bvideo
There's a related strangeness, though. That clip still had more notes beyond the clip boundary. They show up if you drag the clip boundary far enough. But quantizing with the original clip boundary then dragging the clip boundary shows that some notes beyond the original end are still there, while others have disappeared. Haven't figured out how to predict yet, but don't depend on notes outside a clip boundary being preserved.
 

I think this is the effect of me checking "non-destructive midi editing". It seems to only half work now, but in the past you could do slip editing and retrieve some lost notes that got erased with the autopunch. Now if you're lucky you can retrieve some notes, but not all

I don't think it's a bug (the OP). As I said in post #7 the notes were moved outside the clip.

bvideo
I don't think it's a bug (the OP). As I said in post #7 the notes were moved outside the clip.

It's clearly a bug. I've been using quantize for decades with every Cakewalk product since DOS. THis has never once happened until now. And it didn't just happen one time, it happened several times. I created the track by recording until I got to a spot where I didn't like what I did. I stopped, rewound a little to find where I wanted to resume and set up autopunch from there to the end of the song. I did that multiple times and it created many clips.

There does seem to be an issue in this case as one of the "lost" notes is not recoverable by extending the slip-edit boundary, but I'm not sure that's a totally new issue. I'd have to search but I think it's come up before. I avoid Non-destructive MIDI Editing mode because it can be quite destructive - even if only temporarily - and unintuitive in this way.

brundlefly
There does seem to be an issue in this case as one of the "lost" notes is not recoverable by extending the slip-edit boundary, but I'm not sure that's a totally new issue. I'd have to search but I think it's come up before. I avoid Non-destructive MIDI Editing mode because it can be quite destructive - even if only temporarily - and unintuitive in this way.

Let me know if you find anything in your search. It seems in the past non-destructive midi editing actually worked and if I punched in a spot, that if there was a problem I could use slip editing to fix it. Yet now it seems unreliable. How has it been destructive for you and why do you avoid it? What other experiences have you had with it?
 
Thanks for all your insight,

It's all in how you define "destructive". One key effect of NDME is that when you split a clip it effectively creates two copies of the data and slip-edits both to hide the overlapping parts. This means that notes crossing the boundary are effectively truncated at the slip-edit point. For anything other than drums, this is audibly "destructive" in that you lose the full duration of such notes.
 
In contrast, if you split a clip with NDME disabled, the result is two partially overlapping clips with the boundary of the first one ending at the original duration of the longest note that crossed the chosen split point. To me, this is less "destructive", and is usually the behavior I want want when I split a MIDI clip.
 
I just tested your project in some earlier versions, and found that quantizing in X2 and earlier with NDME enabled resulted in the clip's end boundary being extended to restore the original duration of the quantized notes before the slip editing was applied. This is also what happens in all versions when NDME is not enabled.
 
But in X3e as well as Platinum when NDME is enabled the notes get quantized later without the boundary being extended, and one of them is not recoverable. So it appears this got broken some time in X3.
 
 
 
 

brundlefly
...
In contrast, if you split a clip without NDME disabled, the result is two partially overlapping clips with the boundary of the first one ending at the original duration of the longest note that crossed the chosen split point. To me, this is less "destructive", and is usually the behavior I want want when I split a MIDI clip.

So can the resulting clips not be slip edited to expose unmodified data? Otherwise notes that start while that longest note is running could suddenly show up in the earlier clip. (Or maybe you meant "with NDME disabled"?)

brundlefly
I just tested your project in some earlier versions, and found that quantizing in X2 and earlier with NDME enabled resulted in the clip's end boundary being extended to restore the original duration of the quantized notes before the slip editing was applied. This is also what happens in all versions when NDME is not enabled.
...

In this X2 scenario, when the clip's end boundary gets extended by quantize, does it uncover other notes that may have started just after the unextended clip boundary, i.e. while the quantized notes at the end were running? Or did it effectively "trim" the clip so that there aren't any new notes to see by slip-editing?

Yes, I meant "with NDME disabled" in the second paragraph. I fouled up and changed enabled to disabled when I was quickly proof-reading, not noticing that it was preceeded by "without".
 
And, yes, I observed that quantizing in X2 and earlier did effectively apply trimming to notes not originally visible in the slip-edited clip, regardless of the status of NDME. This probably isn't ideal, either. 
 
And I think I was wrong about X3 and Platinum working as expected with NDME disabled; the clip boundary doesn't get extended and the one note is still lost in that case.

You dig deep figuring these things out, Dave (brundlefly). Gotta hand it to you!
 
So I wonder what CW intended to change in these scenarios. Seems like there are strange compromises to be made.

Tried X2 a bit. With NDME enabled:

1. Split a clip that had three notes near the split point, which was at a measure boundary. So two clips resulted, touching on one lane. I moved the 2nd clip down to another track for future reference. The two clips can be slipped to show all the notes from the original clip.
2. The clip has two notes that originally crossed the split boundary and a third that started on the split boundary. In the first clip of the split, the two notes appear truncated, the third one is no longer showing. The third note shows at the start of the 2nd split clip.
3. Quantized (the first clip). The clip lengthened, as brundlefly said, to show the ends of both notes near the end of the first split clip. The notes moved properly. The 2nd note now starts beyond where the clip boundary was, namely at the new measure. Maybe this is what gmp expected.
4. Then I slip edited that clip to look for the third note that used to be aligned at the measure boundary before the split. It was gone! But other notes beyond it were still there, though they had not been quantized. This is a little weird.

Then I disabled NDME and started down the same path.This time, splitting the clip made overlapping clips, just as brundlefly says. On two lanes. Slip editing the clips showed that there was no longer any hidden data. Of the three notes in question, two stayed with the first clip and extended the clip, and the third note started the 2nd clip, as expected. Then quantize extended the first clip further.
 
I suppose this is all as originally intended and expected. As brundlefly points out, this probably got broken in X3. I'll get to X3 a little later.