All these differences have seemed very minor, one of the three main testers sees the wobble on 8K. One of the three main testers sees the small discolored dots on the 5K+.
They all recommended the 5K+ overall at the start, but in general have always said that the three main pillars of VR are better on either headset than any others on the market, that is SDE, resolution/sharpness, and FoV. They all just give the edge to the 5K+ due to the sharpness, but I think for many even the slight reduction in SDE might be worth it, as long as the 8K remains better than any other HMDs on the market, and doesn’t have a terribly pronounced bug or other issue.
If you aren’t backer 200, but much higher, I would just hold off on making a firm commitment until the first 100 headsets start arriving and people start posting. If we see wide variance in responses, it may make our decisions much harder, as that could mean quality control isn’t up to snuff. But if everyone is essentially very happy with their new HMDs, better than anything else on the market, then you’ll be able to comfortably and happily choose between SDE and the ultimate in sharpness/resolution.
the topic is “World moving with you” phenomena, color pattern problems are a different thing and should be discussed in a separate thread
also the description from the Spanish folks and mrtv may sound similar but are different imho
for mrtv everything is fine as long as he moves his head slowly, his problem is only to be seen when he moves his head very fast horizontal and @LoneTech’s explanation was about that
the description from the Spanish is more general and look much more like lens problems, they complain it to be more general , not a jumping object it sounds more like a deformations, not just for fast movement, so might be a completely different problem
imho to even judge about that you need a lot of experience and the right vocabulary to get it transported in the right way, i clearly remember how that was a problem when mrtv started talking about the problem
its like a surgeon talking to a patient and at the same time tries to talk to another surgeon on how to proceed with the operation, all that Latin and stuff, no they are not concealing anything its just the necessity to be extremely exact about what the problem is and how to fix it, if you are sloppy or precise you might end up with the wrong method of how to fix the problem (pertty much the same when to hardcore games start talking about how to tweak a whole system with cpu/ram/gpu/bios and someone just using a computer at work stands near and hears that)
Pimax already said in the KS campaign that the same 8K panels where used in another application up to 90hz, Pimax never exactly confirm why the limit is 80hz except for stability so yes the rest is speculation
but if the 5K+ is using the same Anologix chip do the math…
I’m pretty sure it’s an operation that does not take much more than a minute (if you were talking binning, I misunderstood you :7):
Plug the assembled HMD, into the dual camera test rig. The rig would have parts shaped in such a way that the HMD lines up properly to the cameras, and locks into place.
The rig displays a test image on the HMD, compares it, as shot through the lenses, to an ideal output, and calculates the required distortion parameters to get from the former to the latter.
Optionally, the rig could try out the derivated profile, to make sure.
The rig flashes the parameters into the HMD nvram.
The technician removes the HMD from the rig, and passes it on, down the production line.
Could also be done per assembled, hmmm, can we call them: “eye tubes”? (screen+lens+the plastic pyramid they are set into), instead of the whole HMD.
Any distortion-related wobble effect, you would see as you track some object moving across your view, and it contracts and expands like in a funhouse mirror.
That dual camera test rig you describe would be a VERY EXPENSIVE custom built piece of hardware, with very high/demanding specs, as well as custom software. You would then also need to have profiles linked to the serial number of said HMD thereafter, and integrated into PiTools. It’s a not insignificant piece of infrastructure work. For all we know they’re doing this already, but such details on the industrial process aren’t released to us.
If you go search and watch some of the quality control processes that major manufacturer’s use on their cellphones (Linus did a good in-the factory one)… they have a lot of money invested in all sorts of sensors and scanners and stations to ensure quality. I hope Pimax does too, but we truly have no idea.
I know, I started it however anything can be related. It is how regression happens, e.g. did these dots appear at the same time as the wobble issue. Could be a link. Might not be.
Yeah, as FatalXception suggested we will have 100 reviews/opinions soon and that will give a broader picture of any common issue. Maybe Tested, Linus, RoadToVR etc will get hold of it in that batch too.
