Pimax needs to implement the equivalent of Valve's vertical column correction into Pitool (But horizontally)

A while back, my Pimax headset was plagued with horizontal lines when I moved my head. It was very annoying, and I was trying pretty hard to get it fixed. I tried software adjustment through Pishell, which helped some but didn’t fix it fully, and if I’m remembering right I even got two RMAs. Alas, the problem was never fully fixed. On the third RMA I just decided to sell it new in the box and get an 8k X.

Now the problem appears to be on the Index as well! https://www.reddit.com/r/ValveIndex/comments/cjhndo/index_headset_vertical_banding_through_the_lenses/

However it’s vertical lines instead of horizontal lines. I’m assuming this is because the Index screens are placed vertically instead of horizontally. After a while, Valve added something called column correction as an option. This gives you a range you can set from 5 to -5 which improves these lines.

This is similar to adjustment in Pishell however, Pishell is less user friendly and is not as readily available. Even with the column correction on the Index, for some it was never fully fixed. I think this feature needs to be implemented into Pitool on day 1 of the Pimax 8k X. At the Orlando meetup, I noticed these lines on one of the HMDs. I was hoping it would be fixed on the 8k X, but that appears to not be the case. It was better than before though, I think. Anyway, if these lines are still appearing on the new headsets, I think it’s important that Pimax implements column correction into Pitool. Factory calibration should also be improved, if nothing has changed.

9 Likes

This is a very common issue called “pixel inversion”, and many LCD headsets has this to a certain degree. Also a lot of fast refreshrate PC gaming monitors are dealing with this problem.

Example:

Quote:
"Inversion is also called pixel walk, typically found on higher refresh rate monitors due to voltage control going from positive to negative and back, hence the name inversion (polarity inversion ). It’ll look like the pixels are moving though they’re not, thus why it’s called pixel walk:
Source (one of the posts):

As the Valve Index and HP Reverb panels has vertically aligned subpixels, the inversion or “lines” are vertical, while other VR headsets like Pimax has horizontally aligned subpixels, and these are prone to get horizontal lines. You can clearly see the subpixel alignment in some of my TTL photos from before CES.

I see this problem on both my Valve Index and my HP Reverb V1. At least last i checked, some week ago.

The issue is only seen in movement, in other words when u move your head, and once u stop, the lines disappear. Many people will never even see this, but if you have a good sight and look closely, you can see it, if your headset is affected by it.

This issue can apparently be corrected by voltage adjustment but mostly its fixed with software tweaks.

I have reported the pixel inversion (as Ive seen in it myself on some samples) and discussed this with the engineers already and they are working on a solution to completely remove this problem. Let’s cross our fingers!

8 Likes

Aha, thank you! I was having a hard time figuring out what this was about!

Looking at this video, it seems the pixel inversion effect may be closely related to the RGB stripe artifacts I notice slightly more with especially sharp inputs (VR Toolbox sharpening filter enabled) and the SteamVR filter bypassed (PiTool Render Quality >=1.25).

Personally, I have not noticed this effect specifically. Probably because previous VR headsets including the 5k+, have other problems that are too severe for me, and because the Pimax Vision 8kX may tend to hide such effects better.

@SweViver
Please ensure any filters intended to remove this effect in PiTool can be toggled on/off.

2 Likes

It also depends on the game you play. In Elite D, the only time I ever see this problem is when I’m next to a very bright star. Even then it’s hard to see and easy to ignore.

2 Likes

Not sure whether it’s exactly this effect - but it might be along the same lines: It is most disturbing in Virtual Desktop - edges are “flirring”, apparently I can’t keep my head still enough to stop them from doing that. Highly disturbing, particularly when using the VD internal supersampling at high levels. Without supersampling, lines and particularly contours around characters get more blurred, so the effect isn’t as apparent.
It’s a little sad, because I could technically even read texts on webpages with high SS levels in VD. But with the flirring lines it drives me crazy fast. Antialiasing doesn’t seem to help on first sight.
So if there is a way to mitigate this, I would be highly interested to test it! (My current hope is that the working mitigation strategy will be called 8k X…)

Man I hope so. This issue was very annoying to me on the 8k.

1 Like

I saw exactly this on the 8K+ when I tried it at VRDays, it was the only and very annoying thing I could notice. I hope the engineers can find a solution.
This is not present on the Valve Index any more.

Unfortunately, I can confirm this effect does happen with the NOT FINAL 8kX unit I am using.

