Notes on voice recording:
With some of the better DV cameras such as the Sony VX 1000 and 2000 or the Canon XL1 the camera's own microphone is good enough. For voice recording just set the camera in front of you and speak into it. The camera remote control can be handy for stopping and starting. The smaller handycam and notepad type cameras can record good sound but an external mic is necessary to capture good sound. A piece of thin cloth stretched between yourself and the mic can help avoid pops from plosive consonants such as "P"s and "T"s. Otherwise, speaking just to one side of the mic and not getting too close also helps.
With digital sound of any kind it is important to avoid clipping from recording at too high a level. If you have trouble finding a quiet place an automobile can make a pretty good recording booth.
One of the biggest difficulties for most people is to avoid a self conscious monotone and achieve a natural delivery with good emphasis. Having someone to actually tell the story to can be a big help. Don't worry about getting everything perfect. If you repeat something to get it better it is easy to edit out the bad takes later. Remember, you aren't trying to give a formal speech but just telling a story so a certain amount of the normal, less than grammatically perfect, pattern of speech is OK. Naturalness is more important than perfection and genuineness is much better than trying to sound like some commercial narrator telling us all about things of which they know nothing.
When recording is done transfer the whole thing to a 16 bit 44 KB/sec digital file on your computer and edit with an audio editing program such as Sound Forge or Cool Edit. This makes it easy to cut out the bad takes, tighten up long pauses, and clip out Uhh..s, coughs, and other distractions. If the level is a bit low it can be normalized and depending upon the microphone, recording environment, and nature of the voice itself you may also want to use a bit of equalization and compression to achieve the best sound.
When it sounds right use Windows Media Encoder to compress it. If you don't already have it, it's available free from the MS web site in the Windows Media Player Section. The preset for "Near CD Quality" is probably the best for this purpose. It yields a compression ratio of about 30:1 or around 1.75 MB for 5 minutes of stereo audio. Higher quality is unnecessary for voice but the lower ones (e.g. "F.M. Radio Quality" seem to lose the nuances of voice which provide a close-up and genuine quality to the experience.