Don’t know whether this applies in your case, but any increase to the bitmap size that the game does “on its own initiative”, is not restricted by the SteamVR “maxRecommendedResolution” setting, which limits how high you can go with the SteamVR sliders (the actual values, that are recommended to the game, are limited, but the settings panel still displays the unlimited numbers, alas).
There are also several titles that appear to simply ignore SteamVR’s recommended render target - noteably many, many UE4 ones.
On a side note, that is probably relevant to some observations in this thread, it is worth keeping in mind that SteamVR Home, and parts of the Lab, use dynamic scaling, and will reduce quality (EDIT: automagically, in discrete steps), if they can not keep up the frame rate.
As far as I know, this pipeline doesn’t exist. The game relies on the OpenVR call GetRecommendedRenderTargetSize() which returns two integers pnWidth, pnHeight, there should be nothing else.
I don’t mean that isn’t true that in-game settings give different results, just the reason must be something else, and I guess it depends on choices made by the single developers.
Yeah. Maybe they use different textures for example (more detailed for higher settings) ? While the SteamVR setting would otherwise just scale the same texture. Or it could indeed simply be something like @jojon said. But like others noticed here, I always prefer to use ingame SS if they’re available, seems to yield better results.
the point is that while SteamVR clearly tells me the rendering size, the one in the game doesn’t. What does 1.5 mean in DCS? x * 1.5, Y * 1.5 or 1.5 more pixels?
In Elite: Dangerous is even more trivial because you have:
Yep its all quite confusing I never understood the differene between supersampling and HMD Quality in Elite Dangerous, more than the fact HMD Quality seems more effective
“HMD Quality” in Elite Dangerous is an additional multiplier, on top of Recommended Render Taget Size, and unlike “new style” SteamVR supersampling settings it factors the axes, so x2.0 is X*2 and Y*2, and consequently four times the number of pixels to render (…or 400% in newspeak SteamVR – doubleplus unconfusing :P).
“Supersampling” in Elite Dangerous also multiples the dimensions in the exact same way, and renders that larger bitmap, but then the game on its own downsamples the result to what was originally asked for, prior to that increase in size, before handing the images over to the VR runtime (e.g. OpenVR). This option is also present for rendering to monitor, and that’s really what’s it’s for. It can be useful in VR if one want to push past a combined HMD Q and SteamVR SS of 500%, because at that point, texture filtering will begin to skip samples from the game’s rendered images, rather than including them all in its calculations, so in that case, it can be useful to have the “top load” of detail “prebaked” into the input image – too bad one can not get reasonable frame rates at that point.
(EDIT: …should probably mention that to my eyes, just as logic would suggest, there is no differerence between SteamVR at 100% and HMDQ at 2.0, and SteamVR at 400% and HMDQ at 1.0 – they produce the exact same thing.
What does make a tremendous difference to me, is to opt to use SteamVR’s older filtering method (“allowSupersampleFiltering” : false), which results in a significantly sharper output, at the cost of not smoothing out aliasing as much as the newer one (EDIT2: Some prefer the opposite tradeoff, and will take blurry imagery, if that means less aliasing -. these people are wrong. ))
As @jojon already explained the games usually calculate the SS factor per one dimension, while SteamVR calculates it per pixel count. Which is confusing especially if it is not stated anywhere.
Have you increased the max recommended res in SteamVR?
That’s especially important for the Parallel Projection option in Pimax. If you haven’t changed the max (or haven’t checked it recently, since it sometimes gets reset), SteamVR has a maximum render size of 4096, so VR is running at a lower res than it should be. It makes a huge difference.
Be glad it works for you, but my experience corresponds with what they say (on 5k+). Hz and fps are 2 different things, on 72 Hz I feel uneasy and “something is not right” after some time, even in cockpit games. On 90 Hz all is well even if fps is lower. I have no clue if it will be different on 4k or not though (e.g. if 120Hz is really needed).
At the end it is about what percentage of population you are targeting. I’m sure they did they research/studies back in the time. Did Pimax do any research on high enough population sample? 90 Hz might be enough for 95%, while 72Hz is perhaps good enough for 80% (just guessing, no clue what exact numbers are) which they decided is not good enough for product.
However high fps (for which you need also high Hz) is important for fast paced games. In Eleven table tennis it is obvious that even 90 Hz/fps is not enough (not for comfort, but for fast balls and rapid movements/reactions with paddle). Unless one is playing just cockpit games or slow paced experiences the high Hz/fps is mandatory.
This stuff varies from person to person. I for one have never been able to feel terrible FPS unless it always has inconsistent frametimes every moment and lag or drops below 30.
Wasn’t this confirmed as unnecessary, because the PiTool settings take care of it instead? I did set this once, a long time ago, but not recently.
Also there is a default settings file (which will be overwritten by updates), and a user version, which will not - so naturally it’s the second that should be edited, if at all.
I never understood the differene between supersampling and HMD Quality in Elite Dangerous, more than the fact HMD Quality seems more effective 
))
