Scene Edit Detection only applies to source material.

[Dan Meadows]

I have a question: In order to perform a Scene Edit Detection, you have to make a sequence of a clip and then do it on the sequence, unless I'm doing it wrong and I don't think I am.  So if you're performing it on the sequence, why is the action made only to the source clip?  This seems illogical.  What I mean/what I need to do is sync a clip up with audio (splits), make a sequence of that merged clip and then subclip out only the synced nat sound.  However all that is subclipped is the origianl source material and its mix, which isn't any help to me.  Why go to the trouble of making this a sequence oriented workflow if this happens?  It would seem to be a reasonable and common task that I'm trying to do.

[R Neil Haugen]

First, do not use merge clips to sync separate audio and video files if the camera file also has audio. The intended tool is "create multicam" ... odd as it sounds, when you are only linking one video and audio.

That may be part of your problem here ... or all of it. Try editing as a multicam instead.