Before posting, please read: When to use this forum, when to submit a help ticket

MIDI stops working

Started by go2ldook, December 30, 2020, 07:25:41 PM

Previous topic - Next topic

go2ldook

Arlo:
Seeing a new issue since the last update. Here is my setup:
-Bandhelper plays backing tracks in parallel with a MIDI file that runs my gear
-Ipad connect to gear via RME Babyface Pro
-These MIDI files also send MIDI to Loopy HD on the iPad to trigger looper functions

This has all worked well, few glitches, enough to play live with this setup. But recently the MIDI stream has stopped randomly while I am practicing. Will be in the middle of a song and suddenly gear presets stop switching or loops are no longer triggered. I am assuming the issue is from Bandhelper, since it seems to affect the loops (which is all within the iPad as MIDI passes from Bandhelper to Loopy HD), as well as affecting other gear (MIDI leaving the iPad to the gear via the RME Babyface.

Any thoughts? This would be a set killer to have a loop continue when it is supposed to stop during a song

arlo

Do you have the latest iOS version? iOS 14.2 introduced several MIDI bugs, but iOS 14.3 fixed the ones I know about.

go2ldook

Yes, all up to date.

I had not played in a while. I was having some sound drop out issues previously but traced that to a cracked power supply on my Whammy pedal and fixed that. I reconfigured my pedalboard a little, mostly just trying to add the Aeros Looper so I could go with hardware looping, thinking it might be more reliable than what I have achieved with Loopy HD, but long story short, the Aeros looper does not have enough MIDI control available at this time, and I misread their intent to get this up and working soon. So anyway, I am back to the use of LoopyHD and my RME Babyface.

I am FAIRLY certain I did not have the MIDI dropouts when I got my board going again UNTIL I did the updates, which included updates to Bandhelper, iOs, and also to Audiobus. I do not have any MIDI channeling going in Audiobus...only use it for audio routing settings, so that should not be it.

I have looked through settings and tried various things to see if they have any effect. You can comment on them...some I don't understand their role in my setup:

1) Tried with "Play First 2 Recordings Together" both on and off. Seems to work either way, so right now I have it on, but seems to not be needed as MIDI files runs along with the audio file, so not considered a "recording" by Bandhelper anyway I guess
2) I removed the Tempo output channel. I had it set to Babyface Channel 3...are those MIDI channels showing up? It does not seem to do anything that I can see, and I get tempo info sent from Bandhelper by having MIDI Clock on. Looking at it now, with the RME Babyface unconnected, it gives me options of Speaker 1, Speaker 2 whereas before was Babyface Channels 1-12...so this appears to be selecting output for Bandhelper's tempo sound, not MIDI info. Anyway, I have deselected it.
3) Have tried with "Low Latency Recordings" and "Low Latency Tempos" on and off. Does not seem to matter. I left them on for now, though not clear to me what that does. I have set my "Delay Audio" to 0.05 sec, which seems to provide the best accuracy for my loop recordings.
4) I do not have a MIDI Soundfont file. I am not sure what this is or whether it is relevant to my set-up. I go to Recordings and attach a MIDI File to each backing track and the MIDI File has always run pretty reliably until now, where it variably seems to stop working. In addition to the MIDI file stopping, basically all MIDI shuts down: when I pick a new song, no MIDI Presets are communicated to my gear until I restart Bandhelper.
5) Send MIDI Beat clock is ON and sent to all Ports. I did learn months ago that I could have MIDI conflicts as MIDI clock would go to Bandhelper to my gear, but also through Loopy HD and to the gear, giving me crazy tempo results on my pedalboard gear. That was fixed by having no MIDI flow out of Loopy HD. I think the clock was getting roughly doubled.
6) Start MIDI on Auto-mute is OFF, and Auto-Mute instead of Auto-Stop is OFF. Not sure what these do anyway.
7) Send Custom MIDI is OFF

