Open Radar

 
Steps to Reproduce:
0. Have a Logic file open with multiple speech tracks (in my case, for a 2-hour podcast, with one person isolated on each track).
1. Invoke Strip Silence (^X) on a speech track's single giant initial region.
2. Set the parameters respectively to 1%, 1.2, 0.2, 0.3, Search Zero Crossing On.
3. Click OK to apply. On this first invocation, it should work correctly.
4. Select another speech track that hasn't been split yet. Invoke Strip Silence again (^X), and click OK to apply *without changing any parameter values*.


Expected Results:
The successive invocations of Strip Silence, without changing any parameter values, should apply correctly to their respective regions.


Actual Results:
Successive invocations of Strip Silence omit huge spans of non-silence from the source regions, losing tons of legitimate audio. It seems like it may be doing some kind of union-set with region data incorrectly cached from the previous invocation.

This is new in Logic 10.2.1. In prior versions, the expected results occurred.

Changing any parameter value in the Strip Silence dialog, even if it's changed back before applying, avoids the issue and works correctly (it appears to invalidate cached data when values are changed).


Version:
Logic Pro X 10.2.1