Bug in Midi Transform

[j_e_f_f_g]

I'm writing up the tutorial for how to use Muse's MIDI Transform function. I think I found a bug. It concerns setting the Filter's Event Type operation to "Unequal".

If you choose to match Note events like so:

2.png (3.08 KiB) Viewed 20309 times

Then it properly works. Here's the Event List for a test MIDI track I made:

1.png (13.86 KiB) Viewed 20309 times

The problem is with other event types. For example, select "Pitch Bend" events:

3a.png (3.16 KiB) Viewed 20305 times

[j_e_f_f_g]

I would expect this to select all events except pitch wheel. But it doesn't select Notes either.

3.png (13.82 KiB) Viewed 20321 times

Similiarly, "Control Change" doesn't select Pitch Bend, when it should:

5.png (13.26 KiB) Viewed 20321 times

Attached in a .7z file containing my test midi.

[j_e_f_f_g]

I found something else that seems weird. Let's say I want to select all notes in the range of a2 to e3.

Here's how I set the Filter's "Value 1":

5.png (3.49 KiB) Viewed 20295 times

Here's what gets selected. The a2 is selected, but the e3 is not.

5a.png (12.97 KiB) Viewed 20291 times

That seems strange that the low note of the range I specified is selected, but the high note is not. I would think that logically either both would be selected, or both excluded.

But ok, the low one is, and the high one isn't.

[j_e_f_f_g]

So I just need to increment (by 1) the high note to catch that e3. So here I set it to f3:

6.png (3.37 KiB) Viewed 20290 times

Well, I do catch the e3 now. But I lost the a2, even though I didn't change the low range at all. Huh?

6a.png (14.08 KiB) Viewed 20290 times

Note: I used the same midi test file as before.

[j_e_f_f_g]

Uh oh. I think I found something else, but I believe it's the same bug as above. Set the option to "Lower" than f#3.

9a.png (2.94 KiB) Viewed 20262 times

Apply, and this works at catching the a2 and e3.

7b.png (11.57 KiB) Viewed 20262 times

Lower it to f3 and Apply. Result is same as above (a2 and e3 selected). Good.

Now lower it to e3, and it doesn't select the e3. Still fine.

7.png (12.98 KiB) Viewed 20262 times

[j_e_f_f_g]

But raise it up to f3 and Apply. We should expect to see both the a2 and e3. But we lose the a2.

6.png (12.85 KiB) Viewed 20258 times

And yet when we went from f#3 to f3, the result for f3 was correct. In other words, f3 gives two different results, depending upon whether the previous value was below (e3) or above (f#3).

[Impostor]

I guess it's undocumented for a reason!

But, I (again) fail to see a use case where you'd want to transform events of all types except one specific type, within a single transformation, as you'd apparently want to do if you select "event type" to be "unequal" to the one type you don't want to transform.

Care to enlighten me (again)?

[j_e_f_f_g]

where you'd want to transform events of all types except one specific type

Maybe you have a drum part where the snare velocity is perfect but everything else needs to be increased. Easier to exclude the snare than select everything else.

Remember that there''s lots of other "rules" you can apply att the same time. So the "Unequal" rule doesn't always apply by itelf.

Give me a day or 2 to work on the tutorial, and I'll upload what I have so far. It's complicated, and I want this to be correct and complete.