A simpler explanation is that waves oscillating perpendicular to one another cannot interfere with each other at all. The x axis oscillates at full magnitude, and the y axis does too, regardless of the relative phases. The only thing that's weird here is that the third filter at 45 degrees can "remove" the 90 degree difference in polarization, but it's not that hard to understand, and can be demonstrated with a much simpler experiment by just inserting the 45 degree polarizer between the ones at 0 and 90 degrees. Bonus nitpick: current polarized 3d glasses filter circular polarized light, not linear. If you add two of those together, you'll get linearly polarized light, the angle of which depends on the relative phases. I think (I haven't tried this so I'm not sure) if you put your RealD lenses on the two slits, you would not see interference patterns with the naked eye, but you would if you looked at the screen through a linear polarizing filter.