Jimmy, each method has its advantages and disadvantages you just gotta choose which you can live with better. If Cubase workflow wasn't so darn good I might have switched to Sonar over this integration issue.
Of course other than getting to bypass the Transfer process (with tempo changes and other annoyances), there's are a couple things that make it easier to work standalone for me. First of all the vocal is outside it's processing and backing track which can be good or bad. You can hear more potentially if it's not smothered in reverb. It is harder to tune outside the context of the song, still you can hear it in your head anyway and meters are a nice backup. But the biggie is that when you run Melodyne as
a plugin in Cubase in the conventional manner, you are forced to use Cubase's navigation controls to get around the Melodyne interface. How fast your cursor gets around depends on your zoom level in the host app. So in essence,
the Melodyne cursor and timeline is slave to the DAW, which I find supremely annoying and a bit glitchy when trying to work fast. In stand alone you can drag the cursor along the timeline and navigate promptly, because of course it is the master. When I get it back in Cubase I can mix it with no frills. Because if, for instance, you want to trim off the silence at the beginning of a recording, you can't do that because it still referencing Melodyne. Of course, silence can be deleted in Melodyne, but the same hold true for all other offline processing. There are many other minor reasons but these are the main ones.