Pimax 8K & 5K+ Performance Boost & Improved rendering with new update!

the best possible setting is to find the best resolution (piplay) so as not to fall below 50% in steamvr (recommended value)

Ah but NVIDIA Turing variable-rate shading could be…
https://www.reddit.com/r/hardware/comments/9g0ppv/interesting_new_features_on_the_turing/

“The fun part of variable rate shading, is that it looks extremely easy to implement for VR. So easy, in fact, that it could be a driver toggle that automatically works for all vr games”

2 Likes

Can someone please clearly address what is said in this video? The auto-translate on this one is total garbage.

And seriously- if people start throwing the “scam” word around again at this point, Imma lose my goddamn mind. This forum is a big enough shitshow as it is. PLEASE anyone speaking German- just give us a reasonable TL;DW on this, ya?

5 Likes

I am afraid you don’t miss too much on this particluar 10 minute clip of Thomas. He starts off with the picture comparison, which in my view suggests that the old PiTool version produces a better picture (look at the word Established) while saying that it is absolutely the same to him, “Can you see any difference ? I can’t !”. And then he actually adresses exactly that section where I thought it looks worse now in v.76 and says that blur is likely from holding the camera a bit differently - cool, so forget about your impressions from the lens-shots, they rae random quality. Bottom-line seems to be that he when wearing the headset could not see any significant differences between the PiTool versions.

The remainder of the video is him ranting about the differing resolutions and that he cannot get it to fit exactly the values of his previous benchmark resolutios. Not very helpful, because there may be good reasons for Pimax to play around with the resolution (which resolution will bring the best efficiency between render load and quality improvement with the upscaler & lenses involved? Try & error, isn’t it).
And the main point for all of us is - okay, so if you can get away with only using 3/4 of the previous resolution and get the same picture quality, wouldn’t that mean that we get more FPS and maintain the same quality ? He is aggravated because he dislikes the benchmarks to not be fair this way - I don’t give a shite about being fair, I just want Pimax to use every dirty work-around at their disposal to get me the best picture with a reasonable framerate, and if my PC feels utterly abused by how they achieved it and sulks in the corner of the room for the rest of the night, I couldn’t care less. Foveated rendering will also be utterly unfair - the resolution outside the fovea will be appaling and a disgrace, and still all of us want it today rather than in 2020 - because it doesn’t matter.

19 Likes

That helps. I don’t feel any better :slight_smile: But that helps. Thanks :+1:

1 Like

Again, I think they pursue the wrong way to do benchmarks. I created a thread on that topic but nobody replied.
That may because I am completely wrong or surprisingly it seems common that people do not think it through and have no idea what they are doing. In my opinion problems like that are very foreseeable.
SWE did the right approach in his last video (If I am not the one who is wrong, which I don’t know since nobody seems to like discussing it) but it is hard to compare it with previous results.

2 Likes

In voodoo´s video it shows from minute 1:20 2 comparison pictures, whereby the second is clearly more blurred than the first. His comment: I see absolutely no difference. Ok…

That’s not fair what you’re saying. Like @VoodooDE says, the upper part looks different because of the photo. Like he says it’s hard to make 100% correct photo’s because you have lens focus, lens angle, lens blur and lens sweet spot. Obviously the upper part is not like it is rendered in the HMD, that would be just crazy. He’s talking about the lower part in the photo. According to him that’s pretty equal.

1 Like

I’d encourage you to take a step back and look at the big picture. Ideally, you would want things done differently in several ways, but the testers are all independent “contractors” with different equipment and frankly, this is not their primary mode of employment. What we are dealing with is a scattershot approach, and we just have to look at everything and piece a picture together. It’s dirty, organic and full of mistakes. The only other thing I can offer, is that if you are in a position where you are pressed to decide (like me) just don’t. Wait till there is a preponderance of data to drive you one way or the other.

1 Like

SWE was smart. He used a Pimax 1, Steam 100 as a primary reference point.

  • Then he experimented to find the lowest subjective quality thresholds for his setup and noted them as reference (and fps)
  • Then he experimented to find optimal subjective thresholds for playability and noted them as a reference (and fps)
  • Then he experimented to find upper subjective thresholds for quality and noted them as a reference (and fps)

That meant that he had 4 major reference points at which to compare to. So when Pimax changed the values, he could then find out where the boundaries had shifted and what that meant for overall quality and performance.

It is clear that Sweviver put in a lot of effort into his benchmarks.

20 Likes

voodo says

old pitool at 1.0 and steam at 60% has the resolution at 5443x2418
pitool v.76 steam at 60% has the resolution 2668x24418

if the new pitool give less resolution at 60% of course the fps and perfomance will be better.

he shows to images but says they are equal, but in my opionion i think the second its more blury because the resolution its worst.

I dont know if voodo tried 100% steam and pitool 100% in both pitool version, but i think he should try and compare if the visual quality its the same but with much better perfomance.

maybe even with less resolution, because the render is better the headsets have better image quality and perfomance.

1 Like

Or you could just record the reported rendering resolution in SteamVR (if available) and make it the reference point. Then just setup PiTool and SteamVR SS in a way the rendering resolution matches your target. From @VoodooDE video it looks there is a clear relation between FPS and the total number of pixels rendered.

1 Like

legend. thanks for translation

1 Like

The rendering resolution changes when Pimax makes updates…and it is hard to target exactly the same resolution. This is why Thomas was annoyed. (apart from the other factor that his chosen percentages were rendering at a much lower target resolution)

1 Like

Well he made already load of through the lens photos and is well aware which aspects may deal with the trickyness to do those photos. So I very much trust his judgement on this one, even if the photos do not look alike for us. He also states, again, that it is a sob to do those photos. I also agree on his theory that of course the new way to do the Res-Math is leading to higher FPS. The results though (little less RES) should not result in that much more FPS than shown by SWE. So I guess its a mix of both. Pitool sending a lower Res on 1 + other (real improvements).
And again this way of benchmarking is not leading us anywhere!

1 Like

On my monitor, the lower part is also not the same. The right picture is smeared, all key letters are more fuzzy. I believe that he observed no difference in HMD, but the pictures he used to prove the point were not chosen well.

2 Likes

He may have missed the point that if he saw absolutely no difference, and the newer update resolutions were lower (thereby increasing performance) then it means the updates made improved quality levels as it needs to do less work to provide the same result.

3 Likes

Agreed, he should have shot a better 2nd pic

Sorry, I almost never take a step back, except I know it will take me 2 steps forward later on. I very much look at the big picture, and I encourage everyone to reveal my flaws that I am unable to see myself at times. But I need to know where I am wrong with precise criticism. They have the same equipment to measure FPS, I’m actually suggesting a way to make it easier and proved successful for me many times, which still does not mean that it is better.

I know, I saw the video. My point was about how to set up (somewhat) reliable reference point. The only danger I see is if Pimax starts to change transmitted resolution to the headset (or to be more precise the active area of the input resolution), for example instead of rendering full 2560x1440 they will let the PiTool to render only center 2200x1200 (because what is out of this area is not visible through the lenses anyway).

This way they could reduce rendering burden, while still maintaining the same visual quality. Then using only the rendering resolution as the reference point becomes useless, (at least for the quality), because you could have the same quality while using different rendering resolutions (one would just “better” fit into visible area of the panel). In such case, the testers would also somehow need to assess the image quality and its eventual degradation (or not). Which seems to be much more difficult.