If the audio is generated from a soft synth, then that may also be a reason there is a difference. Especially with drums, as they sometimes 'humanize' things. Meaning, small differences in performance every time, like a real drummer. If these are rendered tracks, imported into each version, then this does not apply. But, when generated in each sonar version, there might be a difference that causes the tracks not to 'flatline'. In superior drummer, there are some settings to tweak what it actually does, but changes in velocity and tempo are definitely on by default.
Also, some effects (especially the ones that emulate vintage equipment) add some random element as well.
While these will cause a null test to fail, it would not mean a lot of difference in sound...