Many thanks, some excellent suggestions!
I'm going to first try bringing them into Ozone7 standalone where they can all appear as separate tabs to toggle through. I'm not sure if all 28 will go in, but even four batches of seven will help greatly.  I have Studio One so their Project page is a good idea also.  Thanks everyone!
 
SimpleM - yes this is our 2x1 hour repetoire and we're bedding in a new lead singer. It also helps us all when practicing. I keep them on a hidden page on the band web site.  The curious can have a peep here:
http://www.thephoenixtrio.com/backing_tracks.html
 

Har-Bal has Match Loudness.  Here's a blurb from http://www.homerecordingconnection.com/news.php?action=view_story&id=463
 
"Use the "match loudness" feature in Har-Bal as a reference for your limiter. When using this feature, write down the gain number. Now pull the fader back to 0.0, record the newly EQ'd file and open the limiter. Set your outceiling to -0.1 and the threshold to the number indicated by Har-Bal. If there was an increase in Har-Bal of 4.3, set the threshold for -4.3. It works every time and the volume level is consistent throughout the entire album. This is an amazing feature in Har-Bal. I know of no other tool that does this."

In another thread I posted that the free offline TT-Dynamic tool will give you an overall RMS level for a song. It should be easy to pick one song as a reference, run the tool on all 28 tracks (takes less than a second per song) and get an rms level for each song. Next, simply take each song and boost or attenuate the output by the difference for that song from the reference song.
 
The nice thing here is that you're only changing the output level to match the reference song,  it is assumed that you've already mastered (or at least have a mastering chain in place) and your dynamics are already handled.
 
Of course, if you wish to take advantage of any added headroom in the case of an attenuation, or a reduction in headroom the case of  increased loudness, you'll need to do it at the mastering chain. The trick is you're matching RMS levels, and then adjusting your dynamics to taste (and available headroom).  
 
Me, I always master for 14db headroom from full swing (K-14 system). So, my RMS levels ares always hanging around the -14db level from digital max.

I think a major problem is that all these songs have been "mastered" without comparison to each other or consideration for the fact that they need to become one "unit". I appreciate the fact that this is just a demo for a new lead singer but on principle it makes no sense to master this way. Turning up the volume on some might reveal way more problems like different peaking behavior or bass discrepancies. You need to have them in the same environment when mastering.

Sanderxpander
I think a major problem is that all these songs have been "mastered" without comparison to each other or consideration for the fact that they need to become one "unit". I appreciate the fact that this is just a demo for a new lead singer but on principle it makes no sense to master this way. Turning up the volume on some might reveal way more problems like different peaking behavior or bass discrepancies. You need to have them in the same environment when mastering.

I agree in principle, but if all he really wants/needs is a loudness match to avoid any nasty surprises in overall loudness from track to track, it really is nothing more than gain matching.
 
Ideally, you'd want to master using closely the same chain on on each song so that they all "sound" the same, let alone their loudness.

mettelus
Hopefully someone with Ozone 7 can chime in.

This is easily done in Ozone by setting the input level fader and checking the results using the Ozone Insight plug-in.  I think that plug only works with the 32-bit version of Ozone, but that's no problem.
 
Indeed, you really need to set the input level appropriately within Ozone if you expect the other Ozone tools (compressor, limiter, and dynamic EQ) to work correctly.

1) Load each song in a separate project at the same time in Sonar
 
2) Find a convenient smallish size for the projects so that you can make a long cascade of all the songs from top left to bottom right. This makes it easier to flip between songs, see 5).
 
3) Add the mastering FX you think you will need to a song that is pretty representative and tweak until you are in the ballpark that you want all the songs to be. Write down which song it is!
 
4) Save as template and load that template into the rest of the songs
 
5) Tweak each song whilst flipping between your representative reference and other songs, starting and stopping playback as needed, until you are satisfied with all your songs.
 

When I receive album cuts for assembly/sequencing, the first thing I do (horrors!) is normalize them all to a peak value of -0.5 dB. Then I listen and reduce the level of cuts that seem loud in comparison to the others. However if there is inconsistency in the mastering, then I fix that...for example if one track sounds softer because it wasn't limited to the same extent as other tracks, I'll tweak that first.
 
I don't necessarily agree that all levels should be matched. Slight volume variations can "lead" the listener along a particular path. On some projects I use short transitions between songs, and sometimes the level of these is ratcheted back quite a bit so that when the next song hits, it hits hard. I also think volume variations within a song can be effective, but this requires starting at a softer level so that the louder sections are more dramatic.
 
As a result in my own albums I establish a baseline with the first cut that's not the maximum possible level but has consistent dynamics. The goal is for the listener to adjust the nominal listening level based on that track. That way I have a little bit of room in terms of having levels that are both higher and lower than the nominal level. I consider dynamics an important part of music, not just within individual songs but in the "lost art" of album assembly in a singles world.
 
Of course, this assumes a listener who really wants to hear my music. It's not necessarily the optimum approach for commercial success...maybe someday there will be a return to listening to music that's collected into an album, and is designed more for taking the listener on an extended journey rather than isolated singles. I've started work on my next album and the opening cut is pretty brutal, but when it goes into the next number, which is basically just a vocal and acoustic guitar, the contrast is stark and (at least I think) effective. The plan is for the songs to tell a story and there are ups and downs. However these days, probably most listeners would listen to the first cut, not like it, and never hear the rest.