MIDI Incorrect On/Off Messages

[ary48682]

Hello everyone,

I'm currently working on a project in which I am sending MIDI data from Pure Data, over the "Interapplication MIDI" setting, into a MIDI track in Digital Performer 9. The main issue here is that DP seems to be misunderstanding the note on/off messages, and the MIDI information, as a result, is a total mess.

Looking at the MIDI Monitor, I can confirm that the note turns on and off, and does not hang. I can confirm that in an earlier version of DP, a number of years ago, I somehow got this to work. I've attached a picture. Note how the rhythm here is constant. https://imgur.com/2faptD7

This was today's attempt. I'd like to emphasize that this is the exact same Pure Data patch. Nothing has changed whatsoever. https://imgur.com/OJOirB5

What I can't seem to understand is whether there is a delay here, and it's a synchronization issue, whether I'm sending some incorrect information (though I did check this multiple times and it doesn't seem to be so), or whether this is a different issue entirely. Any help would be tremendously appreciated. Thank you.

[bayswater]

Your attachments don't appear. You have to post them somewhere, and then put a URL in your message.

First thing I'd check is if there are any MIDI filters on that might mess up the data.

[ary48682]

Your attachments don't appear. You have to post them somewhere, and then put a URL in your message.

First thing I'd check is if there are any MIDI filters on that might mess up the data.

Fixed the link. Don't know why they would be on by default, but will check out the filters.

[ary48682]

Your attachments don't appear. You have to post them somewhere, and then put a URL in your message.

First thing I'd check is if there are any MIDI filters on that might mess up the data.

Fixed the link. Don't know why they would be on by default, but will check out the filters.

Checked the filters, the "Off Velocities" option was off for some reason, I enabled it, but the issue remains.

[stubbsonic]

Those notes in the original example look very short, and possibly delayed.

There is a known quirk with DP where at a given "tick" It will read note-on's before it reads note-off's. That is my theory, anyway. The symptom that I observed was if I had a series of notes of the same pitch where note-attacks AND releases were quantized such that repeating notes were adjacent (i.e., the release of the previous note and the start of the next note were on the same tick); then the first note would play, but the subsequent notes would be very short.

So, in short, if there are three adjacent quantized quarter notes of the same pitch, DP would read it in this order:

Note 1 ON
Note 2 ON
Note 1 OFF (which immediately kills notes 1 & 2)
Note 3 ON
Note 2 OFF (which immediately kills note 3)
Note 3 OFF (not doing anything)

Any subsequent notes are shorted because of this.

This may or may note be related to your issue, but if your notes from Pure Data have no sustain, you could try lengthening them-- even just as a test.

[ary48682]

Those notes in the original example look very short, and possibly delayed.

There is a known quirk with DP where at a given "tick" It will read note-on's before it reads note-off's. That is my theory, anyway. The symptom that I observed was if I had a series of notes of the same pitch where note-attacks AND releases were quantized such that repeating notes were adjacent (i.e., the release of the previous note and the start of the next note were on the same tick); then the first note would play, but the subsequent notes would be very short.

So, in short, if there are three adjacent quantized quarter notes of the same pitch, DP would read it in this order:

Note 1 ON
Note 2 ON
Note 1 OFF (which immediately kills notes 1 & 2)
Note 3 ON
Note 2 OFF (which immediately kills note 3)
Note 3 OFF (not doing anything)

Any subsequent notes are shorted because of this.

This may or may note be related to your issue, but if your notes from Pure Data have no sustain, you could try lengthening them-- even just as a test.

Thank you for your answer. I'm currently working in DP 9.52. Is it known for any MIDI-related issues? I can confirm that I am indeed sending note off messages. I've tried messing around with the duration (to ensure they aren't too short), but that doesn't seem to do anything.

My current work-around has been to create an option for MIDI-file export, and then manually import it into DP. That seems to work quite well, but of course, is more work than I was initially looking for.

[bayswater]

There is a known quirk with DP where at a given "tick" It will read note-on's before it reads note-off's.

What a memory, I'd forgotten about that.

[stubbsonic]

... I'm currently working in DP 9.52. Is it known for any MIDI-related issues? I can confirm that I am indeed sending note off messages. I've tried messing around with the duration (to ensure they aren't too short), but that doesn't seem to do anything.

My current work-around has been to create an option for MIDI-file export, and then manually import it into DP. That seems to work quite well, but of course, is more work than I was initially looking for.

I'm also using DP 9.52. The issue I mention is the only MIDI problem that I'm aware of. Sadly, MOTU support/developers either didn't understand the issue, or just decided there wasn't a good way to fix it.

The only time I quantize note releases like that is when I'm doing MIDI for notation. My workaround is to put a MIDI insert plugin to subtract a couple ticks from each duration. This way playback works as it should.

Please check something. If you drag the long note out of the way, are there other notes underneath? If so, then it seems that DP is somehow missing an occasional note off.

If you generate a longer test, is there a pattern to what it gets and what it misses?

FWIW, I've never had much luck with interapplication processes. But to be fair, I didn't try very hard before giving up.