New shocking discovery 
today on soundstage, showing I’d never really thought about this too deeply: stereo speakers as usually operated are still a ways away from being the ultimate reference for realism of stage reproduction. The inter-aural crosstalk where each ear hears the opposite-side speaker is more a detriment than a benefit to imaging 
, as every ITD and ILD cue that corresponds to sound sources originally not present at the respective speaker’s equivalent position vs. the listener is messed up by the crosstalk, and this smushes many images toward the center, degrading overall realism and clarity.
In this view, cancelling the crosstalk as much as possible should improve the staging realism and image clarity, quite the contrary to what many DSP effects and player plugins have been doing for headphones in trying to create a more realistic virtual stage. 
A little intro (with sound demos) from one producer of commercial crosstalk cancelling solutions: https://youtu.be/lfQPnmy2pxI?t=183
Found out about all this by tripping over a link to this Archimago article:
https://archimago.blogspot.com/2023/10/stereo-crosstalk-cancellation-xtc.html
I see there’s been a single mention of this type of thing in the forum so far, but it’s been restricted to more of a product-specific thread and incorrectly called “speaker virtualization” software (the normal way speakers are used, i.e. without crosstalk cancellation, is not something you want to virtualize as-is, according to this theory): https://forum.headphones.com/t/anyone-try-theoretica-us-baach4mac-speaker-virtualization-software
LE:
BTW, this whole crosstalk-cancellation approach to sound reproduction with speakers has been dubbed “ambiophonics”. Had to throw this in to make searching for more information easier. 