vr Настройки Bignewy Reverb G2 для высокой частоты кадров

[BIGNEWY]

5 minutes ago, Snoopy said:

Thanks Big, my G2 is arriving today and will use your settings as a start!

Enjoy! I am sure after some tweaking you will find what is right for you. It is a great headset.

[USA_Recon]

14 hours ago, Bullant said:

 

Absolutely mate, was just adding some input from the guys who built the headset. To my eyes, and I think that's important because everyone's eyes are different, I see a massive difference in clarity between 54% and 100%. The hornet MFDs are a blurry mess at 54% (without leaning in like I had to on the Rift S), but at 100% I can read them from the normal seated position.

 

With my 2080 I am never going to hit 90fps without turning everything down to minimum, so I'm working with settings that give me a stable 50fps so that reprojection can do its magic.

 

Will be a different story when I get my 3080, at which point I will definitely try your settings.

 

They really have made it confusing in Steam VR SS settings.

 

The first setting is a global percentage. The second allows you to set for each specific game. However, unlike most settings where the game specific setting would override the global setting, the steam SS settings are actually multiplied, so:

 

Steam SS = Global Steam SS x Game specific SS

 

If you set global to 54% and game specific to 120% then you will get an actual SS setting of 65%

 

What the reddit article is saying is that the G2 was designed to need a certain amount of SS for optimal clarity, and that SS happens to be the 100% setting. I guess that makes sense then as to why 100% is the default value!

 

I confirmed last night at 52/54% I was getting 90fps (F-16 quick mission caucas fighter mission).  I switched to 100% and it dropped to 80fps.

 

I would rather though lose frames than have 'shimmering' and DCS has too much shimmering - even BN attests to that shimmering.   Regardless of settings, if I look at the 3 or 9 oclock and roll the aircraft it's not smooth.   I don't know if that is a DCS issue or a G2 issue

 

Note: agree with above post, use what works for you - I don't notice any barrel effect, but I do believe what is being said.

Edited December 9, 2020 by USA_Recon

[BIGNEWY]

22 minutes ago, USA_Recon said:

 

I confirmed last night at 52/54% I was getting 90fps (F-16 quick mission caucas fighter mission).  I switched to 100% and it dropped to 80fps.

 

I would rather though lose frames than have 'shimmering' and DCS has too much shimmering - even BN attests to that shimmering.   Regardless of settings, if I look at the 3 or 9 oclock and roll the aircraft it's not smooth.   I don't know if that is a DCS issue or a G2 issue

 

Note: agree with above post, use what works for you - I don't notice any barrel effect, but I do believe what is being said.

 

Good to hear you are also getting high frame rates, I am not seeing the stuttering on my system however, it is very smooth. 

Only time I have noticed some was if I was recording at the same time.

[markturner1960]

1 hour ago, imacken said:

You're starting up a whole hornet's nest here.  There are many threads all over the internet about this topic.

Some facts.

1) the G2 native res is 2160*2160 per eye.  So, your WMR setting is correct.

2) all VR headsets produce what is called barrel distortion, and there is a certain amount of allowance to compensate for that in the SS value. This is normally taken to be around 1.4 at an axis level, so 1.4x1.4=2, or 200% at a pixel level.

3) in SteamVR, the pixel count for the G2 at 100% is 3172x3096, and at 48% is 2196x2144.  48% is the closest you can get to 'native res'. So, the 100% level has twice as many pixels in total, or 1.4 (approximately) times per axis.  

4) what value of SS you use is down to you, but HP engineers have apparently confirmed that 100% is the correct default value.

5) this is different to the G1, where SteamVR SS reports 100% at pretty much the native resolution, and this is where a lot of confusion has arisen.

6) the G1 100% is equivalent to the G2 48% in terms of pixels and it takes a whole lot more resources to run the G2 at 100%.

Personally, I feel you need to test this out for yourself.  I have asked before how we see the barrel distortion in practice, as I never noticed it in the G1 running at 100%. Can anyone see this distortion in practice?

There is no simple answer to your question.  higher values - up to a limit - give clearer images, but just use a value that is good for you and use that as a benchmark.

 

 Thanks Imacken, that helps. What is the point of having the multiplier app specific setting in Steam VR then? As I currently understand, in this case, The Steam VR home setting of 48% is the G2 outputting at it's native resolutions nearest setting of 2196 x 2144. But this is not recommended by the HP engineers, who say it should be 100%.....If you want better clarity, at the expense of performance, you would simply increase this percentage, yes? So why complicate things by adding another percentage slider for app specific settings? Also, if I set the steam VR home resolution at 100%, I don't get great performance, despite having a very powerful rig. lord knows how people with lesser spec rigs and cards are able to use the G2 at all at anything like what it is supposed to do......If that's the case, then the G2 certainly looks like it has performance that simply cant be tapped yet and is likely to remain that way for some time....

