As an allstar member of team too-lazy, I'll reiterate few obvious things already metnioned that are good to check and play around with before going too deep with the more advanced tools: 1) Adjusting the level of each channel independantly to get them more equal and/or to pan the apparent image somewhat as desired. 2) The 'cross-mixing' mentioned previously, which is equivalent to panning the channels towards center. There is no reason the panning or 'cross-mixing' needs to be done symetrically, you might find that it's best with the dominant side panned toward center slightly and the other side still hard-panned to one side, or vice-a-versa. Play around with it in combination with channel level and use the combination that sounds best. 3) If the channels sound significantly different in timbre, you might try EQ them seperately to match better.
A philosophical insight- We work with 'artificially repoduced sound'. Everything we do through the entire recording chain is 'artifical processing' of past real events, begining with the microphones and going on from there. Panning and level changes are no less artificial than more complex tools, they are just simpler manipulations that have less pitfalls and obscure complications than more complex ones.
Another idea I've been considering for a long time but simply haven't played around with is this, and I think it might work well here:
Use of Mid/Side decode/processing techniques to place the dominant channel in the center and spread the less dominant (more ambient channel) to the sides of the stereo image. The aim being to place the channel with the greater proportion of direct PA sound in the center of the playback image and spread the channel with more crowd and room sound out to both sides.
Not sure exactly how to do that yet, beyond simply assigning one channel as Mid and the other as Side and using a M/S decoding tool. I imagine it might cause some unwanted low frequency cancellation so it may make sense to only do this above a certain frequency. I have a handfull of 4 channel surround recordings where the Left and Right channels are compromised or corrupt but the Center and Back channels are fine, and salvaging those are what got me thinking about this, but it would also apply to other situations like this one where an different 'artificial' stereo image (center/outside) might be more appropriate than the typical left/right paradigm. An application closer to home for many around here would be producing a plesant stereo recording from a single mono SBD and a single mono AUD channel instead of simply mixing them to mono or using the typical pseudo-stereo techniques like hass delays or EQ combs..
I'm interested in your thoughts on that Page, and those of anyone else.