Tweaking the HP Reverb

if the computer is in the process of rendering a a frame and during this process takes the unfinisched frame and tries to show it, it can end up in two different parts of pictures shown together.

Not sure what you mean by this.

Tearing results from the way an image is drawn on display devices, which is from top to bottom. A complete image is not instantly drawn across the entire screen; it has to fill every pixel from left to right, line by line downwards. When the display is out of sync with the output device that is throwing frames at it, in the case of a game being the graphics card, you often get two different frames being drawn on a single screen draw. The top part will be of the newest frame, while the bottom will be the old frame. The line between the two is the tear, and its position can vary depending on what stage the screen draw is at. V-sync and Variable Refresh Rate technology ensures that only one frame is displayed per screen draw, so you never see that tear.

In the case of a camera filming a screen, because the camera is capturing frames that aren't synched to the display output, you'll see lots of these tears. However, they're displayed more prominently as thick black bars scrolling down the screen. This video is a prime example:

As you can see, that tearing is very different to image judder. There's also a more detailed explanation that does a better job than me here:

https://en.wikipedia.org/wiki/Screen_tearing

Well, that's my understanding of what a tear is - can I ask what you think it is?

Tearing results from

I know what tearing is. I've been in the home Cinema bussiness for 20 years and had a lot of problems with video rendering software between the years 2000 and (cirka) 2010.

What you saw on the former 'vid I made was not tearing. It was the problem that comes from trying to catch 45 frames per second source with a 60fps recording.

One of three frames of the source will be chatched two times on the recording. This will make the Movie flow not smooth as the third frame will be visible a longer time.

I managed to shoot the video of the flight you wished with my gopro, at Full HD and 120fps. Still, the source in the VR headset is 90fps and the fram rate of new explicit frams is 45fps. Theoretically, we have a bit of same problem but there is a much smaller percentage of the frames catched 3 or four times. The Youtube video doesnt look as nice as the source, they convert the frame rate to 60fps.

I think it can be seen in the video that it doesnt stutter and that there is no gosting.

If not, I still have the 120 fps source and it vould be put in some kind of cloud service, and viewed directy on your screen after downloading it, there is a need for you to have a 120hz monitor though. (or, play it at half speed on a 60hz monitor).

Did you by the way notice that bn880 solved his stutter problem(or problem of not getting reprojection to work) by updating Windows to 1909 ?

Thank you so much for going to the trouble of filming that. It does indeed appear smooth, so I'm going to capture mine now for comparison. Can I please ask what your SteamVR, WMR and game settings are? Are you using the VR shaders mod? This was captured on a Reverb, right?

So, after 30 minutes of faffing about, I can definitively say... capturing through-the-lens video is an absolute bitch of a job! After about 20 attempts, I got one half decent capture on my GoPro running at 1080p 120fps. And the bad news is that when I watched it back, it didn't really show how prominent the ghosting is in real life. There were a few brief glimpses of it, but nowhere near the same as what I see. It's particularly bad when flying over that town at about 700ft and looking left 90 degrees - every single structure is blurry and fuzzy, as they have a reprojection hovering behind them. My in-game counter showed that the minimum frame rate was 45fps, bouncing around up to 60ish, but only scraping below 45fps once or twice.

I can upload the video if you like, but it doesn't really show what I'm seeing. For some reason it seems to blur motion out, so rather than the stuttery/ghosting that I see, buildings just have a faint halo-ish effect around them.

Anyways, I've tried every single thing you can think of, and then a dozen more ;) From something as simple as updating Windows, to testing the different performance of every combination of the new WMR settings, I've run out of options other than to wait for new hardware, or a dramatic overhaul of the game's VR performance. I'm not phased that I have to wait, as I love the game as it is in 2D. In fact, even if the image in VR was smooth, I'd be hard pressed to go to VR due to the very poor visibility, and the loss of all the detail that you see in 2D. Flying at 3440 x 1440 @ 100Hz looks absolutely stunning, and I'm seeing so many small details that weren't possible when flying in VR. But I'm glad you guys are enjoying VR, as you'll be able to let me know when it's finally ready for my annoyingly demanding eyeballs! A huge, HUGE, HUGE thank you to the guys who spent considerable time discussing it in this thread - it's nice to see some folks are up for a good debate!

I had big problems taking a single photo with my iphone to show the reprojection-indicator.

Gopro, I used my iPhone connected to the gopro so I got the videostream visible on the iphone, that made it much easier to find optimum position an FOV setting to get it clear. Probably very lucky, it only took that one first try.

Anyway, as long as you can believe me/us that DCS actually plays smooth without stutter and absolutely no ghosting we dont really need to see your video.

