Well yes, Dietz, I'd accept that my post above is somewhat "off-label" for this particular forum - apologies.
But "off-label" in terms of practical audio engineering? Well, maybe, if we're to be super-pedantic. The fact is though, my approach works. The results I'm getting are engaging and credible. Why not try it out for yourself in practice, instead of trying to pre-judge it? Eventually I might get around to producing an mp3 demo if I find some spare time.
I'm curious about your statement: "the whole idea of a binauralisation device is to encode the final mix". Well that's not exactly valid in Apple's world where, for example, on every mix channel in Logic Pro - except Stereo Out - the user can choose to use the built-in binaural panner option (ever since Logic 8 about 15 years ago).
My tweaking of Side-amplitude of reverb (in order to fit each reverb sector as nicely as possible in with the binauralised dry signals), is hardly a million miles away from MIR 3D's mathematical manipulation of mic Mid and Side signals in order to produce and distribute the 3D surround sound components to appropriate speakers. (Please do correct me if my theory is off here.) Hate to repeat myself but this approach of mine works nicely in practice. And don't get me wrong, MIR 3D reverb works nicely too; it's just that I prefer algorithmic reverb.
And as for the outcome of tacking a binauraliser such as DearVR Micro onto MIR's output mix, I know you know that a few of us have reported the results as being, sorry to say, not especially praiseworthy. My explanation of that outcome, back then and still today, is that the inevitable crosstalk between simulated speaker channels tends to degrade the psychoacoustic cues essential to binauralisation.
Now although I'm not interested in re-visiting old arguments, I do reserve my right to defend my opinions when necessary. But in any case, can we not simply agree to differ? And for my part, I'll do my best to post only in 'on-label' forums!