Reading around posts on Reddit, it appears that many believe that the source of reported excessive distortion in some Pimax headsets may be caused by a manufacturing flaw and discrepany in the lenses on some devices but not in others.
People have the concern that if the same optics are used in the 8KX, the 8KX will not be well received, and that the advantages offered by the screen upgrade in the 8KX are lessened and not worth the upgrade cost if the same lenses from the 5K+ are used.
I have the same concern, is there any chance a developer from Pimax can comment on whether or not the lenses offered in the 8KX will be improved on from the lenses used in the 5K+ and 8K? @anon74848233@PimaxUSA
Maybe not so much a matter of the lenses as such; More their alignment, as fitted, exacerbated by the fixed-screens-with-moving-lenses fundamental design. This construction, with parts that move in relation to one another, makes it difficult to factory calibrate each individual unit, because even though you can derive coefficients for each end of the IPD sliders, and interpolate for intermediate positions, there will inevitably (unless we’re talking premium build) be enough mechanical “play”, that you can not reliably expect an altogether negligible deviation. (This is one of the many things one might consider relevant for any “enterprise” version of the product.)
Still, I expect it will be just as “good enough”, as it is with current units. As long as one have the HMD fitted as designed, there should be “acceptable” ranges of distortion…
You might have more luck posting on reddit asking them about it if they don’t leave a reply here in a couple of days, judging by their communication recently on issues which they don’t like talking about
Just need to wait on what they say (Thanks for the invite). If you read up on Valve’s explanation on Index Design (assuming not too much is just marketing). It might be a trade off of High FoV design with face shapes as well affecting individual perception through the lenses.
Ie it’s possible with 1 headset tried by a variety of folks some will say very little distortions while some others will report large distortions.
But it is also true how flexible the manufacturing manufacturing process will affect level of varience from one unit to next.
Which goes back to things like Rift S good for a wide variety but is not a one size fits all.
Why it’s important to get headsets into an enviroments for folks to be able to try or have a good no questions ask for a satisfaction gurantee ie 2weeks or 30 daya upon receipt. Even with an advertised small restocking fee.
But this isn’t a problem which has to do with being in the incorrect IPD range. That would be fair to say, as it is for the Rift S and the HP reverb, IF Pimax headsets didn’t have an adjustable IPD.
Unfortunately this is not the case, and many people in the ideal useable advertised IPD range (60-65mm), reported seeing intense distortion on some headsets, but not on their friends headsets or on replacement Pimax headsets.
This has lead many to believe on reddit that the source of the problem appearing for some users in the 60-65mm IPD range, but not for others in the same range, is not due to slight differences in facial shape causing the mass distortion, but a manufacturing error in the production of the lenses, or the placement of the lenses,
which is the real source of the extreme distortion issue reported by users in similar IPD ranges to users who do not experience the issue at all. (Maybe there was a bad production batch, or batches, of 5K+ and 8Ks which had the affected lenses.)
We need to know if they received a thicker face cushion, it could also be caused by the shape of their face.
To be sure we would need these people to try to the headset of someone not experiencing distortion.
As for the “High FOV design” distortion tradeoff, if it exists, may only be the case with Pimax’s fresnel lenses (due to either lens production errors or alignment errors), as extreme distortion has not been reported by StarVRone testers or XTAL users.
But Xtal uses aspheric lenses, StarVR uses fresnel. Neither of those headsets have the same distortion issues to the same degree as Pimax headsets.
Which leads me to think that either a lens redesign for the 8KX should be made (new plastic aspherical lenses or better fresnels), or that Pimax should attempt the same rendering style as StarVRone (Rendering two 4:3 or 5:4 images per eye, and blending them together for a less distorted wide FOV image with 4 images rendered onto two screens) to reduce the extreme distortion in medium and large FOV settings.
The new model of StarVR and XTal do not make as many cost compromises as the Pimax 8k/5k - that’s what allows them their advantages. Other than the more expensive, and properly calibrated, optics, their built-in eye-tracking potentially enables dynamic pupil swim compensation, for less distortion, which everybody seem to assume they are in fact using, but I have seen nobody with any plausible authority saying this for a fact.
…there is of course also the matter of difference, to consider, in sample size of people who have tried the various devices, and can report on them… :7
Yeah I said extreme distortion. I believe some is to be expected on XTAL due to the rendering pipeline also stretching a 4:3 image into a 16:9 image. But distortion may be visible in extremes on some Pimax headsets due to incorrect lens alignment or a lens manufacturing flaw affecting headsets in some batches, which is not surprising as Pimax is starting out and the 5k and 8k are not enterprise level products.
However, visible distortion may be significantly reduced if two 4:3 images are rendered and blended together for each eye like StarVRone, instead of stretching one 4:3 rendered image into a 16:9 widescreen image, but I believe it requires canted lenses as well as canted displays to do.
