@The Man, you're welcome, and I got the notion that you, like I, many times, are a one-man band playing many instruments.
And you're absolutely right, the trick is setting the levels low enough as not to peak out and over modulate, and just high enough to raise them in post in needed.
The on-cammera mic is nice for 'relatively' close audio captures, but not for all things. It has a 'newsy' feel and sound to it.
@Johnnyapple - In broad strokes, and that's where everything really gets distorted (if we want to stay with a tonal metaphor), using something like the H4N becomes really nice as it allows the use of professional XLR microphones with and also provides phantom power to them. That's nice. Sooner or later in any interview, one will want to set up a microphone on a stand or boom close to the subject, and if one is recording to a device, the device you'll want to record to will have XLR inputs with phantom power. I have other tools that work just as well or better, such as the JuicedLink box that has 4 powered inputs, each with discrete preamps and can mix them all into stereo channels. That's convenient for recording larger sound requirements such as a play or multiple speakers, anything where more microphones are necessary. Of course, it comes at a price that I can't really endorse for the average joe.
Naturally, you can also use a mixer board, I have a Beringer, with 16 inputs, for concerts, that go into another recording device. Again, how many concerts a year does one do for a living?
Ultimately, it's much, much easier if you lay your audio track on your video as you shoot. Always. A few tracks here and there aren't too heinous, but any project of size will take its toil on you in the edit bay.
When possible, attach a feed into the camera. And above all remember this, if the audio is separate to the camera, the camera's audio will be of paramount importance to synchronize the captured audio to the video. Make sure it's loud enough to hear on the camera and cue it with a clap and sync point. You should then be able to see the talent or assistant clap their hands (and hear it on the device's tract) and also match the wave forms on the audio tracks.