[dburne]

7 minutes ago, markturner1960 said:

 Thanks Imacken, that helps. What is the point of having the multiplier app specific setting in Steam VR then? As I currently understand, in this case, The Steam VR home setting of 48% is the G2 outputting at it's native resolutions nearest setting of 2196 x 2144. But this is not recommended by the HP engineers, who say it should be 100%.....If you want better clarity, at the expense of performance, you would simply increase this percentage, yes? So why complicate things by adding another percentage slider for app specific settings? Also, if I set the steam VR home resolution at 100%, I don't get great performance, despite having a very powerful rig. lord knows how people with lesser spec rigs and cards are able to use the G2 at all at anything like what it is supposed to do......If that's the case, then the G2 certainly looks like it has performance that simply cant be tapped yet and is likely to remain that way for some time....

 

Having per app settings allows one to easily have different settings for different games, depending on how they run on their system.

I leave my global resolution at 100 percent and use per app setting for each game I run.

 

I have also started setting per app to 100% as well just to try it, and it is looking pretty darn good on some with Motion Smoothing enabled.

I think Steam VR with the G2 handles Motion Smoothing even better than Oculus ASW does.

Edited December 9, 2020 by dburne

[imacken]

The point of the app specific multiplier, is that you don't necessarily want the same level of SS for every game you play, so that gives you flexibility.

Yes, the 'default' 100% is really tough to take. Even with my 3090/10900K combo, I have issues, but only in DCS.  All other titles I run never drop below 90fps.

[netizensmith]

To add confusion, on a Reverb G1 100% is pretty much the actual native res of the headset i.e. not a supersample. If I set my slider to 54% it shows a lower res than the G2 does at 54%.

 

EDIT: I see Imacken already mentioned this so consider this as confirmation as I have a G1.

Edited December 9, 2020 by netizensmith

[Nealius]

How's your spotting/clarity with the low SteamVR resolution? I'm currently running 100% in Steam and 1.2 in DCS and my spotting/clarity is very bad. 

[imacken]

2 minutes ago, Nealius said:

How's your spotting/clarity with the low SteamVR resolution? I'm currently running 100% in Steam and 1.2 in DCS and my spotting/clarity is very bad. 

On a G2?  Either way, you shouldn't use both SS processes. The standard advice is always have PD at 1.0 and do all your SS adjustments in SteamVR.

[Voyager]

Little add-on to cut the GPU Loading: https://docs.microsoft.com/en-us/windows/mixed-reality/enthusiast-guide/using-microsoft-edge#monitor-and-input-handling-issues

 

This is how you can tell WMR to stop generating a bunch of extra virtual monitors whenever you launch it. 

[BIGNEWY]

45 minutes ago, Voyager said:

Little add-on to cut the GPU Loading: https://docs.microsoft.com/en-us/windows/mixed-reality/enthusiast-guide/using-microsoft-edge#monitor-and-input-handling-issues

 

This is how you can tell WMR to stop generating a bunch of extra virtual monitors whenever you launch it. 

 

Thanks for that, I am running 3 normal monitors already, when WMR was opening it was creating 3 more on top, registry edit worked I will test over the next few days to see if I can squeeze some more out of the resolution and keep the higher frames

[MrRoper]

I was playing around with other applications and supersampling and made the mistake of leaving my render resolution at 200% :O sweet jesus it looks amazing. 30 fps on a 3090 not so much though!

[imacken]

57 minutes ago, Voyager said:

Little add-on to cut the GPU Loading: https://docs.microsoft.com/en-us/windows/mixed-reality/enthusiast-guide/using-microsoft-edge#monitor-and-input-handling-issues

 

This is how you can tell WMR to stop generating a bunch of extra virtual monitors whenever you launch it. 

Thanks for that. Interestingly, on my 20H2 version, the default value was 0 not 1, and 20H2 wasn't mentioned in the list. Still, it works, so we'll see what difference it makes.

[BIGNEWY]

2 minutes ago, imacken said:

Thanks for that. Interestingly, on my 20H2 version, the default value was 0 not 1, and 20H2 wasn't mentioned in the list. Still, it works, so we'll see what difference it makes.

Same for me after manually adding the registry entry the value was 0 but it did work and the extra monitors were not present. 

[javelina1]

same with me, running 20H2 and after adding manually the value was set to 0 as well.

[dburne]

32 minutes ago, imacken said:

Thanks for that. Interestingly, on my 20H2 version, the default value was 0 not 1, and 20H2 wasn't mentioned in the list. Still, it works, so we'll see what difference it makes.

 

Same here. I did not notice any difference.

Edited December 9, 2020 by dburne

[Taz1004]

6 hours ago, imacken said:

The point of the app specific multiplier, is that you don't necessarily want the same level of SS for every game you play, so that gives you flexibility.

Yes, the 'default' 100% is really tough to take. Even with my 3090/10900K combo, I have issues, but only in DCS.  All other titles I run never drop below 90fps.

