HP's Reverb VR Pro Headset

[motjoo]

I think most people that plan on getting the Reverb have 2080Ti's. If the one person on Reddit said the 2080ti can barely hold 60 fps with all settings on the lowest in DCS, I don't have too much hope for the Reverb.

 

Which is why I'm holding off on waiting for people to get it and try it out. I'd like to upgrade from my Odyssey but I dunno if the Rift S is a big enough upgrade if the Reverb pans out to be a no go.

 

That person later corrected himself. The low fps was caused by

the selected shadows. With shadows off he could get “rather high” settings

Edited June 3, 2019 by motjoo

[sze5003]

That person later corrected himself. The low fps was caused by

the selected shadows. With shadows off he could get “rather high” settings

That's somewhat a relief. I think soon we will have some people reporting here who have obviously used dcs longer than that Reddit person. He admitted he wasn't familiar with the sim but at least those of us who have had it for years will have a better understanding of what settings to use.

[imacken]

Unless I am missing something here (entirely possible), what are you guys going to be powering the Reverb with? Even a 2080 Ti is going to groan under the load (even with the supposed 50% increase coming to DCS). You will have to turn all the settings way down to have decent framerates above 30ish.

 

Yes, this is what I’ve been saying for a while. Look at my spec. With PD 1.5 and medium/high settings in DCS I get a fairly constant 45fps with my Vive Pro. What hope for the Reverb?

Still, I’ll find out in a few days, I guess.

[9echo]

But is it such a massive difference in performance? I have not tried VR yet, but what sort of framerates are we talking about here for the DCS users on this forum in general? I mean, I can run DCS on my rig with 60fps pretty much at all times after having tweaked some graphics, enabling AA etc in Nvidia Profile Inspector and reducing equivalent settings in DCS itself, all of this on top of having increased loading radius etc beyond DCS standard extreme settings through modifying some .luas. Is this the basis for the performance discussions here where transitioning from 2D to VR literally cuts the performance by more than 50%, just because of the 2D-VR difference itself?

[sze5003]

But is it such a massive difference in performance? I have not tried VR yet, but what sort of framerates are we talking about here for the DCS users on this forum in general? I mean, I can run DCS on my rig with 60fps pretty much at all times after having tweaked some graphics, enabling AA etc in Nvidia Profile Inspector and reducing equivalent settings in DCS itself, all of this on top of having increased loading radius etc beyond DCS standard extreme settings through modifying some .luas. Is this the basis for the performance discussions here where transitioning from 2D to VR literally cuts the performance by more than 50%, just because of the 2D-VR difference itself?

Yes VR is very demanding as now you are drawing a 3D render on everything and on top of that the dcs engine is not the best when it comes to optimizations(minus the 50% vr improvement patch which hopefully will do us more justice).

 

It may not be so demanding on lesser resolution optics but with the reverb pushing the highest resolution, you need more power to drive it. As always shadows will be the first setting people will have to turn down.

[Nagilem]

Ok - I have a 2080ti card with an o+.

 

I have tested MSAA 2x + MFAA with a PD up to 1.5 @ 90hz (60 is too dim for me with LOTS o flicker) with Steam Motion Repro on and have reasonable performance in crowded areas (think Sochi in Caucasus @ a mostly consistent 45fps). My typical setting is same with PD at 1.3 and that gives pretty good results including heavy maps with lots of action.

 

At 1.3 PD setting level, my card must render 5.9M pixels. At 1.5 PD, 6.9M pixels. With the Reverb set to PD 1.0 everywhere, the card will need to render 9.3M pixels. That is over 1/3 more pixels per frame!

 

Not sure how the aliasing will be, but if I need some AA, I'm not sure I will be able to use MSAA and not have a slideshow on my hands. I may have to resort back FXAA with Anisotropic filtering and possibly SMAA to get reduced shimmer.