The only other difference from when this was last working well, I did have the RME Babyface set up with both MIDI Out and MIDI In connected to my other gear. Since I am not currently controlling anything on the iPad with my pedalboard gear, I saw no need to connect a cable to MIDI In. I will try reconnecting that and see what happens, but not sure if that would somehow affect MIDI stability. I have MIDI Thru in Bandhelper OFF, since MIDI flow in my setup is all FROM Bandhelper...nothing going to or thru it.

arlo

If all MIDI functionality stops, it sounds like the iOS MIDI framework is crashing. If you click the MIDI icon in the top toolbar and click the "MIDI Running / MIDI Stopped" button twice to turn it off and back on, does that have the same effect as restarting the app?

Unfortunately if the MIDI framework is crashing with the same data that worked fine before, that sounds like another iOS bug that was not fixed in iOS 14.3. But iOS 14.4 is on the way and we can hope that addresses it.

go2ldook

Quote from: arlo on January 01, 2021, 03:36:13 PM
If all MIDI functionality stops, it sounds like the iOS MIDI framework is crashing. If you click the MIDI icon in the top toolbar and click the "MIDI Running / MIDI Stopped" button twice to turn it off and back on, does that have the same effect as restarting the app?

Unfortunately if the MIDI framework is crashing with the same data that worked fine before, that sounds like another iOS bug that was not fixed in iOS 14.3. But iOS 14.4 is on the way and we can hope that addresses it.

The only way to get MIDI working again is to restart the Bandhelper app. This is very frustrating because, up until now, I have had a set-up that is stable enough for live gigging, but no way I would play live like this. I have done some live streams for fun and might make it 4 songs before MIDI stops and all my gear control disappears.

Note: have tried using Audiobus's MIDI connections but that seems to have no effect. Theoretically Audiobus is supposed to bypadd the weaknesses of iOs MIDI framework (supposedly)?
As far as MIDI issues with iOs, I cannot find anything regarding 14.3 with Google search, just a lot of talk about how 14.3 was supposed to fix issues.

So I guess I am stuck here. Looks like rollback of Bandhelper is difficult if not impossible, so cannot try that. This is beyond frustrating

arlo

Do you have another device with an older iOS version that you can use until iOS 14.4 comes out?

I haven't heard from anyone else who has this problem, though. The iOS 14.2 problems caused the app to crash, and I can see from crash logs submitted through Apple's system that those stopped with iOS 14.3. In your case it seems to be only the MIDI framework that is hanging up, and doesn't reset until the app restarts and rebuilds all the MIDI network connections. In that case, I might be able to change the "stop MIDI" function to clear out more stuff so you could simply restart the MIDI engine rather than restarting the whole app. I could possibly detect when the problem occurs and do that automatically, but that would be a guessing game if the problem doesn't occur on my devices.

Do you see anything in the MIDI status window after the problem happens? Does it still list your MIDI ports, and does it still show your outgoing data?

Do you see any files starting with "MIDI" or "CoreMidi" in the iOS settings app under Privacy > Analytics > Analytics Data?

Ahiru

FWIW we also use BH to play MIDI files in parallel to audio files, so yesterday I tried playing a songlist to see if might also see a similar problem (hadn't tried it for a while due to no gigs!).  It played through 10 songs (typical 3-4 min each) without any issues.

This is with iOS 14.3 on an iPad 6th generation (A1893), reported as Device Type iPad7,5 by BH, with BandHelper v4.1.23. I use an iConnect uAudio2+ for the audio and MIDI to downstream equipment (for MIDI that includes lights and effects). 

We don't use internal MIDI paths as you do @go2ldook for your loops, so maybe there is something unique there. Maybe you could try some tests with no loop triggering (but external equipment triggered) and with loop triggering but no RME interface to see if clues emerge.

I hope you can diagnose it @go2ldook... I can sure sympathize; a MIDI failure for us would mean pack up and go home!