You had the right idea with CEP/AA:
Save Selection should create a new file with only the selected portion of the waveform in it. IME, CEP/AA has an excellent Help file, far better than most other apps I've tried.
I've worked a LOT with Audition, a fair amount with WaveLab, and now with Samplitude SE. While I found Audition the most intuitive, and WaveLab the least, my learning curve with Samplitude SE fell somewhere in the middle. Audition's destructive editing eventually became too big a PITA as I started editing my files more and working with 24-bit.
Edit to correct: the multitrack editor in Audition allows for non-destructive editing and adjusting effects order and properties, but one may only apply the effects to the entire track, not individual clips within each track. WaveLab offers non-destructive editing, but I found it wholly unintuitive, and its linear editing (as best I could figure out, anyway) drove me nuts. Samplitude SE's non-destructive
and object-oriented editing is the bee's knees, as far as I'm concerned.
So why are these two features so important? I think destructive v. non-destructive editing is pretty straightforward. Destructive editing modifies the data itself. So one must always make a backup copy of the master file first, before editing. Okay, no biggie, it's not hard to maintain a copy of the original WAV. But one must also save copies of any interim edits if one wants to roll back easily to a previous state. Gets tedious and slow, after a while, not to mention it eats up HD / archiving media space.
But the real biggie is object-oriented v. linear editing. In both WaveLab and Audition, linear editing restricts the operations one may perform / re-perform / un-do. For example, let's say I apply three different, but consecutive edits: <1> compress, <2> normalize, and <3> fade. Now let's say I want to re-do the compression because I've decided I want to use a different ratio or threshold. In order to re-do the compression in a linear editing model, I first have to un-do the fade, then un-do the normalization, then un-do the compression, then re-apply the compression (with my new settings), re-apply the normalization, and re-apply the fade. Or, instead of un-doing all three, I have to revert back to a WAV file I saved (as an interim state, as noted above) before re-applying all three edits. And I hope I've made notes about any edits
after compression, because if I haven't, I'll need to re-determine the settings I want to use, the portion of the waveform I want to edit, etc. Ugh.
With object-oriented editing, I can apply those same three edits - <1> compress, <2> normalize, and <3> fade - but tweak or un-do or remove any of the three at any time, independently of the others. So in the example above, I would simply modify the compression edit with my new ratio or threshold or whatever, without having to touch the subsequent normalize and fade edits, without having to revert to previously saved states, etc. Beauty!
It's also simplified my archiving. I now just archive the original WAVs + the project file, since in order to re-create both the final 16- and 24-bit versions, all I have to do is simply
File | Export Audio | WAV and select my format (16/44, 24/48, etc.). No more saving interim versions of my edits in case I want to roll back to them. No more saving the master WAV + the edited version in both 16- and 24-bit. Just the master + the project file.
Anyway...I'm long past rambling. I think Samplitude SE v8.3 is well worth the whopping ~$50 or so it costs. Big thanks to Teddy for encouraging me to give it a go.