 

I'll probably have to subsample (pd to <1) to get the game playable UNLESS the holy grail of updates comes this week with the 2.5.5 map optimization in Caucasus and PG maps and we truly see a 50% performance gain.

 

At this point, its finger crossing time...

[imacken]

At 1.3 PD setting level, my card must render 5.9M pixels. At 1.5 PD, 6.9M pixels. With the Reverb set to PD 1.0 everywhere, the card will need to render 9.3M pixels. That is over 1/3 more pixels per frame!

 

Not sure your calculations are correct. The O+ and the Vive Pro have 2*1440*1600=4.6m pixels. At 1.5 PD in DCS, you are looking at 2*2160*2400= 10.3m pixels. More than the Reverb.

I think you have just taken the native res and multiplied by PD. That is not correct. The PD has to be applied to both axes.

[Panthera_Tigris]

Even imacken's number is a bit low. Please see my post : https://forums.eagle.ru/showpost.php?p=3934357&postcount=960

[Nagilem]

Sorry guys - I wasn't clear

 

O+

1440*1600*2 (for both eyes together) = ~4.6M pixels

1440*1600*2*1.3 (adding PD uptick) = ~5.9M pixels

1440*1600*2*1.5 (adding PD uptick) = ~6.9M pixels

 

Reverb

2160*2160*2*1=~9.3m pixels

 

Is the formula right? Its what I've seen so far so I may still have something off.

Edited June 3, 2019 by Nagilem

[Nagilem]

@PT - I think what you are saying with the post link is that additional SS is happening in SteamVR? My assumption is that I would set the SS in Steam to 100% and change PD in DCS.

 

 

 

Is one better than the other?

[imacken]

Sorry guys - I wasn't clear

 

O+

1440*1600*2 (for both eyes together) = ~4.6M pixels

1440*1600*2*1.3 (adding PD uptick) = ~5.9M pixels

1440*1600*2*1.5 (adding PD uptick) = ~6.9M pixels

 

Reverb

2160*2160*2*1=~9.3m pixels

 

Is the formula right? Its what I've seen so far so I may still have something off.

 

No, it’s not right. Look at my post above. You need to apply the PD to both axes.

[Nagilem]

Ok - I think I get it now... So it looks like this

 

(1440*PD)*(1600*PD)*2 for the o+

 

So with a PD of 1.5 - you end up with ~10.3m pixels for the o+

 

Reverb at 1.0 will be ~9.3m pixels which could perform better with better resolution.

 

 

Here's to hoping the reverb will look better at 1.0 PD :)

[imacken]

Ok - I think I get it now... So it looks like this

 

(1440*PD)*(1600*PD)*2 for the o+

 

So with a PD of 1.5 - you end up with ~10.3m pixels for the o+

 

Reverb at 1.0 will be ~9.3m pixels which could perform better with better resolution.

 

 

Here's to hoping the reverb will look better at 1.0 PD :)

 

Yes, exactly. Same for O+ and Vive Pro.

[Nagilem]

@imacken AND PT - thank you for the clarification. That is giving me great hope that the reverb will play much better than I originally thought.

[Panthera_Tigris]

@PT - I think what you are saying with the post link is that additional SS is happening in SteamVR?

 

 

Yes, there is some amount of super sampling that happens by default. When SteamVR is at 100%, it is actually NOT at the native res. For example, my Vive should have a res of 1080×1200 per eye but it is actually at 1512x1680 without any super sampling applied anywhere.

 

You can try it yourself. Reduce the PD in DCS to 1 (or test with any other game) and reduce SS in SteamVR to 100%. Your resolution will be different than the native res. The exact amount of this default SS differs form headset to headset. This is done to counter distortion.

Edited June 3, 2019 by Panthera_Tigris

[imacken]

Yes, there is some amount of super sampling that happens by default. When SteamVR is at 100%, it is actually NOT at the native res. For example, my Vive should have a res of 1080×1200 per eye but it is actually at 1512x1680 without any super sampling applied anywhere.

