, especially if it is rehearsed and practiced ie where the player is not relying on a synchronised backing track. In that case the player is playing from memory primarily and is not relying on hearing the track so any recording will be ok but may be recorded later and capable of adjustment post recording. So for some things we need only measure input latency but for others round trip latency measurement is important.

Real world implementations
These vary from say playing a note now and then or triggering drums with a keyboard to fast Bach type multiple chords and notes with arpeggios etc. If it is just using midi to trigger an effects change then latency could be very high without any negative consequence. Latency may not matter much in non-percussive tracks.

Real World Correction
I have recorded some things using a very high latency but was subconsciously adjusting just as I would for a lousy instrument. You know how a great player can make a crappy instrument sound fantastic. It is because they know its strong points and subconsciously adjust to just play what nice things they can do with it. Likewise I think we can adjust for poor latency or even bad jitter but we are just making the best sound with what we have. This can lead to astoundingly good sounds (such as an out of tune honky tonk piano) in the right material. Likewise severe latency can force us to play well behind the beat but we can adjust our playing for it in some reggae situations for example. Non percussive organ sounds seem to be correctable this way (subconsciously) even if the latency is shocking.

Conclusion
With some material or some applications the latency will not matter much. In other situations it may limit or hamper severely what can be achieved.

Personally
I have used both USB and midi cables going into laptops and desktops both at home and in gigs and have not noticed any limitations with either. Same goes for mid pedals and, I think from memory, pure data applications. I have usually used XP optimised for low latency attached to MOTU firewire gear.

I am sure that measurement would show differences. On the other hand I have usually used the keys for non percussive synths and organs and I am not all that good on keyboards.

Last edited by Gerry G; 03-30-2013 at 05:32 PM. Reason: added bits and pieces
Gerry G is offline   Reply With Quote
Old 06-08-2018, 09:32 PM   #33
Scoox
Human being with feelings
 
Scoox's Avatar
 
Join Date: Jun 2009
Posts: 286
Default

Quote:
Originally Posted by EvilDragon View Post
In addition to the original 31.25 kbits/sec (baud is the signalling rate and is the reciprocal of the shortest signalling element; bits/sec is the data rate) current-loop transported on 5-pin DIN, other connectors have been used for the same electrical data, and transmission of MIDI streams in different forms over USB, IEEE 1394 a.k.a FireWire, and Ethernet is now common (see below).

USB

A standard for MIDI over USB was developed in 1999 as a joint effort between IBM, Microsoft, Altec Lansing, Roland Corporation, and Phillips. To transmit MIDI over USB a Cable Number and Cable Index are added to the message, and the result is encapsulated in a USB packet. The resulting USB message can be double the size of the native MIDI message. Since USB is over 15,000 times faster than MIDI (480,000 Kbits/sec vs 31.25 Kbits/sec,) USB has the potential to be much faster. However, due to the nature of USB there is more latency and jitter introduced that is usually in the range of 2 to 10 ms, or about 2 to 10 MIDI commands. Some comparisons done in the early part of the 2000s showed USB to slightly slower with higher latency, and this is still the case today. Despite the latency and jitter disadvantages, MIDI over USB is increasingly common on musical instruments.


http://en.wikipedia.org/wiki/MIDI#Al...are_transports


Seems like it still transfers better through 5-pin DIN cables
It's good you posted that info here, EvelDragon! I've just checked the link and most of that info has been edited out. Gone. A wake-up call to anyone who thinks Wikipedia is a reliable source.
Scoox is offline   Reply With Quote
Old 06-08-2018, 10:42 PM   #34
mschnell
Human being with feelings
 
mschnell's Avatar
 
Join Date: Jun 2013
Location: Krefeld, Germany
Posts: 14,972
Default

You should be able to reintroduce important infos in the article.

They should be in the back log and can be cut pasted and edited.

Regarding the info ED cited:

If you tunnel 5-pin Miti through USB of course some latency is added.

But if the source device featuring both interfaces generates a bunch of messages at the same time, the latency of the last one rather likely is higher over the devices 5-Pin cable than over USB, simply due to the rather slow bit rate of the Midi cable. (While the latency of the first message supposedly is higher via USB.)

-Michael
mschnell is offline   Reply With Quote
Old 06-08-2018, 11:26 PM   #35
Scoox
Human being with feelings
 
Scoox's Avatar
 
Join Date: Jun 2009
Posts: 286
Default

Quote:
Originally Posted by mschnell View Post
You should be able to reintroduce important infos in the article.

They should be in the back log and can be cut pasted and edited.

Regarding the info ED cited:

If you tunnel 5-pin Miti through USB of course some latency is added.

But if the source device featuring both interfaces generates a bunch of messages at the same time, the latency of the last one rather likely is higher over the devices 5-Pin cable than over USB, simply due to the rather slow bit rate of the Midi cable. (While the latency of the first message supposedly is higher via USB.)

-Michael
I think the real question was not whether there was latency, which is one thing, but rather whether MIDI-over-USB attempts to emulate a regular hardware MIDI connection down to the transfer rate. In other words, ignoring latency, is MIDI-over-USGB capable of higher throughput or does it deliberately slow it down to MIDI speeds?
Scoox is offline   Reply With Quote
Old 06-09-2018, 01:46 AM   #36
mschnell
Human being with feelings
 
mschnell's Avatar
 
Join Date: Jun 2013
Location: Krefeld, Germany
Posts: 14,972
Default

For non-techs it's hard to tell latency and speed correctly apart.

As speed is no issue with USB (some Megabits per second) vs. 5-pin Midi (32 KBits per second) latency is what might be worth discussing.

-Michael
mschnell is offline   Reply With Quote
Old 06-09-2018, 01:47 AM   #37
EvilDragon
Human being with feelings
 
EvilDragon's Avatar
 
Join Date: Jun 2009
Location: Croatia
Posts: 24,812
Default

The problem with USB is not latency, it's jitter. Overall it's going to be higher for MIDI over USB than for regular DIN MIDI (because USB is a serial packet transfer, not serial bitwise transfer).
EvilDragon is offline   Reply With Quote
Old 06-09-2018, 06:34 AM   #38
Scoox
Human being with feelings
 
Scoox's Avatar
 
Join Date: Jun 2009
Posts: 286
Default

Well, I guess we'll never know where the biggest bottleneck is.
Scoox is offline   Reply With Quote
Old 06-09-2018, 09:01 AM   #39
mschnell
Human being with feelings
 
mschnell's Avatar
 
Join Date: Jun 2013
Location: Krefeld, Germany
Posts: 14,972
Default

Jitter is varying latency
-Michael
mschnell is offline   Reply With Quote
Old 06-09-2018, 09:14 AM   #40
EvilDragon
Human being with feelings
 
EvilDragon's Avatar
 
Join Date: Jun 2009
Location: Croatia
Posts: 24,812
Default

Yes, which is much worse than fixed latency, no?
EvilDragon is offline   Reply With Quote
Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT -7. The time now is 09:54 PM.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2024, vBulletin Solutions Inc.