I am putting this into the Cubase category but this also applies to Nuendo and all other Cubase versions.
There seems to be something severely off with how Cubase handles Midi from some WIDI devices, such as the Carry-On 49 bluetooth midi controller and many others.
However, I seem to have no issues with my trusty Korg Microkey 37 which has the most solid WIDI ever.
This video demonstrates the following:
Cubase and Logic simultaneous recording (to my best abilities haha)
Displaying the recorded result and comparison
As you can see, Logic records the signal without any issues whatsoever. It is very accurate!
Cubase on the other hand seems to have issues with note-off events and also the placing of the Note-On events in time. This gets more confusing when it does this seemlingly retrospectively after the events have been recorded.
I also tested retrospective recording and the result is similar.
There is also a 2 year old thread mentioning the issue:
And I found this video in the Gearspace forum, displaying the problem with an earlier version of Cubase:
No wonder musicians seem to consider Bluetooth Midi as mostly unreliable and unstable. The initial experience you make with this stuff is crucial on how to approach technology in general. This is one of these situations. I have found Bluetooth midi always very useful and reliable - thanks to one of my very first devices being a CME controller.
Here is another video with the Midi monitor active, so you can see what’s going on:
I did another test with a Microkorg, connected via a CME WIDI Master plug and the recording is just as good as recorded with a regular cabled controller (except maybe with 2ms of extra latency).
Is the Carry-On really such a bad Bluetooth controller and why does it work just fine with Logic Pro? Does it come down to licenses and patents?
Update:
I think I found something in the Midi monitor data:
The Widi dongle recorded midi shows that all Note-Off events are being reset with Velocity value 64.
The Carry-On controller instead, sends velocity value 0 with every Note-Off event.
Could that be something that Cubase is interpreting wrong?
Video of the data. First 5 notes are Widi Adapter, last 5 notes are coming from the Carry-On 49:
Update #2
OK after setting up recording via cabled connection, the velocity data looks identical. I was even able to record simultaneously from the wired and the bluetooth connection and while pitches overlapped (as intended), you could clearly see the off-timing data, coming from the bluetooth connection. This is so strange to me. Why would Cubase change behaviour of Bluetooth midi sources if the event log looks near identical? It’s like it is applying some sort of prediction/smoothing algorithm but rather than making things better, it just turns everything into a garble.
Well would you look at that. This bug was already reported in 2017 and has since then remained unfixed:
The fact that CME Bluetooth Midi is flawless and the rest is unusable speaks volumes. To me, this looks like this is simply a question of licensing. I’ll use my controller via cable for now and maybe if there’s a smaller “Widi Jack” released by CME, I’ll step over to that.
I’m a long-time Apple fanboy but Mac OS’s built-in Bluetooth has always been a bit skwonky. It’s never been very reliable for me regardless of OS or machine (and there has been a lot). If you can find a 3rd party Bluetooth (like Logictec’s dongles) you’ll have better success.
I’ve always preferred cables as they just work and never bothered me. It’s always been a solution to a problem that I never came across.
I disagree. Apple Bluetooth, especially on newer devices is just very good these days. I can’t tell about older devices but on my two macs (Mac Studio and Macbook Pro M4) there aren’t any connection issues. But like I said, this may not be a bug or an issue with Bluetooth - it is merely a question of software support and licensing.
As you can see in my first post, I used Logic as a realtime comparison - recording the same midi events in parallel in Logic. In the video you can see Cubase on top and Logic Pro at the bottom. Logic, to my surprise, seems to have no issues at all with whatever Bluetooth midi controller I use - no matter if it’s a CME or a noname thirdparty controller. Their midi is solid
I don’t think it’s the Bluetooth standard, but the way Widi messages are being processed. I am not an expert, so I was just trying to make sense of it.
OK, so I just tested Reason 12 and Ableton, and here are my Results overall:
Cubase:
CME, Korg : Widi playback and recording works perfectly
Third party Widi: live playback works but recording is unusable
Reason 12:
same as Cubase
Ableton:
All devices playing back and recording perfectly
Logic Pro
All devices playing back and recording perfectly
The way the events allocate and stretch out in a weird way in Reason looks a lot like how Cubase is doing it. They may be sharing some similar set of libraries here? Again, I am not an expert
Please let me know if you want me to test any more. I have most DAW’s installed
Thank you @dreikelvin for your detailed observations. We were able to reproduce the problem (reported to Steinberg, internal reference: CAN-57796). As we’re just processing the timing information we get from macOS, without knowing whether there’s BluetoothMIDI involved or not, we will have to dig deeper, what Logic and Ableton are doing differently. Thank you!
Yes, the problem isn’t even fixed in Nuendo either - it appears they are still working on it. The bug has been registered though as I have had contact with support. Let’s hope for a near future maintenance update
Nuendo & Cubase 14.0.30 update is here - and I just tested Bluetooth Midi on my foldable CarryOn midi keyboard. I am happy to report that this bug is fixed now!
