Hi guys. I have an 8K X on order and it will be my very first time using a Pimax wide FOV HMD.
Because of the recent changes to the Pimax forums, searching google just gives bad links so I haven’t had much luck researching this.
Could someone please explain, or direct me to an explanation, of the three different FOV settings on these HMD’s? What is relevant to me is the 8K X, but I imagine the FOV settings are the same for all of the different current models except Artisan.
More specifically:
Assume there is no super sampling applied. What is the approximate rendered resolution after the mask is applied for normal and small settings? If this is a dumb question, just explain why please. =)
When using Normal FOV mode (or small) does the in-game FOV get stretched horizontally a bit to utilize more of the displays and lenses? Perhaps rendered resolution is decreased by (this is a simplified example only, I know this number is incorrect!) 20% when you go from wide to normal. Does that mean that the physical screens in the HMD are now 20% solid black, or is it maybe only 10% black with a more stretched image?
I had a third question but can not remember it! I will come back and edit/reply if I remember what it was.
They are basically the same. There are currently 4… my angles are from memory and are only approximates…
Large is ~170° horizontally, which takes longer to render a frame and tends to have some distortion around the edges.
Normal is ~150°, which is what most people use.
Small is ~130° and is useful for under-powered GPUs.
and “Potato” whos name is intended as a joke; it is about the same FOV as a standard VR headset, so it might be useful for games which are incompatible with wider FOVs. It’s main purpose is to demonstrate the limited FOV of other VR headsets.
You can see the currently rendered resolution in SteamVR > Settings > Video. Set the SteamVR SS slider to 100% and read the displayed res. Then switch to Pimax and change your FOV. You then need to restart SteamVR, when you change the headset res.
The image is always stretched non-uniformly to counteract the physical lens image warping. Smaller FOVs do draw more black around the edges of the image.
In addition there is a “mask” option in PiTools which has even more black pixels, which are actually undrawn, so this option can be a framerate speed-up. I think some games are incompatible with the setting.
Only reps from Pimax will be able to answer your resolution questions accurately for the 8KX in Normal and Small FOV at Pitool 1 and SteamVR 100% @SweViver
In game FOV will match the FOV your headset requests.
The lens is fixed and with each FOV setting the FOV is not stretched or compressed relative to other FOV settings. The means an area of the screen is blacked out in smaller FOV settings. However if the 8KX is similar to the 5k+ the rendered resolution when changing from one FOV setting to another does not scale in direct proportion to the FOV being displayed. Of course this is not and issue as this is all relative to the distortion profile and everything is presented correctly on screen but in any case you ideally always render more pixels than the display has because the way a distortion profile must account for lens shape means some areas of the visual area will first render more pixels relative to other areas. This means to achieve 1:1 pixel output in the lower areas you will always need to render more than 1:1 pixels at first. When you switch to a lower FOV you will always render less pixels but you will also render proportionately less pixels each time. From a performance perspective this is a win because it means you can switch to a lower FOV with no loss of visual clarity and see a performance increase greater than the relative FOV difference would suggest due to rendering significantly less pixels comparatively in the lower FOV setting.
Not that it is a big issue because with Pimax the biggest surprise will be how smooth it performs at lower fps. Unlike other VR headsets the Pimax can dynamically run at an fps lower than the refresh rate without any loss of smoothness, without stuttering and without issues like screen tears or hanging and without the need of motion smoothing. In fact motion smoothing seems to make the experience worse not better. This is one of the joys of using a Pimax. You can play a game at much higher setting for much better visual quality than possible in other headsets and not worry about fps drops. A member on iRacing just posted about his experience updating to the latest Pitool and firmware in his 5k+ 2026 BE, he was very excited to be able to set his headset yo 110hz and just forget about fps as the headset can run right down to 60fps with no noticeable loss in smoothness or visual irregularities. If you are coming from another brand headset you will understand how much of advantage this is over the stutters that happen with small frame drops and the slideshow that typically happens at fps in the 60’s.
There is no “mask applied” for the different FOV configurations. Changing the FOV simply instructs the application to render into a different FOV. Which means larger FOVs requires more work on GPU as more pixels and more objects are rendered.
There is no “standard” rendered resolution as the rendered resolution is completely up to the application (as long as it renders the correct FOV) to choose. What SteamVR shows in the UI is recommended resolution, which is a kind of suggestion, which expresses the headset manufacturer’s hope that at this particular resolution most users (with varying hardware) would get satisfying experience (performance/quality wise).
The most important fact to remember is that rendered resolution is in no way directly related to the display resolution.