I don’t think so. Take a 4:3 image and stretch it out to 16:9 to fit on any widescreen display and you will notice the same distortion.
Rendering two different 4:3 or 5:4 aspect ratio images per eye, from different angles, and then blending them together, should siolve the extreme distortion issue.
That or setting Pitool to render either one 16:9 image per eye, x2 blended 5:4 images per eye, or x2 blended 4:3 images per eye, depending on each games chosen aspect ratio chosen in its rendering pipeline may eliminate the issue. (As most VR games are encoded to render images in 5:4 or 4:3 aspect ratios, beause 100FOV lenses, and 1:1, 5:4 or 4:3 displays are standard in consumer VR HMDs.)
Oke but i don’t think heliosurge was meaning extreme distortions as a trade off for wide FOV.
I agree the excesses in distortion reported with some people who tried everything are manufacturing faults. they should RMA instead of shouting how bad pimax headsets are
Sure there may be some to be expected, but things may be done to reduce that distortion issue signicantly, either by identifying the lens manifacturing flaw (if there is one), make efforts to have the lenses better aligned to reduce issues occuring from incorrect lens placement, and look into changing the rendering pipeline to render with two blended 5:4 or 4:3 images per eye, and making cantered lenses for the 8KX if required with a new rendering pipeline, if changing the rendered images to x2 blended 5:4 or 4:3 images per eye works to show a significant reduction in distortion in preliminary testing.
I think if enough of us are vocal about it, Pimax might consider it. It’s important for their success and the success of the 8K X that we do.
Many would be more than pleased to hear and would be willing to look past previous Pimax issues if they found out they changed their rendering pipeline to seriously reduce distortion (which doesnt cost them barely anything to try, no production costs), and/or that they’ve been hard at work developing better plastic aspheric lenses.
Many people in the ideal useable advertised IPD range (60-65mm), reported seeing intense distortion on some headsets, but not on their friends headsets or on replacement Pimax headsets. This was also reported by many users who had reported trying many different face foams and seating options, distances from eyes etc but still reported seeing the issue on their headsets. Also some of those users still reported that their initial headset were affected but their friends or their replacement headsets were not?
Thing is: The games are not restricted to any particular picture aspect ratio - they will render whatever FOV, and size and dimension bitmap the VR runtime (e.g. SteamVR) asks them to, which can be seen in the different sizes you get, when switching PiTool between its three FOV options, as well as Parallel Projections On or Off.
The distortion you get when projecting a large FOV onto a single large flat viewplane is problematic, in the sense that everything gets stretched out toward the edges, which is wasteful, but from the point of view of the camera, which is placed to correspond with your eye in relation to the lens and screen, that projection is every bit as correct, as if you had cut the viewplane down the middle, and folded the outward half toward you (same deal as when they paint extremely tall letters on a road, to make them look correct from the low position of the driver’s seat, or when you render… any FOV actually - not just a huge one, to a monitor; Move to the right distance from the monitor, and the apparent stretching will “go away” – I am pretty sure this a considerable part to why console builds of games are typically lower FOV than computer ones; You sit farther away from the TV than you would from your monitor and keyboard, and therefore see less through the window into the game word it constitutes (That, and it reduces the amount of stuff the weaker hardware needs to evaluate per frame)) – the latter situation will just be more optimal, because you do not need to unnecessarily render as many pixels per degree of FOV, out there in the periphery. (Now: If we had also had physical screens that curved this way, we could potentially also get a sharper image out there, thanks to being able to align them with the field curvature of the lens, but that is a separate matter, as is the fact that if we are going to approach 180 degrees per eye, we are going to have to think about rendering more than a single viewplane per eye (EDIT: …assuming we are sticking with the classic type of transformation used with rasterized realtime 3D graphics). :7 )
oke maybe they are looking into those things already, would be nice to now what they are actually doing, but they have lost interest to inform us on anything it seems
Because they don’t want to set themselves extra deadlines or create unrealistic expectations. Everything they say they get held to account for like a promise. I do agree they should practice replying to issues more without saying too mch or even anything, at least giving an acknowledgement that certain issues/things are being looked into.
Yes they need to examine their production & QA procedures to insure headsets are in tighr tolerances to insure one batch closely matches the next.
However your statement only in pimax fresnel lenses is premature. Only Pimax has a wide FoV headset out in the wild in large numbers. Xtal & StarVR One has no real presence out there in any volume. Both StarVR One & Xtal are also utilizing Eye Tracking.
When the Eye Tracking module is finally released if Dynamic Distortion Correction is in place. We might find that Pimax is more readily used by an even larger audience (meaning those who currently can’t use it well).
Only time will tell. Valve stated specifically problems with Huge lenses & facial profiles. As demonstrated that some are good with default Face Foam, some need overall thicker & some need it thicker at top or bottom.