 

I guess the question is why doesn't the Per Application Setting doesn't "overwrite" the global.  Why does it "multiply"

Like Nvidia settings, per application overwrites global.  This should be proper way.

All in all, SteamVR implementation is just bad.  Not "what" they're implementing, but "how" they're implementing it.

[K1nsal]

On 12/7/2020 at 7:14 PM, Taz1004 said:

 

That's incorrect.  If you do the math, you'll find that 100% in SteamVR is actually 140% of native resolution (in number of pixels).  That is the supersampling Steam deems "appropriate".  That is why you can't reverse it exactly to your native resolution.  You can't reverse 140% reset to 100% back to original 100% without going into decimal value.

 

There is no different set of computations.  Resolution specified is exactly what is being rendered by your GPU and displayed on your headset.

 

It is not a bug but Steam deciding to relabel their own supersampling as 100% is misleading and confusing a lot of people.  Plus this varies depending on system and headset.

It not super sampling per say it Barrell correction so you get close to the native rez across the entire panel instead of just looking at the sweet spot(middle) which is one to one when you under sample like the setting detailed at the top. Which I don't think it a big deal in flight sims when most you attention is hyper focused in one area most the time.

 

 

Edited December 9, 2020 by K1nsal

[imacken]

3 minutes ago, Taz1004 said:

 

I guess the question is why doesn't the Per Application Setting doesn't "overwrite" the global.  Why does it "multiply"

Like Nvidia settings, per application overwrites global.  This should be proper way.

All in all, SteamVR implementation is just bad.  Not "what" they're implementing, but "how" they're implementing it.

Yep, that makes sense. Would be so much easier for users to understand. 

[Taz1004]

36 minutes ago, K1nsal said:

It not super sampling per say it Barrell correction so you get close to the native rez across the entire panel instead of just looking at the sweet spot(middle) which is one to one when you under sample like the setting detailed at the top. Which I don't think it a big deal in flight sims when most you attention is hyper focused in one area most the time.

 

 

 

Again, that video is saying same thing I've been saying.  I know why it is being rendered higher than native resolution.  Actually almost everyone who has VR set knows this.  But that is not the issue we were talking about.  I love how these videos talk to you like a child.

 

Lens distorts the image.  Renderer reverses the distortion.  Pixels in the center gets magnified and you'll see more jaggies.  Renderer renders images 1.4 times larger to compensate for the magnified pixels.  Then downsampled back to native resolution.  When you render larger and downsample, that is called supersampling.

 

Let's say you buy OC graphic card which is factory OC'ed to 120%.  But their software says 100% even tho the clock speed is higher.  Wouldn't you be confused whether the GPU is overclocked or not?  So you ask the question "Why does it say 100%?"  But they keep answering "Clock speed is higher because it is OC card and it is not a bug".

Edited December 9, 2020 by Taz1004

[Voyager]

1 hour ago, imacken said:

Thanks for that. Interestingly, on my 20H2 version, the default value was 0 not 1, and 20H2 wasn't mentioned in the list. Still, it works, so we'll see what difference it makes.

 

Looking at the timestamps, I think not is just that the article was written prior to 20H2 being released, and has not yet been updated for it. 

 

If I had to guess, someone in software decided to make the default, once you create the registry key to be zero, probably on the assumption that if you are adding it, it's probably because you want to turn that feature off. 

 

I am glad they added that control to in though. Those extra monitors have been driving me nuts the entire time I've had my Reverb G1.

[worldrecordaviator]

Great stuff. Thanks for putting in the work. Any thoughts on default.vrsettings - renderTargetScale & motionReprojection Mode?

[imacken]

1 hour ago, worldrecordaviator said:

Great stuff. Thanks for putting in the work. Any thoughts on default.vrsettings - renderTargetScale & motionReprojection Mode?

Don’t touch that file any more as it tells you in the file first few lines. All MP settings are within the WMR tab of SteamVR dashboard, and SteamVR itself. 

[worldrecordaviator]

2 hours ago, imacken said:

Don’t touch that file any more as it tells you in the file first few lines. All MP settings are within the WMR tab of SteamVR dashboard, and SteamVR itself. 

Thanks imacken, must've missed that. Would you recommend leaving them as I have set (renderTargetScale = 2.0 & motionReprojection = auto) or returning them to something else (don't remember what stock was)? Or doesn't matter?

Edited December 10, 2020 by worldrecordaviator

[imacken]

6 hours ago, worldrecordaviator said:

Thanks imacken, must've missed that. Would you recommend leaving them as I have set (renderTargetScale = 2.0 & motionReprojection = auto) or returning them to something else (don't remember what stock was)? Or doesn't matter?

Here is the default file.  Note what it says in the first few lines! 

Have you seen this Using SteamVR with Windows Mixed Reality - Enthusiast Guide | Microsoft Docs 

default.zip