There is one thing that we would need to know, if the reprojection really is active. If I disable reprojection I only see the FPS number steady when flying where the system delivers steady 90fps. In all other cases it will vary constantly between 90 and lowest fps it bogs down to.

With reprojection on it should only show 90 or 45, or a lower value when it doesnt reach at least 45fps. For brief moments it can show values between 45 and 90, when it switches between reprojection or true 90fps.

All settings is per the Thuds guide at:

https://vr4dcs.com/

Windows 10 version 1909. The guide at vr4dcs.com said 1903 a few weels ago, so I guess that version also works. I actually think that make sure you have version 1909 would be the first point to success. Means going in and reading version no number in windows system.

Windows mixed reality, I use highest setting of quality(I dont think it matters in gaming though, the steam vr and the game settings is setting the levels).

Theres a choise of 60 and 90 fps, of course use 90 if trying to get settings like I do ( and the guide at vr4dcs.com).

In nvidia control panel there is a setting for pre rendered frames. Should be set at 3. All other settings default to begin with. There are some settings that help some people but makes problems to others so they should not be messed with initially. Basically, the idea is to get it working good with smooth gaming first, probably some settings a bit low, and when it runs good then increase settings step by step until the individual mix of smoothenessand quality is fullfilled. After this, maybe start trying the odd settings that may be helpfull or harmfull to fps.

”Windows Mixed Reality for SteamVR” should be i stalled via Steam. One setting needed should beta-public version or something like that. This is to properly see the correct resulotion settings: 100% equals native of the reverb.(actually not exactly native but 2207x2160).

”SteamVR” should be installed, and NOT use the beta on this one.

In steam VR you set the resulotion to 100% ( and the redolution at the slider have to say 2207x2160 to begin with. This runs the reverbs at normal resulotion without supersampling).

In DCS settings in the VR-tab, pixeldensity/PD should be set to 1.0.

You should use supersampling in steamvr i stead of DCS because it draw less calculation power that way. And, do not start to increase supersampling initially, it will most probably stutter.

The specific dcs graphiccsettings I dont have them in the head so Ill get back later.

I also overclock my 2080ti with ( in my case) asus gpu overclocking tool. It may give you around 10% extra GPU power. I have read that msi afterburner is really good on this, so i think others would recommend that one.

There is a cheap software, fpsVR ( like 4usd) that run inside the vr headset and show you in real time the liad of gpu, cpu and the time it takes to produce the frames. Priceless for tuning. The best tuning tool I ever used. Incease ome setting and in game immediatly see that for example the GPU frame time rises too much, or that the GPU RAM fills to 100%. A must to save time in vr tuning.

The activation of reprojection by going and editinc a setting document:

All settings are in this link, actually a few mire that I forgot about also:

https://vr4dcs.com/2019/09/10/reverb-settings-for-dcs/

I did follow Thuds guide in that link, just as he said. All credit should go to Thud!

Hey mate, thanks for the suggestions for tweaks - Thud's guide was my first stop for tweaking the HMD, and then moved on to many, many more. I ended up spending over 100 hours trying to get VR working to my liking, all to no avail. No dramas, as I love 2D.

Windows mixed reality, I use highest setting of quality(I dont think it matters in gaming though, the steam vr and the game settings is setting the levels).

Just a quick note - the bit you said about WMR settings actually has a huge impact on performance. Unfortunately they didn't label it very well, so it's quite confusing, but if you opt into the Windows Insider program you can get it updated to better reflect what each setting does. Rather than try to explain it myself, you can read about it straight from one of the WMR developers at Microsoft.

I believe you when you say you don't see any stuttering, but I'm sure that if I was to use your PC and HMD I'd notice it. I think I'm just very susceptible to things like this. Out of interest, do you notice the rainbow effect on DLP projectors? I sure do, but many people don't. If you've never heard of the rainbow effect, you probably don't notice it - lucky bugger!

Thanks again for going to so much effort to try to help me get my Reverb stutter free.

It's particularly bad when flying over that town at about 700ft and looking left 90 degrees - every single structure is blurry and fuzzy, as they have a reprojection hovering behind them. My in-game counter showed that the minimum frame rate was 45fps, bouncing around up to 60ish, but only scraping below 45fps once or twice.

I can recreate what you describe by dropping below 45 with reprojection on, or dropping below 90 with reprojection turned off. 80 fps looks nasty moving around buildings with reprojection off whereas 45 looks super smooth with it on. The only time I see ghosting at 45 with reprojection on is when passing a fast moving aircraft coming the other way.

