Varjo Aero

So, is this mysterious extra performance despite higher resolution *not* because of the Aero's dynamic foveated rendering tech? Is there even more performance to be gained from that?
 
So, is this mysterious extra performance despite higher resolution *not* because of the Aero's dynamic foveated rendering tech? Is there even more performance to be gained from that?

The Aero dynamic foveated rendering is a hook that a software developer needs to take advantage of for it to have any effect. Below states that the software needs to have a low and high resolution frame provided. So yes, there could be more performance gained from that.

"Applications should be prepared to render the frame into high resolution and lower resolution textures, so that the application can fall back to fixed rendering if necessary (for example, if the gaze is not available). There is no need to create a separate swapchain, however, since the same texture can be used for both fixed and foveated rendering with the help of viewports. A separate swapchain is only necessary if the engine does not allow you to set a custom viewport and always renders into the full texture."


There is overhead involved with foveated rendering, so there will be a break even point and then a point where it helps, but it won't be the cure all some people think it will be. There are estimates that it could help by 50-70%, but HTC's first attempt was closer to 30%. I would think the best case scenario would be for a wide FOV headset.
 
Last edited:
mysterious extra performance
WMR was originally intended to compete with SteamVR;
that adding software "glue" to use WMR headsets is less efficient than
Varjo's native SteamVR headset implementation is no real mystery;
Microsoft has a history of collaborating with Intel
for inefficient feature implementations
which want increasingly powerful CPUs for acceptable performance.

Good performance for Sony's new VR headset
using considerably a less powerful CPU and GPU than comparable PC VR
should not be considered mysterious, simply less inefficient.
 
WMR was originally intended to compete with SteamVR;
that adding software "glue" to use WMR headsets is less efficient than
Varjo's native SteamVR headset implementation is no real mystery;
Microsoft has a history of collaborating with Intel
for inefficient feature implementations
which want increasingly powerful CPUs for acceptable performance.

Good performance for Sony's new VR headset
using considerably a less powerful CPU and GPU than comparable PC VR
should not be considered mysterious, simply less inefficient.
A while back when StarVR headsets seemed more important they too were surprisingly efficient.

Of course they are still showing $3200 for a developer version.

Here’s the full list of specifications for the StarVR One as listed on the product page for the headset:
  • Panel: 2 x 4.77” AMOLED
  • Display resolution: 16 million sub-pixels
  • Refresh rate: 90Hz low persistence
  • Lens type: Custom Fresnel lenses
  • Field of view: 210-degree horizontal FOV, 130-degree vertical FOV
  • Eye Tracking: Fully integrated Tobii eye tracking
  • IPD measurement: Yes with automatic SW adjustment
  • Dynamic Foveated Rendering: Yes
  • Tracking; SteamVR Tracking 2.0
  • In/Out: 2 x 0.9m Type-C cables, 2 x 5m Type-C extension cables, 1 x 3.5mm stereo headphone jack with microphone, Optional cable adapter box
  • Weight: 450g without head strap/headband and cables
 
WMR was originally intended to compete with SteamVR;
that adding software "glue" to use WMR headsets is less efficient than
Varjo's native SteamVR headset implementation is no real mystery;
Microsoft has a history of collaborating with Intel
for inefficient feature implementations
which want increasingly powerful CPUs for acceptable performance.

Good performance for Sony's new VR headset
using considerably a less powerful CPU and GPU than comparable PC VR
should not be considered mysterious, simply less inefficient.
OK, i want one now :D

There is only the 3000EUR price tag including lighthouses and Index controllers plus tax, that is holding me back...a lot...
 
Last edited:
OK, i want one now :D

There is only the 3000EUR price tag including lighthouses and Index controllers plus tax, that is holding me back...a lot...

I've got base stations and controllers and I still haven't pulled the trigger. I'm still hoping Valve hits another home run. I'm not expecting it to be cheap, but I do expect it will be something I'll be happy with for a few years.

I'm still hoping the rumors of having on board processing to take the load of the GPU are true. That would be a huge deal and would allow them to have foveated rendering in the headset which would be critical if they increased the FOV.

Unlike HoiHman, the Index fits my face extremely comfortably and I have the lenses set all the way back to maximize FOV, so I expect the next Index model to fit be just as well.
 
Are you using anything besides sims?
I initially set my Index up over 2.5 years ago and haven't had to touch it since then. Perfect tracking all the time. It initially calibrates with the headset on the floor and after that you never have to adjust anything. It just works perfectly.

There is something to be said for that. But I understand your comment.

The problem is that other than Valve's co-venture with HP, the entire high end enthusiast headset market belongs to SteamVR tracking with base stations.

I seem to remember making a comment about how this would be how things would go 3 years ago.
 
  • Deleted member 197115

High end enthusiast headsets?
 
Last edited by a moderator:
Does that not kind of contradict your dissapproval of inside-out tracking?

I have had zero issues with the tracking of my G2 with race and flight sims.

