Cable length in itself will not add any noticeable delay. Speed of light and all that. If you are using crap cables that cause errors, then the longer they are, the crappier they will be. USB specification maxes out at 9’ (~3m) so if you are talking about cables much longer than that, fuggetaboudit. The longest cable on my setup is 1m.
Even at USB 1 speeds the raw data rate is nearly 50 times faster than old-school midi. I’m not sure if there is additional overhead introduced by USB protocols. I plan to investigate that. I would like to know where the 5ms figure came from.
On the other hand, after all my praise for USB, my system has recently developed a problem where most (all?) of my controllers are re-enumerating themselves. I have to delete them and then add back in. I’m not 100% sure, but I think I see a number in the PID that is moving upwards. I’m currently looking for a pattern to this before starting a thread. Or maybe there is already something along these lines already in the forum. Looking…
Specs tend to be pretty conservative so you can push things at your own risk. In my studio I am using 5M cables with no issues (other than some devices being fussier than others as mentioned). But for some reason the live environment seems less forgiving and it is here I am going back to DIN MIDI for peace of mind
The golden test is the good old microphone over the keyboard. Record the sound of your finger hitting the key while simultaneously recording that midi note. Zoom in on waveform and compare the time of the transient to the midi event.
Agreed, the 9ft limitation in the USB spec is very conservative.
You gotta go with what works, but if you are using a PC (or mac). Midi will be USB at some point.
You are assuming that I would be allowed to be anywhere near a microphone
That seems like an interesting test. Have you actually tried it?
I spent part of my lunch break googling USB midi latency. Hard data seems hard to find. I did see some mention of the 5ms mentioned in this thread, but nothing to indicate it was actually scientifically tested. I kind of ignored any post that postulated that longer cables increase latency. a 10ft cable has about 10nS latency. That is many orders of magnitude faster than midi. I can confirm that USB does do some erro checking and resending on bad data. So if we are talking about actually dropping bits, then, yeah, latency goes up. But by how much? I dunno. I wonder of there is software out there that can report USB error rates, retries, etc.
One guy did a test similar to your microphone test, but using the start of audio from internal sound from one hardware synth to vs the start of sound from another hardware synth connected to the first over midi. He concluded that USB and DIN were the same. This test and your microphone test don’t account for delays in the instrument sound engine itself. But if it’s all low enough to be undetectable by hoomans then I guess it doesn’t matter.
You use a transient waveform in a sampler.
Any decent DAW places audio with sample accuracy, and the tolerances are far tighter than midi, obviously.
If you swap USB cables and consistently see a 5ms difference, you’ve established something.
I would not suggest using any hardware synths in such a test, unless you wanted to test the interface of that hardware instrument. Different problem.
MIDI delay is VERY detectable by humans.
In our particular use, where live delivery is the norm, it’s even more crucial to optimize these components.
Keep in mind that the USB cable is just a bit of copper. It’s inherent delay is about 1 nanosecond per foot. That is not detectable by anyone. If you are seeing cable differences measured in milliseconds it is because the connection is failing. USB is capable of detecting transmission errors and retrying. If the USB cable is the only variable in your experiment and one cable has conistently longer delays, then the cable with longer delay is defective. It makes a poor connection or the wire itself doesnot have the correct 90 ohm impedance for USB. The delays are caused br retries. Throw it away.
An old-school midi cable has no error correction. If defective you would have noticed missing notes, stuck notes, etc and immdiately thrown it away.
Differences in USB cables is not an argument against USB midi as long as the good cables are no worse than DIN midi. Have you run thus experiment with your good USB cable vs DIN?
The one thing that bothered me a bit for live situations is that DIN midi is a fairly robust connector. You are not likely to knock it out walking past it. USB A-B is a bit less robust. One needs to be a bit more careful about knocking them around. I find that using straight cables is not always optimum. I have a collection of up, down, left, right adapters to prevent cable loops from sticking out into space.
I just added my first devuce that requires USB-C. That seems like it could be even easier to knock around. I don’t know yet.
HaHa… kinda’ agree with Derek on the USB/DIN idea… although I currently am using a four port USB hub to drive my UR-22, nanoPad2 and mouse as well as DIN Quadra Merger for the pedals and DIN MIDI. The USB hub is attached to my laptop stand with the cables inserted to the hub. I use fingernail polish of different colors to denote where the USB cables plug into the hub. It has been very reliable this way and really has let me down only once in 3 years…(AFTER A WINDOWS UPDATE).
One utility that I’ve been using for USB outlook is 'USB Device Tree Viewer v4.3.0. https://www.uwe-sieber.de/english.html
It has a lot of info regarding USB ports and hubs and where they are connected. Right click on hubs and devices gives you some functionality. It has helped me in regard to keeping things right for finicky Windows more than once!
I’ve been using DORE MIDI FC-3 for a while now. 3 switches and I have an expression pedal plugged in. Works great for my setup. I left the center programmed as a sustain pedal, although I remap it in Cantabile a lot. And the expression pedal sends CC1 (mod wheel) but you can remap that too. A little quirky to get setup the first time, but I’m pretty happy with it.
It has been very reliable this way and really has let me down only once in 3 years…(AFTER A WINDOWS UPDATE).