It sounds like you have no control over the video/music. If you do, see my notes below.
Record at least 24bits at 96Khz for video and 88.2 for audio. That'll easily divide by 2 to achieve the final correct sample rate. This will theoretically give you the cleanest possible sound.
Record with external compression (in the analog stage) if you have the equipment. Have your narrator say the loudest thing s/he'd say, and adjust your levels so that it's just below clipping.
Make a rough edit. Once your narration is recorded, the best way to match it to your existing video is with a multitracker that has a video channel, or a NLE with multiple audio tracks.
Adjust each narration cut so all the levels are similar. Manually create envelopes, separate cuts and gain settings if necessary.
Make a detailed edit. Cut out the breaths, smacks, ums, make him say something different, etc.
Use envelopes to manually duck the background music. This part is especially creative. As a slow fade vs. a quicker one may have a different emphasis on the visual or narration.
Compress the narration track. It's a good idea to keep all the narration cuts on the same track so you can apply a uniform compression to it. I personally use Waves Ultramaximizer. Sony/SF has a similar plugin. I'm sure there are others out there. I try to make the V/O hot, but it should not sound breathy or unnatural. I don't go over -3db.
Mildly Compress the master channel. Your original background music track should not be compressed. If it is, and you can't get the studio to send you an uncompressed track, turn the gain down, maybe to -1db.
There should be no sound distortion and of course no clipping. The narration should be clear and easy to understand. The music should not obviously duck in and out. If it is obvious, you may need to work on your envelopes. It's OK to start bring in the music before the narrator is finished. This creates somewhat of a crescendo affect and adds emphasis.
Here is how I do voice overs.
Let me first note that it is important to not think of the voice over as a separate project or a separate job. It's the same with background music.
I don't let the narrator hear the music or watch the video - unless they have lots of experience. What usually happens is that they get distracted by the video and timing, or they follow the mood of the music. I want the narration to be a clear, expressive representation of what the client wants to convey. I usually have the client sign off on the narration before I begin editing. This way I can adjust the feel of the piece by varying the music and timing of the narration cuts.
Narration should be done as early into the production as possible. The reason for this is that it gives you more flexibility with your visual and audio edits. If you've already finished all your visual edits, and your client really wants the narrator to say 1 and 2 and 3, but you only have room for 1 and 2, you have to make some compromise on your visual edit, which might have tied in perfectly with your music.
- make a rough (visual) edit based on the story board. (You do have a story board, right?)
- record the narration.
- score or choose the background music.
- make a semi final edit with the music, video, and voice over.
4a. continue to make edits in communication with your client.
- edit the final - color correction, titles, FX, fades, flying turtles, etc.