Indeed “reasonably cheap” is very subjective.
Well, it also gives you the option of extending the X’s cable by a few meters… and it’s stable. Not cheap, but not huge either. I find the water block for my 4090 more expensive. Nearly 450 euros. 
crazy
Godd Luck!
3 Likes
Don’t put too much stock into the exact details of what’s been said to customers about this issue by Pimax and nVidia. This is a complex engineering problem, and we on the sidelines aren’t going to be told a play by play on the internal details of what’s going on. Is it nVidia’s fault? Are they fixing it on their side? Is it Pimax’s fault? Are they fixing it on their side? The answers to these questions are probably not simple, and it also may be a moving target as the two companies work on the issue.
What matters from our perspective is that they are working on the issue. I suggest that we let their engineers do their jobs and not second guess them from the sidelines based on the scant second hand information we hear about.
We should not assume that Pimax’s situation is actually the same as Varjo’s situation. It’s likely that they are both caused by the same changes nVidia made for the 40 series, but how exactly this affects their VR headsets doesn’t appear to be the same, and the remedies may not be the same.
The Aero seems to have been hit much worse by the changes nVidia made. Reportedly this caused problems that were immediately obvious and untenable across their entire customer base. In some ways this is actually better because reproducing the problem was easy and obvious. Whereas for Pimax the effects have been more subtle in most cases and harder to reproduce and analyze, seemingly only affecting a small portion of their customer base. That actually makes the issue relatively more insidious and difficult for Pimax to resolve.
That’s very interesting. I don’t know about “smoothing the signal”, but the presence of the repeater in the chain does clearly have effects on video transmission. It might be causing a different mode to be selected or something along those lines. This would be inline with the observation that this issue doesn’t seem to happen to all Pimax customers. The problem may only occur in specific scenarios, and the presence of the repeater may change the scenario such that it’s using a different branch of the code that avoids the issue.
That’s also inline with the theory that this may really be an nVidia issue and other companies are just implementing workarounds that avoid it. Their “fixes” may just be avoiding using the particular mode that has the problem. Putting the repeater into the chain might accomplish the same thing.
All of this is just speculation though, but it’s plausible. The upshot is that based on your results, experimenting with the Lindy repeater is worth giving a shot as a potential immediately available workaround.
I don’t think you’re right about this. This is almost certainly lag on the video side and has nothing to do with tracking. Specifically, it is a lag somewhere between reprojection of the frame and display on the physical panel. It becomes human perceivable by making fast turning head movements, but the lag is actually always there.
1 Like
FYI: I have the exact same “ghost frame” issue, Gainward 4090 Phantom, I7-13900K, 8kx. I didn’t have this issue with my 2080TI. When I swap the gfx cards and leave everything else untouched, the problem appears only on the 4090. I will issue a ticket to both nvidia and Pimax as suggested here.
2 Likes
Sorry but no, I have tested this. There is no lag in the video. Quoting my previous example, I flick a switch on my joystick and the switch in cockpit flicks with no lag. It moving my head causes lag in the tracking of that movement.
As such the whole video feed is not delayed, just head tracking movements.
1 Like
Maybe it is new chipset problem? I have z490 and 10700k and no this problem, I paly only ED and don’t now how about in other games, while the 4090 didn’t improve the fps at all in this game, same was 3090
1 Like
It’s been established across a pretty wide variety of hardware so far.
Also elite dangerous requires very slow/minor head movements so it would be practically imperceptible at 75hz.
But objectively there is latency in the video. I believe what you mean is the latency is below your perceived threshold. You’re testing this with human perception alone, and so some things about how human perception works need to be considered.
The threshold of human perception of latency for the cockpit switches is not the same as the threshold for human perception of moving your head in VR. The latter is much more sensitive.
So if the amount of latency falls between these two thresholds, it would appear to you the way that you’ve described even though the latency is actually constant for both of tests.
The VR rendering pipeline accounts for these differences in human perception thresholds. When you click a switch in game, the game will represent this in the next game frame. This will appear instantaneous to the player. But head movements are another matter. Waiting until the next game frame isn’t fast enough to avoid causing nausea in humans. Head tracking data is sampled near the end of generating the current game frame and reprojection is applied at the last moment before display.
