Jump to content
Sign in to follow this  
chiliwili69

Supersampling ratios Rift vs VivePro (vs Odyssey & WMR)

Recommended Posts

Just to compare the net effort demanded to the GPU for each VR device in terms of pixel to be rendered I created a spreadsheet were those are compared:

 

In the new SteamVR terminology SS=1.0 is 100%, SS=1.3 is 130%.

I added a column called Absolute ratio which compare the absolute SuperSampling with respect to the physical display resolution. As shown in another post with samples pictures, going in the Rift to absolute SuperSampling ratios 3.3 (SS_SteamVR=2.0) doesn´t bring a significant added details. With the VivePro I will do some image comparisons with different SS ratios.

 

It is important to remark that:

150% in Rift = 128% in Vive = 128% in VMR = 128% in Odyssey = 72% in VivePro = 40% in Pimax 5K+ = 44% in Pimax 8K

 

This spreadsheet is accesible in the tab "SS per device": https://docs.google.com/spreadsheets/d/1gJmnz_nVxI6_dG_UYNCCpZVK2-f8NBy-y1gia77Hu_k

 

188407355_SSratiosperVRdevice.thumb.png.9ff08c05a33f608847ee342bfb7959fd.png

 

 

 

I included the Odyssey in the comparison but in the present day it still have the scale problems with IL-2.

Edit 1st-June-2018: Odyssey, like all WMR, are working well now with IL-2 VR.

 

Edited by chiliwili69
  • Thanks 3
  • Upvote 3

Share this post


Link to post
Share on other sites

Still don't know how Chilli is getting that 1.7 SS result for the Rift. My 1080 TI barely manages 1.3 with the Rift in BoX. This is for 90 fps though, maybe he's going for 45 fps.

Share this post


Link to post
Share on other sites
On 27/4/2018 at 2:17 PM, TimJ147 said:

Out of interest what steam ss are you running. Are you using a 1080 card?

 

I use a 1080Ti card.

With the Rift I used 1.7 SteamVR ratio.

Now with VivePro I use 130% SteamnVR which is 1.3

My settings are HIGH, although for Kuban or complex scenarios I use BALANCED.

Still need to play around for the best settings for my CPU&GPU&Vivepro

19 hours ago, BeastyBaiter said:

Still don't know how Chilli is getting that 1.7 SS result for the Rift. My 1080 TI barely manages 1.3 with the Rift in BoX. This is for 90 fps though, maybe he's going for 45 fps.

 

I don't know what you mean by "is getting that 1.7 SS results for the Rift"

 

It depends heavily on the ULTRA/HIGH/BALANCED/LOW presets. 

I don't know why but Kuban is in general more demanding.

In the past I made some experiments to determine when the GPU is the bottleneck for different SS ratios and presets>

With the old Balapan track, suing the VR test settings and with a 1070 card I was close to 90 most of the time.

 

 

Share this post


Link to post
Share on other sites

I have updated the table in the first post to include the Vive as well. I also corrected some cells mistakes of previous table.

 

Now we can see that Vive and Rift uses different internal SS ratios for SS=100% (Absolute 1.66 for Rift versus 1.96 for Vive).

So, SS numbers apply differently to Rift and Vive. So when making performance comparisons of Rift vs Vive this should be taken into account.

We can also check that Vive and VivePro use same internal SS ratio for SS=100%.

Share this post


Link to post
Share on other sites
2 hours ago, chiliwili69 said:

I have updated the table in the first post to include the Vive as well. I also corrected some cells mistakes of previous table.

 

Now we can see that Vive and Rift uses different internal SS ratios for SS=100% (Absolute 1.66 for Rift versus 1.96 for Vive).

So, SS numbers apply differently to Rift and Vive. So when making performance comparisons of Rift vs Vive this should be taken into account.

We can also check that Vive and VivePro use same internal SS ratio for SS=100%.

 

You may want to clarify, I assume you are talking about SS settings in Steam VR.

I do not use the settings in Steam VR, I only set SS in the Oculus software.

Course I may be the only one doing so.

;)

Edited by dburne

Share this post


Link to post
Share on other sites

No, you are not the only one who sets SS in the third party oculus software (Oculus Tray Tool) or oculus software (Oculus Debug Tool), I believe quite a few use that.

 

The column at the right shows the SS values in terms of SS used in OTT/ODT, even if they are not used for the rows of Vive, VivePro or Odyssey.

The column of the absolute SS ratio uses the SteamVR convention which is based in the ratio pixels_rendered/pixels_display.

 

I find the SteamVR convention more useful in order to have a direct idea about the extra effort in terms of pixels required to the GPU (ie 1.3 means 30% more)

Edited by chiliwili69

Share this post


Link to post
Share on other sites

I think the more I read the more confused I get.

 

Ok so I have been running Steam VR, with my SS set in Oculus Tray Tool for my Rift.

Having a look in Steam VR Video Settings, the box for Manual Override of the Application Resolution is not checked.

But then it also states "Steam VR Automatically sets your application resolution based on the performance of your GPU".

So is SS truly not being used in Steam VR? Is it really adjusting resolution based on performance? I want to make sure the only SS used in BoX for my Rift is what I have set in the Oculus software via the Tray Tool.

 

Alternatively, if I click on Manual Override in Video Settings, it sets to a recommended level of 200%. I do not use Manual Override, just took a peek in there.

 

Man I so wish 1CGS would offer native support for Oculus and I could get rid of all this Steam crap.

 

Edited by dburne
  • Like 1

Share this post


Link to post
Share on other sites

I beleive automatical detection simply means steam vr adds an ss value to the exe file after you have launched it once. If you then reset the value to 100 % it is not going to change it again.

