Halflife Alyx tested on 8 headsets included the pimax 5kplus,see here

VR content
22

It is indeed. Best not to listen to people who give off vibes of any sort of affinity with any particular brand.

(Kind of amusing when “fanboys” from every “camp” accuse of you of being a shill from any one of the others, just because you are not down with hiding any product’s warts. :7)

I wonder how Steve drew that, though (yes, I noticed you mentioned Tiltbrush), because that is not the perceived shape of the index FOV, when you’ve got it all dialled in, on that picture: The numbers are right, but you get a saddle type shape, that bulges in along the edges, in the cardinal directions (EDIT: …because the screen ends before the lens does, so to speak).

(EDIT2: Never mind – the answer is probably obvious: He likely looked in the direction he was drawing at any given time, and as we’ve all discovered, using HMDs; The pupil offsetting when one look around, means that you see farther in the direction you are not looking, and lose FOV in the direction you do look, because you effectively “move around in front of the keyhole you are peering through”, changing the occlusion conditions. (Thankfully the 8k/5k lenses are rather wide, so that doesn’t happen so much with those devices - at least laterally :7).

Even though those bulges are so much farther out than their siblings on the Rift CV1, I still feel they “box me in”, just like the CV1 did. Much of this is probably because in both cases they are in focus – The 5k+ focus falloff could actually be helpful here, when in small FOV mode, making the crop softer, and less intrusive, and at any rate going a little farter out, and ending convex. :7

Because I’ve got a low and prominent brow, I need to “swivel in” all large lens HMDs under that brow (kind of like a rock climber holding on for dear life under an overhang), to get close enough to the lenses for optimal optical conditions – this led me to the situation with the index, where I can do that and still get good focus when looking straight ahead, and up (thanks to aforementioned decent edge-to-edge clarity, which I for one just. couldn’t. get. with any previous HMD I’ve had, including my 5k+), and gain FOV downwards, where I am more likely to look, at the cost of FOV upwards, that is occluded by that pesky brow anyway. :stuck_out_tongue:

That was kind of neat; Too bad about the incessant stray light in the two-Fresnel-lens macaroon, and the fact that tilting down also pitches the forward vector down, so anything that does a raycast on facing direction, forces me to e.g. look down my nose at a UI gaze target, or otherwise roll my head back, to avoid swimming down, if in water. :stuck_out_tongue:

1 Like
23

he explains it at this timestamp

24

Hmm, yes, nothing that I didn’t surmise there, it turns out.

I’m assuming for now, that the specific aspect of his methodology, that produced the too round shape, is probably the guess I edited in, into my previous post (the bit about looking in the direction he was drawing)…

So what I get from the conclusion Steve seems to be drawing there, about that the index FOV feeling Pimax-small-mode-ish to him, was not in direct comparison with his Pimax, but relative to other sub-120 headsets, and the way the ehm… “interaction” between their per-eye views, especially on the vertical - and the width of the vertical peak, gives him kind of a sense of that silly little mask they put over the frame in some old movies, when the director wanted to show us the view of a character looking thorugh binoculars… :stuck_out_tongue:

25

Hey Risa I was just reading back your posts. But if I understand correctly your conclusion is to just start at your actual IPD and then just experiment, sliding it higher and lower, right? To get the geometry right, pretty much as you said here: https://community.openmr.ai/t/pimax-response-required-is-there-an-official-solution-for-eyestrain-and-ipd-issues-in-the-works/17913/9

However the vertical position of the headset on your head is at least as important in my experience to fine tune geometry distortion. Or in other words, to minimize distortion you’ll need to move the lenses to your distortion ‘sweet spot’ both horizontally or vertically. Agreed?

And I couldnt agree more to what you and @jojon said about how this is actually pimax task to educate their users about this. Well of course they can just ignore it but reviews as the above are the result and this is NOT good for Pimax at all.

1 Like
26

It’s the truth though. Pimax 5K+ is an extremely distorted mess, to the point where having a small FOV like on the Vive but with a clean undistorted center image, is less distracting than Pimax’ high FOV.

