Well, of course there's such a thing as peak average power vs average power - it's usually referred to as the crest factor. You have to know that, the longer a power level stays at a particular level, the more it influences the long term average - so peak average levels can negatively influence the average (assuming you want a greater crest factor).
 
If anyone is up to date on the loudness 'wars' and has taken a look at the  BS1770 standard, you get schooled in the knowledge that: 
 

• loudness is perceptual
• loudness is frequency dependent
• loudness is measured in units of time

There's nothing wrong with, and it's the norm for most, with measuring average loudness the full length of the song. You can still note peak average levels if you want, and if you're interested, you can get your hands on a meter that does crest factor metering. Blue Cat's DP Meter Pro does this, shows crest factor, and has the necessary histograms to view averages vs time.
 
But, as some ancient wit said, "This is all academic.."
 
 
 

Again thanks for the info, I've got a brief breather today before moving on to other projects (thank you Lord for the projects!) and I hope to digest some of the amazing material above.
 
I finished the song mix and "little m" master on Friday morning but had some other pressing work come up so the singer did her own video late last night. The song is just a Karaoke download with her voice dubbed in. It is 2 steps two low for her so we had some real challenges with that but for a rush job, "we got 'er done".
 
I made a number of versions playing around with added harmonic content and increased upper mid EQ's and I favored the one without both (her voice sat down in the mix sweeter, I thought), but she favored the one with both and that is her prerogative ( I suspect she may have listened on a laptop).
 
The church will play it today from YouTube rather than a DVD.
 
http://www.youtube.com/watch?v=7ri_WN22eJo&feature=youtu.be
 
Thanks again so much for the help!

He seems to have really enjoyed life. Nice job with the editing!

Mike I just noticed something. In your post #5 where you did some (Adobe Audition) analysis of the various tracks (from O'Hawke) these are the results you get for Track 8. (I only mention track 8 because as I said it was one of the louder ones)
 
Track 08
Average RMS Power: -11.68 dB     -11.92 dB
 
When you did this check what are your rms settings down below. You will see you have them set for Square wave. If you switch this to sine wave you get an average result of -8.6 instead and a max rms level of -2.3 as well. Just another place for error don't you think! It can be all rather confusing.
 
If you look at what Ian Shepherd has been saying regarding rms measurements he definitely states that you should be using sine wave for music NOT square wave. This also applies to generating test tones too. There has been some discussion as to what type of wave should be used for test tone generation and from what I have observed lately it should be a sine wave.
 
It clearly demonstrates that the average for the track is -8.6 and not -11.8 as you are suggesting. I was aiming for -7 actually when I mastered that. I can see where the -8 comes from because of the quieter sections averaged over time but for most of the track it is hitting -7.
 
Not that I am a fan of these high levels either but sometimes the clients just want it this loud.