Bizarre staff silence depending on notes input earlier

Bizarre behavior, I’m getting no playback for a single staff, but only in very specific circumstances.

After a great scavenger hunt, for my score I find that adding notes in one measure causes playback to go completely silent for that staff for MUCH of the rest of the flow, but ONLY a number of measures have passed where the playback still works. I tried using the default sounds (instead of my usual Noteperformer), creating a new staff and copying everything from the old staff, and changing the instrument to something else. It’s very reproducible, but not making a lot of sense so I assume something got corrupt. I noticed a similar behavior sometimes where a staff would be silent after merging flows, but in those cases closing/reopening Dorico would get things working again. It also happens when I export the flow’s audio to a WAV file. I created a little test file from the project I was working on, which contains only the staff in question, which I can share…

Yes, do post the cut-down file. Words can get us only so far.

I didn’t see a way to attach a file in the forum (otherwise I would have done so when creating this issue). Here’s a google drive link to a folder containing the dorico file and a text file with the details about the issue:

https://drive.google.com/drive/folders/1FwWdav2lHRrWnwXCUTSDQfKysZWwJvo-?usp=sharing

In my original project, if I deleted the notes & dynamics in measures 32-38, then it seemed to restore the playback for the rest of the movement for this staff. However when I try that in this test file, there is still no sound at measure 76 and most measures after that.

I have been examining your file.

Three observations:

  1. In Play mode, when viewing in the Key Editor, at places where the sound stops (measure 76) or resumes (the measures preceding measure 220), the velocity is 0%, whereas nearby notes which do sound have a velocity of 100%. Using the various tools (Draw, Line, etc.) to increase that value above 0% proved completely futile.

  2. Adding, removing, and changing the position of slurs in such measures did affect where the playback resumed.

  3. After doing Edit > Select All and then Edit > Reset Appearance, playback appeared to be normal for the whole piece. As expected, resetting the appearance moved the notation around in some places.
    I have no idea whatsoever as to why this would fix the playback problem. Maybe the Dorico development team might be able to offer an explanation.

= = = = = = = = = = = = = = =

EDIT:
I have narrowed it down even further.

When I selected ONLY measure 243 and did Edit > Reset Appearance, playback for all measures returned to normal. The trill in measure 243 also played, whereas previously it had not.

Viewing measure 243 while making that edit, I noticed that the trill line changed. Its original properties were Start speed: Slow and End speed: Fast. Working through the 9 possibilities I found that three combinations did not play the trill:
Start speed: Slow with End speed: Normal
Start speed: Slow with End speed: Fast
Start speed: Normal with End speed: Slow

With those three combinations, playback in the rest of the piece displayed the problem as described by the OP in the first post. With all the other combinations, playback was as it should be.

For the purposes of further testing, I changed the trill in measure 243 to Start speed: Normal with End speed: Fast.

The trills in measures 23 and 27, both of which were Start speed: Slow with End speed: Normal, played as expected before and after changing the trill in measure 243.

The trill in measure 231 which is Start speed: Normal with End speed: Fast did not play before the change to the trill in measure 243 but did play after that change.

So, until the Dorico development team has a chance to examine the situation, the short-term fix will be to change the properties of the trill in measure 243.

1 Like

Unfortunately for some reason I’m unable to download the file from Google Drive (which is throwing some kind of error). Jesse, I believe your trust level on the forum should now be sufficient that you can simply drag and drop the Dorico project into a reply, and it will be uploaded properly. If not, you can email the file to me at d dot spreadbury at steinberg dot de and I’ll take a look when I get a chance.

@dspreadbury Here’s the file, which I had downloaded.

playback issues start measure 76.dorico (846.3 KB)

Also, here it is downloaded just now from Google Drive.

bizarre-staff-silence-depending-on-notes-input-earlier-20230917T221509Z-001.zip (839.6 KB)

Thanks for attaching the file for me, Steven.

The workaround proposed by Steven of unsetting the Start speed and End speed properties of the trill in bar 243 is a good one; alternatively, if you want to preserve the appearance of the trill line in that bar, try instead activating Suppress playback for that trill.

What’s happening is that in order to fit the desired notes into the playback of the trill, Dorico is ending up with a very tiny fraction for the difference in the duration between the range the tuplet occupies and the notes it has generated to play the trill. In reconciling the difference, it’s using an incredibly precise value that seems to be causing an overflow in the arithmetic elsewhere, or something along those lines. It will require a bit more investigation than I can muster at midnight on a Sunday night to fully solve it.

1 Like

Thank you! With these workarounds I should be good for this project, it helps knowing what the actual cause was. Indicating the speed of the trill may be a bit over-prescriptive anyway, but I thought it was cool to be able to be so precise so I thought I’d try it out.

By the way, another workaround I found works well is to move notes that aren’t playing into another voice (like a second upstem voice). At least for cello I don’t need that voice for polyphonic writing. But knowing that the issue is the trill, and considering the silent passages are hard to predict, I’m just going to avoid the issue altogether by not overriding the trill speed. I’ll let you know if I still have these issues after removing those trill speed overrides.

Thank you very much for looking into this for me – especially at midnight on a Sunday night!