Stuck Notes w/Imported MIDI

Ran across an interesting one.
I’m just going to leave this as a ‘hmmm…’ unless someone else has noted something similar, and then maybe we can get some info together to see if it’s us or Pyramid.

Was just playing around with a generative plugin and made some MIDI Clips in Ableton, bounced 'em down, loaded them into a new project (rather, a template project) for the Pyramid.

Noticed I was getting stuck notes with at least one synth.

I checked the MIDI output and noticed that there was no 8x nn vv MIDI Event (Note Off) to correspond with one specific instance. Checked the Pyramid MIDI Monitor, and yup: no Note Off Event (even though the Note Event was finite - I think 2 steps long).

Note: I use the Chance 0% technique to nullify MIDI ouput and this would occur when I set the Chance 0% “On” (thereby nullifying). This technique does not nullify corresponding Note Offs…or rather, it hasn’t in the literal years I’ve been using this technique so I wasn’t sure if it was a “latest OS” issue.

Well, seems it wasn’t - I think.

I opened up Step Mode and noticed that, and I’m probably misrembering now, that the Note Events were displaying as being 2 steps long on the right side of the screen when holding down a pad corresponding to an active event. But the dot display showed 3 steps.

I ‘fixed’ this, or rather: I changed the Note Event to a different Note Length, then returned it to what it should have been (2 steps, displaying as 2 steps). Then everything played fine.

I’ve never had an issue importing MIDI Clips from Ableton.
The software plugin to create the clips is new to me, but I’d think once the clip was in Ableton, modified, and exported that should be a moot point.

I’m thinking perhaps some bizarre export/import(copy-load) thing, or data corruption.
But unless someone else has run into this, I’m not going to pursue documenting/troubleshooting.

Unless you’ve read this far and realise it’s something you’ve run across.
If so, let me know your experience(s) with this curiosity and let’s see if we can figure it out.

sorry, not quite following…

can we simplify to the underlying problem.

are you saying that if you import a midi file from ableton live, that when you play it, you do not see any note offs? does it go this on every track? just one track?

have you looked at the midi file thats generated by live… does it contain the note offs?
are they using 0x8x note off events or 0x9x note on events with velocity = 0. (either are valid)
do they look ok? e.g. correct channel/notes, whats the velocity?

obviously there are two steps to this…
a) checking midi file is what is expected.
b) check that pyramid is interpretting this correctly

if (a) is valid, and (b) is ‘wrong’ then the question is what is different in the midi file, compared to what the pyramid is expected… and under what situations does this happen.

hopefully the simpler the case we can reduce this too (without speculation :wink: ) the more likely you can supply a ‘test case’ to Squarp, and they can determine if they can ‘easily’ fix it or not.

I have seen this with a few tracks in one specific project.
I’m still modifying my workflow, so although I have “imported” Ableton MIDI clips, it is not something i’ve been doing often and with regularity.

Although I noticed this with a few tracks, it was bothersome one one track specifically because it was a bass line - the rest were percussive sounds that were finite sounds. The bass sound (destination: Minitaur (monophonic)) would come across as ‘stuck’.

Part A of your troubleshooting steps: I’m not familiar with reading hex in a .mid file as I don’t think I’ve run across a need to interpret them. There’s some stuff that looks like a header, then what I recognise as sequences of 3 bytes that are Note On & Off Events, with an extra byte between (timing, I’m assuming). You’d know more about this than I, of course. But trust me: I’m familiar with what Note On Events in hex look like, so I immediately recognised Note On and Off Events.

So, the note that’s getting stuck is a G#1 MIDI Note 0x20 I believe.
This is a curious string (formatted for better reading):

90 25 74 30
90 24 5a 00
80 25 00 60
90 20 6b 00
80 24 00 81 40
80 20 00 00

Also, the extraneous 0x40 in there occurs a few times across the file. I’m assuming you know what this is - the extra bytes are sometimes 00, 0x20, or 0x40 which are some kind of timing thing…? maybe? whut?

I’m assuming that .mid file data would be sequential - or if I’m using the wrong words or context, that each Note Event is read/processed start of file to the end. Note Off Events are present, but occur after subsequent Note On Events in this above snippet.

I’m thinking this is the problem and I should have cleaned this up in Ableton ahead of time. I’d swear I quantised start and end notes.

Note: This MIDI clip was created by a generative plug-in. I’m also thinking that with a normal polyphonic synth this would not be an issue (I should bop this over to a poly synth to see if it’s an issue - wait, I have this bass line layered and the poly synth layer plays fine), so I’m guessing:

  • This Plug-In made Note Off Events occur after subsequent Note On Events
  • I failed to rectify this in Ableton because the issue and potential problems weren’t on my radar
  • Monophonic synths don’t play nice with these kinds of sequences of notes (the poly synths on other Tracks apparently fixed themselves due to finite envelopes or just polyphonic note stealing forcing previous notes to cut off)

Still, watching the Pyramid MIDI Monitor I did not see the MIDI Out showing the 80 20 00 Events so…hrmmm. Monophonic synth destination should have no issue with this. (/thinking out loud)

Truly this is a curiosity for me: I resolved the front issue of getting things to sound correctly. I don’t know if this is indicative of an actual issue with Exporting MIDI Clips from Ableton for use with Pyramid, or an issue with data corruption (I haven’t heard of this on the Pyramid yet, but obvs something to look out for), or just a random stupid from me (honestly always the first option I try to check for, but dunning-kruger effect suggests I’ll never see my own stupid), or an actual failing of the Pyramid.

I just wasn’t sure how to troubleshoot further or if it was warranted.

Is there anything I should check further before filing this away as “curiosity” to be used later if a similar issue arises, or would you counsel picking up the pieces of my shattered life and moving on?

Funny: I tried to attach the .mid file but it’s not allowed.

This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.