This is why I say that the latency being introduced by the issue of this thread is probably occurring after reprojection has already been applied. Reports are that this is related to changes nVidia made to their video driver for the 40 series. If there was latency introduced to the head tracking, that is through a completely different subsystem (USB) which is not associated with the video driver.
What you have perceived in your testing is consistent with this hypothesis due to limitations in human perception. What you suggest could also be true just based on those tests alone (since humans can’t tell the difference), but it doesn’t match with the other things that we know about this issue. So that’s probably not what is actually happening.
2 Likes
So I have been testing this for a month, extensively, but your opinion, having not observed it yourself, is that you know better than me what I am seeing?
Your nonsense is not helpful. Disregarding people’s testing and making assumptions on what I am able to perceive is loose conjecture at best, and rather insulting given all the testing I have been doing.
If you don’t have anything helpful to offer, then please ignore my posts.
1 Like
Ok. I understand your frustration. But being offensive to people who are genuinely trying to help you - even if you vehemently disagree with their observations- is really counter productive. It seems that you have already dismissed any possibility that does not entirely blame Pimax.
If that’s what you truly believe, then there is no point posting here because either none of us will be able to help. If you want to vent, fine. But try to be civil. Nobody here has been deliberately rude or mean to you.
2 Likes
I was perfectly civil and didn’t tell anyone to shut up etc, I simply said ignore my posts. I was not being offensive.
I am however not welcoming of my testing extensively then being told by someone who hasn’t seen it, that I am not observing what I am observing. What is the point of sharing my testing results to be told those results are ‘wrong’.
I haven’t at all blamed just Pimax, I have complained about the poor responses I am getting from Nvidia.
I have shared lots of details of my testing so that we can work together to find a solution. Not to be told that my testing is not valid or the results I report are not actually the results because I can’t perceive the difference in my results…
1 Like
Further to this I have tested the latest 2 Nvidia driver versions, including today’s release. I’m both the lag is worse, not better. The least lag driver version is still 522.25.
1 Like
I have also found the stuttering worse on more recent drivers. You could be a little more courteous in general.
1 Like
To be fair, when you’ve dropped big bucks on a new GPU and you get this kind of issue- it’s understandable that sometimes the frustration affects the tone of your writing here. Especially when you may feel that readers aren’t really ‘hearing’ what you’re saying.
This is exacerbated by the fact the the problem is very technical.
So I apologize for mistaking your tone for rudeness.
I wish I could be more help.
1 Like
I have observed the issue myself, and I have performed my own testing, too.
I also have considerable relevant professional background. My posts are intended to share my insight, research, and findings, and to try to help.
2 Likes
I don’t see it as nonsense at all. The fact there is no issue on the 30series speaks for itself. It is indeed related to the video driver changes for the 40series cards.
He never dismissed what you perceived. He had only added context to the differences between head movement and input latency with switch presses. While somewhat related they are indeed not the same.
I had a p1 4k and Motion to photon latency was not good as you have motion blurring. However in Ethan Carter and other games I found using a gamepad. for turning while moving vs using head to change direction. You would not get the ghosting.
I do really appreciate your frustration with buying the latest “Best” video card and having it have bad issues that are still not resolved.
Hopefully the wait on a fix will not be too much longer.
Fair enough, no offence intended either way I guess. This is an emotive issue as I have a 4090 and 2 Pimax headsets unusable for a month now. That equates to nearly £4000 of gear I cannot use. So it is extremely frustrating.
I am doing as much testing as I can to try and get as objective as possible information to help with debugging this. And so far no progress at all.
Whereas it might only be a sub-section of users have the issue, for those that have it, it is pretty catastrophic. And in the meantime I am letting people down as I cannot maintain my commitments as part of a virtual squadron.
So apologies if I am a bit reactive.
2 Likes
I know your pain bro, I bought this card exclusively for VR and now I can’t play at all, I’ve been so f*cking bored it’s unreal.
I also can’t believe I’ve gone through two 4090s both acting up, yet somehow this is still considered a niche issue.
Hopefully we can get you back to your squad soon 
3 Likes
Cheers dude. So you’ve had 2 4090s…that must be doubly frustrating. Were they the same model?
1 Like
Not that it seems to be model specific