4 Likes
27

Man I have to agree. I just did again what I’ve been doing countless times. Adjusting IPD higher/lower, moving the hmd higher/lower/left/right/tilt. And if I mess long enough with it I can get it to a point where the world doesn’t feel TOO wobbly, but it still is somewhat wobbly. And then I switch back to the index, I just put it on and BOOM everything is perfect. And besides the distortion, the world just feels much more real. Maybe due to more sensation of depth?

At this point though I do consider that I might just have a bad unit. I have one of the first ones so it could be that the panels are just badly aligned, I don’t know. So I’m going to wait till I receive my 8k-X to confirm that.

But if the guys in the video see what I see, I can totally understand their reaction.

2 Likes
28

Same experience here. No matter how much I play with ipd and placement on face, the world bends as I move my head. It’s a zero immersion experience. And as soon as I put on another hmd, boom, rock solid world that I can get immersed in.

2 Likes
29

I bounce between 5K+ and Quest and I see it right away. Something is always a bit “off” about the Pimax no matter what I do to make it “normal” and so I just accept it and try to make the best of feeling immersed like I do in an oculus hmd.

I will qualify that by adding that the slight feeling of it being off is not a deal breaker at all, the 5K is still my preferred hmd and I continue to hold out that future versions of Pitool,and finally when the MAS and comfort kit are available, that the effect will be eliminated.

4 Likes
30

Well it’s at least comforting to know there are several of us out here, LOL. To me it seems we’re a minority, I think 70-80% of the users is actually quite happy about their hardware and would rate it about 8/10 (if it’s not being RMA-ed LOL). While I’d probably rate it 4/10. So I think this must almost indicate that we have just faulty units. Maybe badly aligned panels or something. I do know that I have the old panel version including black dots so that’s of course also not helping

3 Likes
31

Starting with the nominal IPD (i.e. the one you measured or get measured) is the safest way one can take, because one needs to start somewhere and the current design does not guarantee the exact value will be right. The chances are however it should not be far away.

Then the additional fiddling with an IPD setting is needed to align the eye depth with the view axis. Unfortunately, depending on the eye depth, the correct IPD setting may imply a positive offset (to the nominal value), which would make the lenses offset so much to the sides that the straight view will already suffer from the lens distortion (as the eye will be looking through the lens peripheral, when already in the “natural” position - i.e. looking ahead).

This is the reason why I suggested that eventually one needs to find a trade off.

Now, while I wrote that, Pimax added another way to tackle the horizontal fitting by using the software IPD offset. This allows moving the lenses in one direction (e.g. for compensating the distortion (blurriness)) and moving the IPD offset in the opposite direction, to compensate for the previous offset and thus keep the rendered IPD correct.

I never addressed a vertical placement (or the importance of), because there was not really much left to be set (this has changed with the “vertical offset” option, but only slightly). One needs to find the correct vertical placement himself by just trying.

The general problem with those settings (and tweaks) is that people sometimes find some exotic combination and think that this particular combination is the general solution, while they just found the workaround fix for poorly set or misaligned lenses and/or panels in their headset, or something particular to theirs facial features (eye depth, eye asymmetry).

This is the reason why these posts should be taken apriori with a grain of salt.

4 Likes
32

If I understood him correctly he used RoV tool to get the vertical and horizontal range and then superimposed the shape from Tiltbrush on top of it. I have two points:

  1. RoV tool has a small error in measuring the horizontal angles because of the way it measures them. This explains why Steve believes he saw more horizontal range than the headset renders.

  2. The diagram is displayed in “equiangular” scale, which is not what the human eye see normally. So this representation might be confusing when compared to the direct experience.

2 Likes
33

I still rate it an 8 out of 10 despite the slight feeling of the image being “off”.

34

I used to think like you in july 2019. Then in august Pitool came out and I haven’t seen distortion since.
I cant explain it really. Pimax is so random and this is the problem.

35

notice at the end of the article they don’t even mention the pimax. They went out of their way to exclude it.

1 Like