8.5 PE export mixdown sounds different

Greetings to y'all, it's been a while. :)
 
Long story short: I'm still on PE 8.5 (because reasons), and I'm starting to realize my mixdowns does not sound 100% like the playback sound. The differences are subtle but audible. I would suspect that this is due to plugin behaviour in some way, I'm using lots and lots of plugins and instruments that are newer than the Sonar platform itself, in theory this maybe shouldn't be a problem but maybe it is anyway.
 
One example: I did a mixdown before work today of a song I'm working on. Opened the file in Wavelab and immediately thought "too much guitars". Went back so Sonar to correct it, but the guitars were fine. Two tracks of Revalver 4 btw, the level difference is probable somewhere around 1-1.5dB.
 
Is there anything I can try besides fiddling with the settings in the export dialog? Are there any "deeper" settings that could affect the mixdown results? I tried searching for related topics but didn't find anything particularly useful.
 
Thanks guys. :)

There are a thousand potential reasons for minor differences, so any hypotheses offered here would be pure speculation. However, it's unlikely that the SONAR version or possible version mismatches of plugins/instruments are to blame.
 
When investigating such matters, the first thing I do is narrow the suspect list. Export your project as a 32-bit wave file and then import it back into the same project. Use the solo buttons to switch between the in-project mix and the imported file. If differences are heard, use analysis tools such as SPAN to determine what's different (spectrum, levels, average RMS or dynamic range). If no differences are heard, then there is something about the WaveLab environment that's altering the sound or altering the way you hear it.
 
To eliminate plugin behavior that might change during export, try exporting the wave file with all plugins bypassed. Yeh, it'll sound pretty awful, but it should sound the same after importing back into the project.

Yes, relevant question. But: to be able to actually compare the differencies after I discovered them I imported the exported mixdown file (32 bit btw) back into Sonar, into my original project like bitflipper also suggested. That was a natural first step for me. :) And that's basically where I left everything and ran off to work. Now I'm home and I'm gonna start fiddling with the different export settings.

That can happen when you have a mono effect (like an amp sim) on a stereo track, or if you insert a stereo effect into a mono track (which forces it to become stereo internally). Essentially, you're adding two mono tracks together, which increases the level by 3 dB. This may not be your problem, but it's worth checking to make sure everything in the signal path is mono throughout.

I work for some studios and there is a STANDARD:
Every track, synth with plugins has to be bounced to an audio-track.
Then only the audio tracks are mixed.
First - this should solve the problem.
Second: This has also the advantage that everyone can repeat the mix 20 years later (without needing the plugins).