I was unaware of this (not a big "read all the docs guy) but I found this:
From some experimentation:
Assume you have a sample, where you might want to use slices to get different distinct sounds (e.g. a drum break).
Once you make your slices, each slice gets assigned to a single key (in the keyzone).
Each assigned key plays that slice at whatever pitch it has in the original sample.
Often that’s exactly what you want.
Now, if you want your sample to play at a different pitch based the key, you might (typically) not slice it at all. (Maybe more useful with short samples.)
But if you use that “octave for transposing” option you get to mix these two behaviors.
That first octave is reserved for the first “slice”, which is really just the whole sample.
When you hit a key in the range of that octave you get “key drives pitch” behavior. Once you go above that octave you get the “key selects slice” behavior.
You can get this same behavior if you do not slice your sample and instead use the Sxx command to play a sample at the given offset. (This is mentioned in that Reddit link.)
The use of the explicit octave-for-transposing is very handy when you want to carefully select slices by hand (and perhaps add looping, etc.).
I’m mainly a fan of explicit slices it makes it easier for me to interpret what’s happening in the pattern editor, and easier for to improvise and record using a MIDI controller.
Having the option to do this and have some re-pitching from different keys is something I need to play with.
Thanks for asking this question!