Audio levels and processing for broadcast

I have to deliver video programming for a cable show that is broadcast from mpeg files. The guy doing the encoding wants the audio to peek at -3db and average between
-10db and -6db. The programs are typically live presentations using wireless misc and some forms with table mics.

My questions are
1. How hot should the audio levels be set in camera? We shoot with DSR 250?s, and I have the audio peeking at -12db
2. I want to use soundtrack Pro to condition the audio, what processing should I use. The audio always looks anemic on the time line before processing it in STP. I have used both the limiter and the compressor to boost the volume.
3. What does normalize audio do, I have been told to normalize sets the 0 db level.
4. How can I get audio levels peeking at -3 db without distortion
5. What is the scale on STP referring to

I understand that a limiter can boost the gain x dbs and not exceed a determined db level. I thing a compressor can make audio sound louder with less dynamic range. I don?t know what Normalize audio does. I have noise reduction working. Im just not sure how to put this all together. Please give me a work flow and settings. Thanks

Never give up, there is always a solution