Patterns 5, 6a, and/or 6b, at some distances, viewed through Virtual Desktop, show rainbow color artifacts. These remain visible with the headset not moving, on a desk, pointed directly at Virtual Desktop.

However, the observed artifacts move to patterns 2a, 2b, 4a, 4b, 7a, 7b (disappearing entirely from 5, 6a, 6b), when the headset is rotated near 90deg. This leads me to conclude the artifacts are indeed pixel walk, or at least unique to the horizontal direction.

I now realize this is the same as the RGB stripe artifacts I observed in the Pimax 5k+, so at least I can attest this is no longer anywhere near as severe.

http://www.lagom.nl/lcd-test/inversion.php

Previously, I just assumed these were purely spatial moire patterns, I had not expected a temporal component…

6 Likes

I’m not sure, does mirage have the final or not final 8kx unit, asking for a friend :wink:

1 Like

Ha ha, yes, I have been emphasizing that often. But really, I am making a distinction here.

  • ‘Not final’, fixable (usually software) flaws. Every 8kX that ships should be compatible with a corrective update.

  • ‘Pimax Vision 8kX’, expect at least the early units, if not all to behave this way. DSC for instance, is a hardware requirement, for good reason, and is not going away anytime soon for any high-resolution headset.

I am also very much grateful for the opportunity to get a head start on all this extensive testing, configuration, and actual use of the 8kX headset. I definitely don’t want to create any confusion over any inaccuracies for @SweViver and company to deal with.

1 Like

We have been reporting this issue since March 2019…

I have provided many screen pictures in May 2019 but we have never had any feedback from Pimax after Kevin had said the tech team was “not able to reproduce the issue…”

https://community.openmr.ai/t/8k-horizontal-lines-problem/16652/123

And in August 2019 the thread as just been locked, despite last answer was from someone saying the replacement of his HMD had not solved the issue…



It is very important to note the issue occurs only on 1 panel for me (as for others who also noticed this issue with their HMD), the other panel is NOT affected AT ALL ! :face_with_monocle:

So I don’t see why you would have to try to solve this with a “software workaround” when it is proven the current software is perfectly able to display images not affected at all by this issue :roll_eyes:

As others I have tried the flicker parameters with pishell tool provided by support with no success. Not only it was not solving the line issue on the affected panel but it was also not creating the issue on the not affected eye.

So I am still under the feeling this is a hardware issue. Why is the other eye displaying perfectly fine ? In the other thread I was suspecting hardware interferences coming from the integrated gyroscope, as ROTATING your head (up/down rotation) is what makes the issue occur .

So it “could” be a shielding issue. IIRC back to the CRT days I think I remember seeing a similar issue when using a cheap poorly shielded VGA cable.



Now in the old thread Kevin also claimed later that over 5 reported cases, 3 were solved through remote access and the 2 others have received a replacement from which one said the issue was just “reduced”:

https://community.openmr.ai/t/8k-horizontal-lines-problem/16652/209?u=neelrocker

One sure thing this line issue has a HUGE impact on visual quality as it HALVES the vertical resolution of the affected eye as soon as there is a vertical rotation of your head (which happens almost all the time when you use the HMD).



Here is the original (and now locked…) thread:

https://community.openmr.ai/t/8k-horizontal-lines-problem/16652?u=neelrocker

and here the 2 sets of pictures I had provided in this old thread:

https://community.openmr.ai/t/8k-horizontal-lines-problem/16652/128?u=neelrocker
https://community.openmr.ai/t/8k-horizontal-lines-problem/16652/127?u=neelrocker

It’s not the gyroscope. These lines appear with a moving image, not head movement alone.

1 Like

Oh I have never thought about it that way but you are right indeed :slight_smile:

I have just done the following test with virtual desktop:

  • with follow head unchecked, rotating my head up/down makes the lines appear (on white text for exemple)
  • when checked (=the desktop mirror now stays fixed in the HMD) there are no more lines during vertical head rotations
  • checked + moving the mirror window up/down with mouse, the lines appear again (they are just less easy to detected compared to when you make the image move by rotating your head)

So it can’t be interferences from the gyro indeed, but more likely the “pixel inversion” as said by sweviver.

3 Likes

Yea, this problem was quite the annoyance on the 8k. It really bugged me, and Pimax should really try to completely eliminate this issue before releasing the 8k X. I don’t want to go through the hassle of getting replacements only to see the issue again. The 8k X may have super low or basically nonexistent screen door effect, but what does that matter if horizontal lines plague your screen during movement.

1 Like

This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.