I’ve recently been noticing occasional “hangs” on Pianoteq where, when using the sustain pedal, every once in a while a single note seems to sustain even after the sustain pedal is no longer pressed. (Eventually it fades out.) This is effectively similar to what you would expect from the sostenuto pedal, but my piano doesn’t have a sostenuto pedal. The phenomenon is difficult to provoke intentionally. I “unlearned” all pedals except the sustain pedal from the instrument chain just in case, to avoid any kind of distraction.
This is with the current Pianoteq 8 on a Zynthian 5.1 connected by 5-pin DIN MIDI to a Roland FP4F, with the sustain pedal hooked up to the FP4F in the usual way (with a 6.3mm jack). According to the Zynthian’s UI, the sustain pedal works as normal – you can see it being engaged and disengaged. On the off-chance, I’ll be trying USB MIDI next to see whether that changes anything, and I might also try to get at the native Pianoteq UI in order to find out what Pianoteq’s idea of the situation looks like (I presume I need to use VNC for that).
Has anyone had a similar experience and, if so, what did you do?
Note that we’ve got two intermittent issues with Sustain Pedals - admittedly different symptoms, and they could well be unrelated, but I wanted to point that out.
That occurred to me too, but I think the symptoms are sufficiently different in either case that it is better to not handle them in the same discussion.
The issue doesn’t seem to occur with an USB, as opposed to a 5-pin DIN MIDI connection – I just had my first live gig with the Zynthian and it performed flawlessly. I’m starting to suspect that there might be something not 100% right with the MIDI cable I was using, and may try another one just to see whether that changes anything. But as it is perfectly possible to use USB-based MIDI the original point is now kinda moot.
For the purpose of reporting a possible similar issue (but different, all notes affected) as @anselm had, I thought I would explain my issue from a intensive practice session, last week.
Connection: Midi Keyboard with Sustain Pedal, connect via 5-pin on a Zynthian V5 (pi4) running ORAM (first practice session with oram, otherwise pre-oram stable with no issues).
After ~2 hours the sustain release was often not being recognized/received from the engine, this was not restricted to just Pianoteq but also other synths. Changing snapshots would remove the sustain (of course) but pressing and releasing the pedal again brought back the issue that releasing was not recognized. Testing was limited but adding and removing the Pedal Jack showed no change but depending on the mechanical wiring this does not exclude that it could still be the cause.
I’ll put together a list of tests which could help to identify the cause, this may help others with similar issues, I need to think about it first.
If no one has opened a “GitHub issue tracker” then I do it, I couldn’t find one.
Both my official hardware 5.1 and a bandmate using a DIY 5.0 on a Pi4 also experience this. For me, it’s associated with specific MIDI sources. I can easily, mostly reproducibly make notes stick with MIDI send from the Octatrack (not by flooding it with data, it only takes a handful of non-overlapping notes and the most replicable scenario for me is to start playing a sequence on the containing one note on the Octatrack and then stop playback while the note is still sustaining, especially using Helm or OB-XD (Mimi-D is unique, in that just two or three sequential MIDI notes sent from the Octatrack will almost always make it completely stop producing sound until I actually remove the chain completely). I’ve been using the Octatrack since 2017 and the Zynthian is the only piece of hardware I’ve used with it that has ever had an issue like this.
However, I’ve tried other DIN MIDI sources (from 80s keyboards to modern controllers) and haven’t been able to get a note to stick even when I deliberately try.
I’ve never had the issue with controllers conncted directly by USB myself, but my friend’s DIY Zynthian only uses direct USB connections to the RPI for MIDI and he also gets stuck notes.
There are a few threads on here from 2019 about this issue, too, mostly attributing it to specific synths (especially Helm, which is the most prone to it in my experience too but it’s definitely not specific to Helm).
At some point today I can use the Octatrack->Zynthian setup to produce some stuck notes while also recording the MIDI data directly from the THRU jack into a DAW. That way we would have a standard MIDI file that had never touched the Zynthian software at all but was known to cause stuck notes, which could be useful for testing.
I do not see a ticket in the issue tracker for this. Please raise a ticket using the Report Issue button in webconf. Please try to provide a workflow that reliably triggers the problem so that we can reproduce it.
I’ll put a ticket together for my version of the problem as soon as I have time, but I can already say one thing:
turning off multitimbral mode (for the onboard DIN input) makes it stop happening (so far).
EDIT: spoke too soon, I hadn’t soloed the MIDI track (on the Octatrack) that was actually causing the stuck notes. When I solo it I still get the stuck note when multitimbral mode is off.
I do suspect this has as much to do with the Octatrack as it does with the Zynthian, but since the Zynthian is the only hardware I’ve ever encountered that behaves this way with the Octatrack it seems like it’s something that could be overcome on the Zynthian end.
EDIT 2: I’ve been able to consistently replicate it but one of the two variables that determine whether it sticks is whether or not it receives a transport stop message. When I file a bug report I’d like to be able to include standard MIDI files that will replicate it, but I don’t know of any software that will let me record transport messages. I can monitor them easily enough but actually recording them isn’t something that any software I know of seems to do - they all either route transport messages to their own transport or filter them out (which is what makes sense 99.9% of the time of course).
Anyone have any recommendations for software that will record raw MIDI data from an input, including system realtime messages? I’m going to keep looking for a little longer but I have to start work soon.
EDIT: I haven’t had any luck at all finding a way to record a MIDI stream including system realtime messages. I’m sure it’s possible to build a tool in Puredata or Supercollider or something but probably not worth the trouble. I’ll write up a bug report.
EDIT 2: more like writing a novel than a bug report but I think I got some pretty useful information in the end.
Thank you for creating the Issue report, I’ve seen the zip file and I’ll try and reproduce the issue this week. Your method seems to reproduce the issue more reliably than mine so let’s take this route. I seem to vaguely remember your video (I could be totally wrong here) that you initially made, can you point me to it again, I can’t seem to find it again.
FWIW I just finished practice and didn’t have a single stuck note in almost three hours of continuous playing, from the Octatrack sequencer AND a KMI L-Board at the same time.
The other guy with the DIY Zynthian didn’t seem to have any either, but he used nothing but Fluidsynth tonight and it seems to be immune to the problem for me.
EDIT: I hope you saw my followup post in the bug report, by the way. I got to spend a little more time testing and corrected some stuff.
Just chiming in here that I also get stuck notes almost daily.
Usually requires a reboot to fix it. I haven’t been able to see a pattern to what causes it. It happens with many instruments so not related to something specific
New 5.1 user with official hardware, nothing special about my setup and not sending crazy amounts of midi.
Will post back here (or create bug report) when I’m able to provide steps to reproduce it.