I never see values between 45 and 90 other then a rapid blip as it switches between. I wouldn't expect to see it bouncing around to 60ish.

Again I know you have spent ages on this and i perfectly accept the resolution argument but I still feel something isn't quite right with your headset be it settings or some physical issue with the device or PC.

I tried a few different visual quality settings but I dont think I saw any difference.

I left it at ”ultra(beta)” or something like that.

Ive read earlier that it only affects Windows Mixed Reality Home and nothing else, trying to find that reference I found this:

YangFromMSFT2y

Hi Gj80, My name is Yang and I'm an engineer working on the Windows Mixed Reality team at Microsoft.

 

The visual quality settings only affect the rendering of the WMR Home experience.

He basically says it doesnt affect games

Yes, I can see the rainbow effect if I make actions to see it. They doesnt disturb me and the two first projectors was DLP( but chosen with extra DLP-wheel speed to minimize these effects.)

I know you have spent a lot of time with this and tried loads of things but something simply doesn't sound right with this.

Seems that we might be thinking the same thought?

Reprojection maybe not working.

First thing is to make sure reprojection really is activated and works.

We need to know the windows version, I think

Btw.

this is nice guide also how to tweak DCS for VR

DCS: VR Optimization Guide

ywGhE1vnRVA

but with his shader mod I have distorted graphics on the edges of VR headset (HP Reverb v2), any idea about settings of MASKSIZE for HP Reverb?

I could never get a masksize that fit, so I turned off the mask...

I could never get a masksize that fit, so I turned off the mask...

This is ok for me:

#define MASKSIZE 0.690f

Here is JSGME version with all tweaks like no doble flare ect, ready for HP Reverb:

https://drive.google.com/file/d/1rza3qFjmAjx5gIjU2YAG15bDqaL-fAXp/view?usp=sharing

I find #define MASKSIZE 0.600f to be ok for me in the Reverb

thanks guys, i think i gave up at about 595 before ... i found a tiny border at 600 and 630 worked out fine for me, thanks

I never see values between 45 and 90 other then a rapid blip as it switches between. I wouldn't expect to see it bouncing around to 60ish.

This is a result of the motionvector=auto tweak that is mentioned in Thud's guide and which many of the DCS VR community swear by. Many VR users claim motion reprojection isn't enabled unless this tweak is made, which is utterly false. It seems that even among those who have a "perfectly smooth VRexperience", the settings used to achieve this are rather different. Which again suggests I have a very different idea if smooth to them.

BTW, can I ask which aircraft you fly most? Possibly ground pounders, WW2 or choppers?

This is a result of the motionvector=auto tweak that is mentioned in Thud's guide and which many of the DCS VR community swear by. Many VR users claim motion reprojection isn't enabled unless this tweak is made, which is utterly false.

No, actually its not false at all. You can't get it to work, so you insist that it doesn't do anything. :megalol:

Just a quick note - the bit you said about WMR settings actually has a huge impact on performance. Unfortunately they didn't label it very well, so it's quite confusing, but if you opt into the Windows Insider program you can get it updated to better reflect what each setting does. Rather than try to explain it myself, you can read about it straight from one of the WMR developers at Microsoft.

That link you refer to actually says the same as I say:

Visual quality of my home - changing these settings affects only the WMR home environment (aka "cliff house").

Adjust level of detail... - this changes some of the rendering affects we use in the home Environment. In particular, the visual quality of different materials (wood, concrete, etc.) will scale as you change the setting from low to high.

Change app window resolution - By default, most 2D windows launched in the home Environment are launched with a 720p resolution. You can of course manually resize them horizontally & vertically to change that. However, you can opt to have them all start at 1080p. Previously this option was available as the "Very high (beta)" option under Visual quality. We've appropriately split it out as a separate setting now.

So this is only in the "Home Environment room" where you start you WMR experience for WMR...something...only was there 1 minut during setup.

This is not important for me as such, but I Think its good for other reading this thread so that they dont use that much time believing that that setting is important.

This is a result of the motionvector=auto tweak that is mentioned in Thud's guide and which many of the DCS VR community swear by. Many VR users claim motion reprojection isn't enabled unless this tweak is made, which is utterly false. It seems that even among those who have a "perfectly smooth VRexperience", the settings used to achieve this are rather different. Which again suggests I have a very different idea if smooth to them.

If I understand you correctly you mean that the difference between this two lines doesnt make any difference?

//"motionReprojectionMode" : "auto",

and

"motionReprojectionMode" : "auto",

Not true. Not for me at least. When I first started DCS with the reverb, I got around 75 to mostly 90fps with the settings I had and in that/plane/Quick mission I run.