Share this post


Link to post
Share on other sites
25 minutes ago, -[HRAF]Black_Sab said:

I beleive automatical detection simply means steam vr adds an ss value to the exe file after you have launched it once. If you then reset the value to 100 % it is not going to change it again.

 

Yeah that is kinda what I was thinking, when I was looking around in there yesterday I played  with moving that SS slider around.

I moved it down to 100%, but still the resolution it listed per eye was higher than the native resolution of the Rift, which is what I find confusing.

 

I just want to make sure Steam VR is not doing any SS change to my Rift display, and at this time after looking around in those settings I am not sure it is not.

Share this post


Link to post
Share on other sites

From the little I understand I believe chiliwili found that both oculus and vive are using hard coded super sampling. Probably that is the reason steam vr is showing higher values than expected on 100 %.

Share this post


Link to post
Share on other sites

Thanks - I guess for now I will just continue to leave that option unchecked...

Share this post


Link to post
Share on other sites
On 30/4/2018 at 1:14 PM, dburne said:

I moved it down to 100%, but still the resolution it listed per eye was higher than the native resolution of the Rift, which is what I find confusing.

 

In other to be sure SteamVR DOES NOT introduce any undesired added SS you must set the checkbox to manual override to 100%. (this is the same than the previous steamVR was doing).

Then you can control the SS from OTT.

 

If you set SS=1.0 in OTT and Manual Override with SS=100% in SteamVR you will see that your resolution is 1344x1600 per eye which is exactly what the table shows above.

 

If you set SS=1.3 in OTT and Manual Override with SS=100% in SteamVR you will see that your resolution is  around 1752x2086 per eye which is exactly what the table shows above.

 

You can always active the HUD of Oculus Debug Tool to effectively know the final resolution used by the Rift. (it should match the one reported by SteamVR)

On 30/4/2018 at 3:00 PM, -[HRAF]Black_Sab said:

Probably that is the reason steam vr is showing higher values than expected on 100 %.

 

Yes, all VR devices use an internal SS ratio. That´s why using 100% does not gives you the physical display resolution.

 

The Rift uses a 1.66 internal ratio. it is shown in the table above in the right column in green (absolute ratio)

 

The Vive and VivePro uses the same 1.96 internal ratio. it is shown in the table above in the right column in green (absolute ratio)

 

Share this post


Link to post
Share on other sites
5 hours ago, chiliwili69 said:

 

In other to be sure SteamVR DOES NOT introduce any undesired added SS you must set the checkbox to manual override to 100%. (this is the same than the previous steamVR was doing).

Then you can control the SS from OTT.

 

 

You are right on that, after playing around with many settings today because of the shimmering and roughness I was getting on the landscape now, I went in and put a check in manual override and set to 100%. 
Now the SS I set in Oculus Tray Tool seems to be working properly again, and the terrain is gorgeous again.

 

Thanks!

 

Man I sure wish 1CGS could add native Rift support...

Share this post


Link to post
Share on other sites
On 5/2/2018 at 6:53 PM, dburne said:

 

You are right on that, after playing around with many settings today because of the shimmering and roughness I was getting on the landscape now, I went in and put a check in manual override and set to 100%. 
Now the SS I set in Oculus Tray Tool seems to be working properly again, and the terrain is gorgeous again.

 

Thanks!

 

Man I sure wish 1CGS could add native Rift support...

+1. I made these changes in Steam VR last night to allow OTT exclusive control over SS in game and it made all the difference!  For me, the game is back to being one of the best VR flight experiences out there.

  • Like 1

Share this post


Link to post
Share on other sites
On 5/8/2018 at 7:53 AM, GIJoe said:

+1. I made these changes in Steam VR last night to allow OTT exclusive control over SS in game and it made all the difference!  For me, the game is back to being one of the best VR flight experiences out there.

I can not find the manual override option in Steam Settings.

I have a total of 25Gb VRam available and should have much better VR experience than present.

Share this post


Link to post
Share on other sites
On 4/23/2018 at 5:58 PM, chiliwili69 said:

I included the Odyssey in the comparison but in the present day it still have the scale problems with IL-2.

 

 
You may be able to fix this by editing the startup.cfg file ([game]/data/startup.cfg) and manually setting the IPD there.

Share this post


Link to post
Share on other sites
20 hours ago, Gordon200 said:

I can not find the manual override option in Steam Settings.

You have to look in Steam VR, not in Steam.

In Steam VR you have to go to Video.

14 hours ago, SCG_Fenris_Wolf said:

Odyssey and VivePro, which have the same default res at 1.0

 Odyssey and VivePro have same physical display resolution.

The res at 1.0 is different depending on the internal supersampling every device introduces.

For the VivePro I have check that at SS=1.0 the values are the ones displayed in the table above.

For the Odyssey, another user of this forum who owns Odyssey was reporting the values of the table above.

Share this post


Link to post
Share on other sites

Oh, all right, that is indeed very interesting. Thanks chili! 

Edited by SCG_Fenris_Wolf

Share this post


Link to post
Share on other sites

By the way, in your first post you still say Odyssey had a scaling problem, that's fixed. 

 

 

Thank you, the numbers are correct and I confirm them.

 

That means different values for SS must get put into SteamVR to achieve the same result - just as you said. The low value at 100% thus was the reason for the amazing performance/fps when I first connected the Odyssey 😂

Edited by SCG_Fenris_Wolf

Share this post


Link to post
Share on other sites

I have made the spreadsheet available online in the link of first post. Just for future reference and future devices.

 

I have realized now! that 170% SteamVR_SS is equivalent to 144% StreamVR_SS in the standard Vive. So when we use the IL-2 benchmark we should take that into account. I will modify the instructions to run the benchmark according to the device.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
Sign in to follow this  

×
×
  • Create New...