MOTUnation.com

I like to use MIDI scrubbing when I compose (i.e.: dragging the play wiper across passages). But MOTU/DP broke the way MIDI plays back, it doesn't work right. It's been broken for years. DP "won't face", they blame it on the VI (even though it is also a problem with DP's bundled VIs).

I'd like to know if it was fixed in DP10. But, since MOTU support refused to "face" I'm doubtful. However, if someone could confirm that it was fixed, it would be worth it for me to upgrade from 9.

A brief description of the issue:
"There is a known quirk in DP9 with a series of repeated pitches of full duration, where the first note plays normally, but the subsequent notes are played very very short (if at all, depending on the VI).

How to test in DP10:
Reliably fails with Pianoteq, create an instance of PT. Create a MIDI track with a series of quarter notes of the same pitch. Now quantize both attacks and releases to quarter-notes. Play. First note will be quarter note, subsequent notes will be hyper-staccato.

Select MIDI notes, REGION/Change Duration... remove 10 ticks from all notes. Playback. Should play fine. Now use scrub to see if notes play as they should.

As I've mentioned in other threads, it used to work correctly but was broken in some release of DP. I had a theory that DP had changed the way it dealt with note-ons and note-offs. The workaround of removing ticks of duration works for normal playback but not for MIDI scrubbing, even removing as many as 30 ticks won't fix it.

Is this the same thing as event priorities -- the rules governing the order of transmission when multiple events happen at the same time? When a note on and a note off happen at the same time for the same note, which one is sent out first to the VI or MIDI device? Presumably, you want the Note Off for an earlier note to go out first so the Note On for the later note with trigger the sound.

If this is what you mean, which way do you think it should happen? And why would the problem be specific to scrubbing?

I can recall in the past having repeated notes cut off, probably with overlapping notes. I either used the legato setting in the duration dialog, or reduced note length by a tick in the case of quantized MIDI. I vaguely recall reading that DP follows Note Off priority but I can't find any mention of this in the V10 manual, so its difficult to know if anything changed.

What do other DAWs do? This comes up when I work with others using FL, The Grim One, Studio One, GB and Logic, who ask for help in getting their MIDI to work. If I import their MIDI into DP, I don't get the cut off notes they get, so I have assumed DP does it right and the others don't.

How to test in DP10:
Reliably fails with Pianoteq, create an instance of PT. Create a MIDI track with a series of quarter notes of the same pitch. Now quantize both attacks and releases to quarter-notes. Play. First note will be quarter note, subsequent notes will be hyper-staccato.

In DP10.3. All notes play back normally when I do this operation - so maybe an improvement here for you

Select MIDI notes, REGION/Change Duration... remove 10 ticks from all notes. Playback. Should play fine. Now use scrub to see if notes play as they should.

I can confirm the second note scrubbed is usually hyper-staccato (as if Pianoteq is getting a note-on/off a the same time), but not always. The speed of the scrubbing and the zoom level can affect the scrub playback...if you're zoomed in and scrub slow enough the notes will playback normally, but yeah scrubbing is indeed effectively non-functional like this. Taking 10 ticks off helps btw. Taking 30 off pretty much eliminates the problem unless you're zoomed way out - so again, maybe an improvement?

A drag fo you I'm sure but this is a pretty specific bug stubbsonic. You know the drill. If lots of people needed this fixed it might be more of a priority for MOTU. This particular bug has never come up in my workflow, but I've had to make adjustments to accommodate other bugs that MOTU has yet to get around to...which isn't to say you shouldn't continue to ask. Good luck!

Thanks for responding @dix, and for thoughts, @bayswater.

As I compose I may slowly drag the wiper across a phrase and then let my "inner ear" whisper to me what comes next. This is related to melody, harmony, voice-leading, and voice allocation. Hearing all the notes stop suddenly is so irritating. I'm usually zoomed in so that I'm seeing about 4 measures in the screen (give or take).

I do think it is related to priority. I.e., it should always read NOTE OFFs before note ons. DP used to work properly, that is where I started to use my MIDI scrubbing method for composition. But at some point, MOTU must have switched this priority. (I don't know when, maybe v7 or v8?)

As to why it remains an issue during scrubbing, I can only guess. Perhaps scrubbing involves using a lower resolution time-stamp so it follows the event order, but again, if a note off and note on are too close together, the priority bites us in the butt.

I did start a techlink about it, but got no further than MOTU saying I should contact the VI developer (i.e., of which MOTU is definitely one).

Sounds like DP 10 hasn't addressed this at all, and I'm not surprised since I'm the only person bothered and MOTU seems to think it is someone elses' fault. Eeergh.

I've considered adding a second DAW for MIDI editing, but that just seems like I'd trade one headache for three others.

This puzzle me so I tried it out in 10.13. Just playing normally, it looks like Note Off priority is in effect. Scrubbing, the second note was cut off as you reported. But what's odd, with the same note played six times, only the second note was cut off. Notes 3 to 6 played properly. There is a copy of the simple project with six notes playing MX4 here in V9 and V10 if you want to try it out. If you see the same thing, it looks like a bug, not a Priority thing.