I didnt even really have a smooth gaming even though the DCS counter said 90fps. there was some small stuttering regardless of what at least seemed as 90fps on the counter.

Only change I did was remove the // from that line (according to the difference between the two quotes above), and I got very smooth gaming with either 90 or 45fps on the DCS fps counter.

I have resett the REM-marks( // ) a couple of times for some setting/tests, and the counter did start hunting up and down and I also got that stutter again, each time. After each test I removed the REMs ( // ) and ech time it Went back to 90 or 45 with smooth gaming.

If I understand your statement, I Think you should consider the possibility that thuds guide is correct and if it doesnt work the same for you you still have a problem with the setup. Maybe, after more than one install, you have two or more different

default.vrsettings files

I find #define MASKSIZE 0.600f to be ok for me in the Reverb

For me from 685 is ok, between 600 and 680 I see still issues on edge on FOV at some angles when you move your head.

If I understand you correctly you mean that the difference between this two lines doesnt make any difference?

 

and

 

 

Not true. Not for me at least. When I first started DCS with the reverb, I got around 75 to mostly 90fps with the settings I had and in that/plane/Quick mission I run.

I didnt even really have a smooth gaming even though the DCS counter said 90fps. there was some small stuttering regardless of what at least seemed as 90fps on the counter.

 

Only change I did was remove the // from that line (according to the difference between the two quotes above), and I got very smooth gaming with either 90 or 45fps on the DCS fps counter.

I have resett the REM-marks( // ) a couple of times for some setting/tests, and the counter did start hunting up and down and I also got that stutter again, each time. After each test I removed the REMs ( // ) and ech time it Went back to 90 or 45 with smooth gaming.

 

If I understand your statement, I Think you should consider the possibility that thuds guide is correct and if it doesnt work the same you can find a WMR for you you still have a problem with the setup. Maybe, after more than one install, you have two or more different and maybe the WMR/Reverb do not use the one you are changing…?

Yep, that's the line I was referring to. You can find a WMR paper by Microsoft explaining exactly what it does with each of three settings; my results are inline with that.

If I understand you correctly you mean that the difference between this two lines doesnt make any difference?

 

and

 

 

Not true. Not for me at least. When I first started DCS with the reverb, I got around 75 to mostly 90fps with the settings I had and in that/plane/Quick mission I run.

I didnt even really have a smooth gaming even though the DCS counter said 90fps. there was some small stuttering regardless of what at least seemed as 90fps on the counter.

 

Only change I did was remove the // from that line (according to the difference between the two quotes above), and I got very smooth gaming with either 90 or 45fps on the DCS fps counter.

I have resett the REM-marks( // ) a couple of times for some setting/tests, and the counter did start hunting up and down and I also got that stutter again, each time. After each test I removed the REMs ( // ) and ech time it Went back to 90 or 45 with smooth gaming.

 

If I understand your statement, I Think you should consider the possibility that thuds guide is correct and if it doesnt work the same you can find a WMR for you you still have a problem with the setup. Maybe, after more than one install, you have two or more different and maybe the WMR/Reverb do not use the one you are changing…?

My bad, your description of the line is what I meant (am currently writing from hospital on my phone!) and it can be enabled by commenting out the //, or the value changed to motionvector.

. You can find a WMR paper by MS or Steam (can't recall) explaining what it does; my results are inline with what Ms describes.

This is a result of the motionvector=auto tweak that is mentioned in Thud's guide and which many of the DCS VR community swear by. Many VR users claim motion reprojection isn't enabled unless this tweak is made, which is utterly false. It seems that even among those who have a "perfectly smooth VRexperience", the settings used to achieve this are rather different. Which again suggests I have a very different idea if smooth to them.

 

BTW, can I ask which aircraft you fly most? Possibly ground pounders, WW2 or choppers?

No it's not false. Have you checked every single user's VR performance? Er, no, you haven't so go find another hobby other than VR trolling. You aren't needed or wanted here by the majority.

No it's not false. Have you checked every single user's VR performance? Er, no, you haven't so go find another hobby other than VR trolling. You aren't needed or wanted here by the majority.

Not trolling, just correcting incorrect statements so others don't waste their time. Thanks for your constructive input.

Here is that guide:

https://steamcommunity.com/games/719950/announcements/detail/3229520292654229103

Note that commenting out the line doesn't disable reprojection though. There is currently no way to fully disable reprojection with WMR. Btw, if you're not adverse to tweaking this, you should try the latest Windows Insider build, which changes up the WMR quality options for potential performance gains. There's a guide in this forum.