 

You can try it yourself. Reduce the PD in DCS to 1 (or test with any other game) and reduce SS in SteamVR to 100%. Your resolution will be different than the native res. The exact amount of this default SS differs form headset to headset. This is done to counter distortion.

You say a certain amount of SS goes on by default, but this discussion has gone on many times before re Steam settings. Steam has changed that pixel value more than once. Still, the general consensus afaics is that 100% is basically still no SS on the part of Steam.

[boedha68]

:D all theory. When the mailman drops the boxes, we are all in another world! :D

[Harlikwin]

Ok - I think I get it now... So it looks like this

 

(1440*PD)*(1600*PD)*2 for the o+

 

So with a PD of 1.5 - you end up with ~10.3m pixels for the o+

 

Reverb at 1.0 will be ~9.3m pixels which could perform better with better resolution.

 

 

Here's to hoping the reverb will look better at 1.0 PD :)

 

Yup, thats the right math. Your earlier math might be how steam SS numbers work though.

 

And of course it will. Real pixels are way better than super sampled scenes that are then down sampled.

[Harlikwin]

But is it such a massive difference in performance? I have not tried VR yet, but what sort of framerates are we talking about here for the DCS users on this forum in general? I mean, I can run DCS on my rig with 60fps pretty much at all times after having tweaked some graphics, enabling AA etc in Nvidia Profile Inspector and reducing equivalent settings in DCS itself, all of this on top of having increased loading radius etc beyond DCS standard extreme settings through modifying some .luas. Is this the basis for the performance discussions here where transitioning from 2D to VR literally cuts the performance by more than 50%, just because of the 2D-VR difference itself?

 

Yeah, because it has 2x more draw calls, which is why CPU is as or more important than GPU.

[Harlikwin]

What is side effect or symptoms of getting 30 and below fps. My system only get around 17-30 fps depending on the situation.

I still am able to enjoy the VR on a medium to low settings. No low latency or stuttering here.

Using odyssey +.

 

 

When I was running my low end rig with a rift I could reduce settings to get a stable 45 most of the time (no shadows minmal eye candy) and for the most part it didn't get bad unless things dropped into the low 20's or teens. Like I couldn't tell a difference between 30 and 45 most of the time. But it would certainly stutter/slide show from time to time. And I think thats way more CPU limiting than GPU.

[Harlikwin]

That's somewhat a relief. I think soon we will have some people reporting here who have obviously used dcs longer than that Reddit person. He admitted he wasn't familiar with the sim but at least those of us who have had it for years will have a better understanding of what settings to use.

 

I think with the VR boost it will be fine for high end guys (i.e. I9x00k and 2080ti). No promises on anything less.

[Nagilem]

@Harlikwin AND 9Echo - Actually nevermind. I'll do more searching and see what I find.

Edited June 3, 2019 by Nagilem

[John Hargreaves]

:D all theory. When the mailman drops the boxes, we are all in another world! :D

 

Hehe I hope he's a bit more gentle than that with it :music_whistling:

 

Just got my confirmation email from HP, Reverb expected next Tuesday :pilotfly: excited.

[dburne]

Hehe I hope he's a bit more gentle than that with it :music_whistling:

 

Just got my confirmation email from HP, Reverb expected next Tuesday :pilotfly: excited.

 

:thumbup:

 

I look forward to hearing your thoughts on it.

[Panthera_Tigris]

You say a certain amount of SS goes on by default, but this discussion has gone on many times before re Steam settings. Steam has changed that pixel value more than once. Still, the general consensus afaics is that 100% is basically still no SS on the part of Steam.

 

Like I said, you can verify it for yourself in SteamVR. I just posted a screenshot with my values.

 

Also, I am not sure if you are talking about the interal SS OR the recommended SS for your GPU. They are not the same things and its the latter which was changed by Steam. If the former was also changed, I am not aware of that.

Edited June 3, 2019 by Panthera_Tigris