Nonsensical Event List Readings in Logic 10.7.8

[Plowman]

The left-most column is the start of the note. The right-most column is the end of the note. Apparently notes in Logic 10.7.8 go back in time. 

I cannot recall a more disappointing update than 10.7.8. 

[des99]

The right most column is the *duration* of the note, not the position of the end of the note.

It does actually tell you what that column is in the header:

Edited July 4, 2023 by des99

[Plowman]

It's set to "Length as Actual Position" which tells you the end of the note. 

[David Nahmani]

I can't reproduce this here, even with the project start marker before (or after) 1 1 1 1. How did you end up in that situation? 

If I try to place the note end before the position, the note end automatically jumps back to 1 tick after the position here: 

[des99]

9 hours ago, Plowman said:

It's set to "Length as Actual Position"

Oh I see... well that's pretty bad, then!

Like David, I can't reproduce this here either without definite steps...

[Plowman]

Thanks for reading guys. 

The hope here is not for a solution. This stuff just happens, and I cannot invoke it. Be assured I didn't drag the note to a minus setting, as that (like David said) just reduces it to a tick. 

No, this is how Logic chooses to render any given note length in what I can only call random, regrettable acts. But I did prepare the following, just as a study in observation.

Note the initial rendering of a region: See how the end points of the notes are strewn across impossible positions. Note as well that the D6's end point stands alone as correctly stated (65 2 4 176).

Now after I delete all *other* regions and go back to the example region, things change.  As you can see, most but not all note ends are rectified. The initial F5 and A5 are still wacky. 

Here's the GIF. 1. Show the example region. 2. Shift I to select all other regions and delete. 3. Go back to the example region. 

I'm choosing not to post this file, as it is long and important to me. If I catch it in a newly inaugurated file, I may post. I'm just documenting here. 

This file does have a modicum of time signature changes. I suspect that's why another bug -- Set Optimal Region Lengths -- doesn't work well. And that may underlie this issue with note ends. 

In another thread des99 posits that the press to include Logic as an iPad program may have lead to our current issues with 10.7.8. That is exactly what I have been thinking. Yes, all versions have bugs, and they affect users disproportionately. But 10.7.8 is just a cut above -- or below -- for a third digit update.  

[des99]

Crazy stuff indeed! When the note end positions are non-sensical - if you flip the event list to show lengths instead, what does Logic display for the lengths, out of interest?

Also how are those notes rendered in the piano roll?

(ie I'm trying to work out whether this is a display/info/text formatting bug in the event list, or whether the underlying MIDI data is actually corrupted - which should show up in all MIDI editors...)

Edited July 4, 2023 by des99

[Plowman]

17 minutes ago, des99 said:

Crazy stuff indeed! When the note end positions are non-sensical - if you flip the event list to show lengths instead, what does Logic display for the lengths, out of interest?

Also how are those notes rendered in the piano roll?

(ie I'm trying to work out whether this is a display/info/text formatting bug in the event list, or whether the underlying MIDI data is actually corrupted - which should show up in all MIDI editors...)

These are the same questions I asked myself. Piano Roll and Score are fine and accurate. 

Firstly, you gotta love it: I re-open the same file and the numbers are different. Wrong, but different. When it says 240 1 3 77, that is still wildly inaccurate -- even though it's not going back in time. This note doesn't last 175 bars.

So the GIF shows a coursing through some display options for the end of the note. Relative Position seems to be okay,. Length / Info looks fine too. Going back to Length as Absolute Position is still wrong. 

Then at the very end of the movie, I click off and on the region, and now Length as Absolute Position is correct. 

So be it. If I really need the right reading, apparently clicking off and on -- or deleting and un-doing-- gets me somewhere near accurate. 

Thanks for your interest. 

[rAC]

@Plowman has this been remedied in 10.7.9?

[Plowman]

I dialed back to 10.7.7 so I never did see if this bug had been addressed. My guess though is that it was adjacent to the time signature bug that corrupted the "optimal size" commands. That has been fixed, and likely this issue too. 

Did you ever see it? I seemed to be alone in this particular matter. For other 10.7.8 bugs I found myself in good company.