Audio: Better to Normalise in Peak or Max levels in FCP?

Hi,

I have some good quality voice overs in my FCP HD project. They average about -24 on the FCP input levels when set to their default 0db on the timeline.

I need to boast the volume of these tracks.

Am I better off raising the timeline levels to it's maximum +12db (aiming for a average/top input level average of -12db) or using Peak to Normalise these clips upwards?

Or would I be better to leave the levels alone and duplicate the clips (stereo pairs) onto 2 more tracks and thus have 4 tracks instead of two?

It's for PAL MiniDV thats going onto DVD.

Many thanks,
Darrin

Powerbook 17 1.67 512meg 100gig FCP HD