Renze,
There is some chance that the 10x50 has a greater inter-objective distance (IOD) than the x42 or x32 models, i.e., creating more retinal disparity. For monocular cues, 3-D perception is also enhanced by greater FOV, other factors held constant.
[See my experiment 12 yrs. ago.]
Incidentally, did you ever measure your NV's IOD when the instrument is set to your IPD? Whenever (IOD — IPD) > 0.0, the 3-D effect is enhanced.
Although I've never considered transparency being an aid to 3-D before, it is possible that it allows greater impact of subtle visual gradients. Certainly, as opacity increases gradients become less visible and depth perception would diminish. The general hypothesis would need to be studied in greater detail experimentally.
Ed