Yeah, I was thinking how software could do that. It would need to apply some serious image learning and pattern matching for the entire Fresnel area. The camera would have to move around too unless it can swivel like an eyeball.
It would have to swivel like an eyeball/scan like an eyeball, as that’s how we look through the lens ourselves, no point in correcting from angles where your eye wouldn’t be.
I would actually be very surprised if HTC took pupil swim into account at that particular point (although the lenses are designed with it in mind to begin with - it is ostensibly (EDIT: …as per Alan Yates, of Valve) the top reason we ended up with fresnels at all) - don’t think there is anything in the parameters (you can read them from your Vive) that sounds related.
With the 5k/8k things become more complicated, due to the, in several ways, more complex optical setup – not least of all that is sounds like the distortion over the entire lens can never be decribed using a single simple function.
Aye, based on progress I imagine it is not simple. Others have managed it though. The optical properties for the Pimax lenses are already known so any function correcting distortion would have that data to calculate from. It is just math, advanced.
that whats the idea of using eye tracking for correction is about
starvr has wide fov and eye tracking but its not clear that its really been used for correction, its just what people think
and for the moment that’s not relevant as there is no eye tracking from pimax available and there is nothing that shows that the tracking possible atm is fast enough for this and that the software (like streamvr) can use it, lots of code to write in the future
my guess would be that even if you would have eye tracking now it would at least take 4-6 month so see the results in games
but thats not helping here for this topic and i dont think pimax will have any offical statement about this
i guess as mrtv was skype’ing with the pimax founder that was something that came up and i guess that will be handled by pimax dev team
I guess HTC might have something custom built.
Heck - we’ve seen Pimax’s lens-glue-ing stamp devices on picture - wonder whether those are off-the-shelf devices, or in-house or commissioned work… :7
I wondered what you were talking about. I was not talking about eye tracking. We were talking about about how Lens Distortion Profile equipment can automate bespoke profiles and how it works, needs to work. At the Manufacturing plant.
i thought my reply was linked to the message, maybe i will quote more in the furure
the device in the factory for me was just fact (Alan Yates comments about the gearvr mod), nothing to discuss, we know that its used with htc vive and the doc ok homebrew device was on debate for a while in the vive reddit when the gearvr topics where a its high - the only way to get a good lens profile
tinkering by hand did not turn out to be of any use on a larger scale and every vive lens modder building its own measuring equipment …
Maybe so Lillo, but your well know position is in contrario to all spec comment made by companies and PImax, in this case the 8K would not be a 80hz nor the 5k+ a 90hz LCD headset.
So to rephrase Pimax have limited the commented specification of the 8K to 80hz because of stability problem like said in the risks and challenges section, “probably” (read speculation) link to the upscaler over heating or performing badly when higher signal clock are used
From the Kickstarter risks and challenges info!: ( @D3Pixel here I found where I saw the info on the 8k’s panels)
The display panel supports 90Hz according to it’s specification, and
it has proven a stable 90Hz in another project. The display port is
DP1.4 and the chip we used supports DP1.4 as well. So theoretically,
Pimax 8K can support 90Hz.
But the stability is much important because our HMD should work out
of the lab in thousands of computers condition. In the past two weeks,
we have done more stress tests from 75Hz to 85Hz base on V3 hardware
(latest DP 1.4 solution). Up to now, the 80Hz is confirmed stable. 82Hz
just pass the initial test standard. And now we’re trying to optimize on
85Hz and 87Hz. Since 90Hz is the upper limit of the whole system.
Several hardware reasons might make the refresh rate drop: the clock
rate, cable length, display adapter FPC, EMI desense issue. Contact
impedance from different Video Card’s DP connectors, etc.
however anything can be related. It is how regression happens, e.g. did these dots appear at the same time as the wobble issue. Could be a link. Might not be.