Thanks, everyone, for all these replies. A lot to process here.

A step and a half is a large move.  Also--and I don't know if you did this--using too much Variation tool (the thing that looks like a triangle wave) or too much of the pitch drift tool can affect the sound; and sometimes, oddly, the effect is most noticeable on the note following the one edited.
 
I am a heavy Melodyne user but I don't move pitches that far.  I always record more than one vocal track so I can copy a section over if I have to move a note too far.  Also, I take the best chorus in Melodyne and paste it into all the chorus slots.  (There is a trick to this if you need to know it.)
 
If a note has a lot of slope, I sometimes do one or both of the following:
- Split the note after the slope so I can pitch-center the main part of the note.
- Use some judgment to center the note so that its average pitch sounds right.  For example, if a note bends up and down and you center it right on pitch, it will sound off-key.
 
About formant, as much as I love Melodyne, I have had better results with the V-Vocal formant tool than with the one in Melodyne.  On the other hand, V-Vocal needs formant adjustment more often than Melodyne.  So, I seldom mess with formant in Melodyne.  If I need to, usually the pitch has been moved way too far.

Thanks, Konrad. That fills in a few blanks.

konradh
- Use some judgment to center the note so that its average pitch sounds right.  For example, if a note bends up and down and you center it right on pitch, it will sound off-key.
 

This is an important one if you're working on pitch correction and want the "best" results - centering blob position on the pitch grid often isn't the best option. 
 

konradhIf a note has a lot of slope

I was re-reading this thread and making notes when I noticed this statement and got to wondering what 'slope' is.
Are you referring to what Eli Krantzberg, in his Melodyne Explained videos, calls 'pitch drift?'

I too, do not use the formant tool. If I can't move the note transparently, it's time for a punch recording.
 
Something else I have started doing now on almost every vocal song.... I record the vocal as perfectly as possible and do it in 3 unique tracks/takes.  I melodyne one of them and call it the main vocal. If I run into a glitch of sorts, I have 2 other tracks available and some times there's just enough of a difference, or I sang that one slightly better so I can cut and paste the part as needed.  But after I have a finished main vocal track I use the other 2 lead vocal tracks as well. I lower them to -16db to -22db or so, pan them widely opposite and don't pitch fix them on purpose. This gives a slight widening and a fatter sound to the main vocal with the subtle differences in timing and pitch. The other 2 tracks are only audible when the vocal bus is in solo mode. In the mix, technically, they are still there, but covered by the instruments. The vocal just sounds a bit "larger".

Guitarhacker
I too, do not use the formant tool. If I can't move the note transparently, it's time for a punch recording.
 
Something else I have started doing now on almost every vocal song.... I record the vocal as perfectly as possible and do it in 3 unique tracks/takes.  I melodyne one of them and call it the main vocal. If I run into a glitch of sorts, I have 2 other tracks available and some times there's just enough of a difference, or I sang that one slightly better so I can cut and paste the part as needed.  But after I have a finished main vocal track I use the other 2 lead vocal tracks as well. I lower them to -16db to -22db or so, pan them widely opposite and don't pitch fix them on purpose. This gives a slight widening and a fatter sound to the main vocal with the subtle differences in timing and pitch. The other 2 tracks are only audible when the vocal bus is in solo mode. In the mix, technically, they are still there, but covered by the instruments. The vocal just sounds a bit "larger".

Interesting technique. Very interesting. I must give that a try sometime.
 
When I'm recording vocals, I usually try for three acceptable takes. That is: three takes I think are good enough to do the job. I do that for each vocal part (I almost always do the back-up vocals as well). And since I almost always do two back-up parts, I end up with three-part harmony comprised of nine complete and usable takes. So, following your technique shouldn't be a problem.
 
Thanks for passing that along, Guitarhacker.