No, just smaller part of the display is used. The FOV the headset can display is defined by its optics, which does not change when you change the FOV settings.
I know there is an option to toggle the mask on and off, (Hidden Area Mask: on/off) so I assume you are saying the mask is the same for all FOV settings?
As far as arguing the “real” FOV you can’t really define the physical FOV of the HMD as seen by the wearer’s eyes. The number will vary significantly depending on the shape of the wearer’s face, inset of their eyes, how tight they make the strap, etc.
The only thing you could do is to create a standardized test that would give you a “relative” number for each HMD to allow a somewhat more objective FOV comparison. If you were only talking about the FOV software side, that makes sense. They really ought to make a standardized test so buyers could compare apples to apples FOV numbers. List it as something like “FOV for average user” and have it be a specific rig used to test it. If nothing else, one of the VR websites really should do this. In fact, I think I’m going to make exactly that suggestion to my favorite VR news site to see if it might be a practical idea.
That said, the FOV spec claims by everyone seem to be bonkers. Reverb and Index are both listed in many places across the internet as 115 degree FOV. That is laughable! A week or two ago I had them both side by side and spent six hours switching back and forth (in the same setting using DCS instant action) to compare readability of fine print, clarity, ability to spot distant objects, stereoscopic effect, etc. and I’m confident the Reverb had the smallest FOV of any HMD I’ve ever owned. Index is the largest I’ve owned so far. The difference was not subtle!
That is the reason why I was not talking about physical FOV, but “rendered” FOV. If you check the link I posted above (HMD Geometry Database | Collected geometry data from some commercially available VR headsets.), you will see what it means. While it cannot guarantee that the user will see the whole rendered FOV, it is still the best I can collect about the headset in pretty formalized way. And it also guarantees one thing, the user cannot see more than what is rendered.
If you check the link (and the visualization of different FOVs,) you will also see what the hidden area mask (HAM) means in relation to the FOV. It is the area which is not visible due to the optical properties of the headset and therefore not needed to be rendered. The HAM is just further refinement of the rendered FOV (which is by definition rectangular) to save some rendering resources.
Reverb has smaller rendered FOV than Index, but neither is close to 115° (see the link).
Risa, the rendered FOV isn’t what I’m so much interested in. I’m interested in how much of my actual peripheral vision is the VR environment vs. what is black/dark.
I think that for immersion or presence the actual perceived physical FOV is more important. I’m aware of the importance of the rendered FOV as well, but that’s not what I’m looking for here.
There is no way to get the “perceived physical FOV” from the headset alone, because it depends also on user’s facial features in particular eye depth (or eye relief). So it is highly individual.
Plus, if the application renders the scene at particular FOV and your eye is closer (to the lens) than for what the distortion / optics was calculated (designed), you would perceive higher physical FOV because of how optics works. In other words the rendered FOV (assuming the full rendered FOV was visible before) will be stretched over the physical FOV giving you a false perception of the bigger FOV which will however display only the rendered (smaller) FOV content.
On the other hand, if you get your eyes further from the lens (than for what the design was calculated) you will see the original FOV in smaller physical FOV and again the scene gets deformed (now it will look smaller).
Both these effects are the consequence of the optics and the fact that the headset manufacturer can make the system work for only one (model) eye position correctly.
Which means, measuring the physical FOV, even if you have tools/apps for that, is tricky business.
I am not sure what do you mean by “absolute” perceived FOV, my reply was commenting on “perceived physical FOV” you mentioned previously, which I intuitively defined as an angular range where you are able to see the scene (or where you can see the rendered image).
It is hard to know what he is talking about but in that last post it seemed he was asking what the physical FOV is. I actually wanted to try and measure this the other day but could not find my protractor. I considered that as we see edge compression perhaps our physical FOV is smaller. But then I worked out how we can dial edge compression out so now I’m not so sure.
I’d imagine the physical FOV ideally matches our rendered FOV. If every headset did this then we know the answer. Even in the Pimax I would estimate out physical FOV to be about 160 degrees based on my experiments.
Absolute Physical FOV is what is impossible to measure because it varies depending on the physical characteristics of the person wearing the HMD. You cannot say that HMD “A” has “X” FOV.
What you could give is a “relative” physical FOV where you have a test that measures the physical IPD of each different HMD when worn by the same (artificial) person. This would allow the various HMD’s to be compared against one another in an objective way in spite of the problem described above.
The physical/perceived FOV is what is going to make people feel like VR is closer to reality.
In a perfect world these two things (physical FOV and rendered FOV) would be equal.
.