How to change the directions of virtual cameras?

Hello guys. The pimax 8k has a 10 deg. canted design. If my undsertanding is correct, the L/R vrtual cameras also rotate outward accordingly by 10 deg.
Is there any way to modify the rotating angle of virtual camera? (from pitool or steamvr side?)
I am interested in the impact of such a dismatch on user’s perception. Thx.

Unless you intend to experiment with more granular mismatches, just run any application that does not work with canted lenses, without the “parallel projections” workaround enabled – that will immediately give you those 10 degrees off.

You could no doubt e.g. hexedit the angles in piserver.exe, or intervene somewhere else along the line, to spoof whichever invalid value you desire, but I wouldn’t know where the transformation matrices in question may be located.

…otherwise you may need to create your own test-app, with e.g. Unity, or Unreal Engine. :7

EDIT: By the way: I don’t remember exactly, but have a recollection that either the screens or the lenses are canted an additional 2.5°, on top of the 10 (EDIT2: …the lenses I think, as one can see by the wedge-shape of the frames they are fitted into.)

1 Like

Thanks for the info! This is exactly what I need.

Based on previous observation if the dismatch is above 5 deg., I will have double vision; If the dismatch is 1-2 deg., everything looks no difference (I guess it will impact perception of depth to some extent). So I am trying to find the boundary for double vision (some where between 2 - 5 deg.).

The 2.5 extra lens cant is new to me. Do you have source for this number?

Well, I would say it is zero, but the human mind is undeniably flexible, and will make the best it can out of anything it is presented… :7

Nothing better, I’m afraid, than my own clumsy trying to shove a protractor into the HMD, at a time when I had it taken apart, in order to modify the IPD range – hardly a precise figure. :7

…as to why it is that way: I don’t know, but it could be a similar purpose to when I did the same thing myself, with a little test box I botched together, where I did it in order to gain some ok-ish focus over a wider FOV, through a compromise which did not let the center axis tangent of the field curvature of the lenses I had graze the screen surface, as is usually done, but instead moved a bit closer, and did the cant, sacrificing a little bit of focus in the middle, and instead getting a ring around it, which had the sharp vision…