Thanks for taking the time to test it. It's helpful to confirm that it is not just me. I wish it was fixed in 10 (as surely that ship long sailed for 9).

MOTU won't fix it if they don't think it is a problem. And since it's a sneaky little bug for one person, I guess that's just how it is.

I filed a tech link with the sample project attached. Clearly not a problem specific to 3rd party VIs.

I appreciate your taking the time to do that, Bayswater. Very much.

I've spent a little more time with Pro Tools First. And I am impressed...

by what a steaming $#!+ pile that software is.

On the pros side, it does seem to have many of the features I'd like, and is generally pretty intuitive to figure out.

As for the cons: it's crashy/buggy, MIDI lasso selection is stupid, the software is intrusive (it adds a launch agent for AvidLink without asking-- you have to hack the launch agent to disable it), it doesn't show any 3rd party instruments, takes forever to launch (then it often crashes), there are NO preferences in ANY menus-- so I assume no custom key commands, and did I mention it is crashy?

BTW, I did have to trash preferences, but it didn't seem to help.

If you want to feel better about DP, try Pro Tools First for a couple days. Oooph.

I'll probably need to uninstall it, which will be a project.

On the brighter side, I have a workaround for my composition conundrum, I can use my PC3K as an external MIDI sound module for edits, composition, scrubbing. Then when I'm ready to render, I can switch over to VI's when/if I need a rendering stage.

I did reach out to Modartt support and they suggested looking at the MIDI monitor that is built into their plugin. I had done this in the past and comfirmed that DP is indeed sending note-ons before note offs on the same tick during scrubbing.

They also suggested trying the VST version (which is a good idea). I'll follow up on that.

They asked if it is still an issue with an exported MIDI file. Since MOTU seems to have employed some kind of fix for regular playback, I suspect the exported MIDI will work, but I will test it, anyway.

Not sure what to make of this. Are you saying that the note order for note on and note off on the same tick is different depending on whether you scrub or use regular playback?

bayswater wrote: ↑Thu Jan 14, 2021 8:33 am Not sure what to make of this. Are you saying that the note order for note on and note off on the same tick is different depending on whether you scrub or use regular playback?

Yes.

First, I failed to mention that Modartt claims it is DP's issue and not an issue with their VI.

When it was an issue with normal playback, it seems MOTU did somehow address it, but that routine that handles scrubbing didn't get the same correction.

The VST version of Pianoteq made no difference with the scrub issue.

stubbsonic wrote: ↑Thu Jan 14, 2021 9:21 am

bayswater wrote: ↑Thu Jan 14, 2021 8:33 am Not sure what to make of this. Are you saying that the note order for note on and note off on the same tick is different depending on whether you scrub or use regular playback?

Yes.

First, I failed to mention that Modartt claims it is DP's issue and not an issue with their VI.

When it was an issue with normal playback, it seems MOTU did somehow address it, but that routine that handles scrubbing didn't get the same correction.

The VST version of Pianoteq made no difference with the scrub issue.

Does Modartt know that this doesn't happen with other VIs, like Falcon? Not to say there is nothing wrong at MOTUs end, but how would they explain the lack of a problem with a number of other VIs?

They don't mention it as such, but I don't mind asking them.

If I was to venture a guess, when scrubbing, DP sends the data out on the wrong order, and there may be an obscure "rule" that tells VI developers to re-prioritize note-offs.

I wish I understood more what the original change was at DP that caused the issue, and then how they addressed it. At first I thought they may have changed their tools (like quantize, pencil, step) to shorten notes by one tick. All guessing.

Modartt was testing DP 10, and did get back with the final word:

There are no rounding errors. DP is not ordering the note-on and note-off as it should. Here is a list of events received when scrubbing a quantized MIDI sequence:

t=38285 : new MIDI event: note ON chan 0 note 45 vel 62 delta_frame=0
t=38331 : new MIDI event: note OFF chan 0 note 45 vel 64 delta_frame=0
t=39678 : new MIDI event: note ON chan 0 note 41 vel 62 delta_frame=0
t=39678 : new MIDI event: note OFF chan 0 note 45 vel 115 delta_frame=0
t=41024 : new MIDI event: note ON chan 0 note 42 vel 62 delta_frame=0
t=41025 : new MIDI event: note OFF chan 0 note 41 vel 115 delta_frame=0
t=41628 : new MIDI event: note ON chan 0 note 41 vel 62 delta_frame=0
t=41628 : new MIDI event: note OFF chan 0 note 42 vel 115 delta_frame=0
t=42371 : new MIDI event: note ON chan 0 note 41 vel 62 delta_frame=0
t=42371 : new MIDI event: note OFF chan 0 note 41 vel 115 delta_frame=0
t=42975 : new MIDI event: note ON chan 0 note 41 vel 62 delta_frame=0
t=42975 : new MIDI event: note OFF chan 0 note 41 vel 115 delta_frame=0
t=43532 : new MIDI event: note OFF chan 0 note 41 vel 64 delta_frame=0

At t=42371 a new note-on is received for note 41, immediately followed by a note-off for that same note.