Inside-out tracking is good enough for most people as evidenced by the completely stand alone Oculus system that completely outsells all the PC VR headsets combined. So whether it is viable or not is not in question.

He has used a variety of headsets both with inside-out tracking and base stations and prefers base stations. It would seem that he has more comparison data to draw a preference from than you or I have. There is good enough and the current best tracking on the market.

In this discussion there have already been performance issues mentioned using WMV vs. a more direct SteamVR implementation. How can the Aero perform well with more resolution than the G2?

All that said, there is no point getting into a heated tracking method discussion(not saying you've gotten heated). We've already seen that some people have decided base stations are evil or too expensive and want nothing to do with them and that's fine. For me it was a one time cost and I fully expected that my next purchase will likely be just a headset and I will keep the controllers and base stations I already have.
 
Last edited:
All that said, there is no point getting into a heated tracking method discussion(not saying you've gotten heated). We've already seen that some people have decided base stations are evil or too expensive and want nothing to do with them and that's fine. For me it was a one time cost and I fully expected that my next purchase will likely be just a headset and I will keep the controllers and base stations I already have.
I am definitively not heated, as i don't have a heated opinion on that topic :)

That aside, while not having had the pleasure to try out lighthouses yet, i actually have the experience, that i find the inside-out tracking of the G2 a bit lacking and unprecise with HL: Alyx as well as the controllers not exactly oozing quality.

Apart from that in my main addiction at the moment, race and flight sims, the inside-out tracking is perfectly fine and i am enjoying the less complex setup and lower cost of not having to buy additional HW.

I think for a headset that is supposed to be premium and has a premium pricetag, the additional components you have to buy are actually very worthy of a discussion. The Index stuff e.g. would set me back another 500EUR plus tax on top of the headset, which is almost as much as the complete G2 set.
 
Does that not kind of contradict your dissapproval of inside-out tracking?

I have had zero issues with the tracking of my G2 with race and flight sims.

Not really no. The picture in the G2 is not as stable as the picture in the Index when you combine VR with motion.

A lot of SFX users complain about microstutters in the Reverb G2 and and an "unstable image"

In common words a lighthouse tracked VR HMD handles heave much better than an inside out tracked VR HMD. It's something i noticed when i tried and compared the G2 to my Index

I already have the lighthouses and they are known for the best tracking available.

Yes of course the lighthouse tracking is the more expensive option, but when you already have them it's a different story. Also let's be honest, when you are considering buying a Varjo Aero, we can agree that money is not the issue.

Any way...... I just ordered one ( locally ) with an expected delivery date of 4 weeks and a 14 days returning periode.

Already excited to get my hands on it :D
 
I'd be excited too!
I look forward to hearing your impressions as always!
But even if you love it, I'm still waiting for the Index 2 to drop and I suspect my 2080Ti would be lacking with the Aero and I'm holding on to my 2080Ti until the 40 series arrive.
 
I'd be excited too!
I look forward to hearing your impressions as always!
But even if you love it, I'm still waiting for the Index 2 to drop and I suspect my 2080Ti would be lacking with the Aero and I'm holding on to my 2080Ti until the 40 series arrive.
Probably the wiser decision

But... .i have been waiting for Index 2 for too long now.

Knowing Valve and with all the chip shortages, it could easily take another 2-3 years.

Right now all Valve's focus is on the stream deck and i wouldn't be surprised if the Index 2 will be add-on for that and not really high-end PCVR focused.

Time will tell.

One thing is for sure, for that kind of cash, the Aero better blows me away or it's going back:)
( Like the G1, the G2 and the VivePro 2 )

There's absolutely no way i can justify this amount a cash for a VR Headset, but all those concerns might fade away when i put it on for the first time.

(PS: I got confirmation for a delivery in week of April 18th )
 
Last edited:
I was having big tracking issues with my first CV1 and my motion system . I had to use my “ recenter” button 2-3 times every lap and sincerely it was impossible to enjoy my racing and I almost was thinking to go back to monitors but then I tried the Oculus S of my friend( 2 years ago) and with my surprise tye tracking was much much better without almost any view drifting . With my G2 it’s almost the same thing. Maybe I have to use the recenter button once every 7-8 laps.
 
With my G2 it’s almost the same thing. Maybe I have to use the recenter button once every 7-8 laps.

Just by way of comparison. I can reboot my computer and my Valve Index is still centered.

In fact I've rolled my rig to the side of my room and then rolled it back to the center of the room, reconnected everything and started it back up and realized after a race was over that I had never re-centered my HMD because I put my rig back within an inch or less of where it was last.

That's for iRacing but it shows just how solid the tracking is. If my rig wasn't mobile, I would never have to recenter my HMD.

The same goes for every single room scale game I have. Recentering the HMD is never required.

For Dirt Rally 2.0 there may be a setting I need to change because it will "auto-center" based on where the headset is facing when Dirt Rally first comes up. So that requires a single Recenter click when the game first starts up.

Probably the wiser decision

It depends on whether the Index 2 is actually announced this spring and delivered by Fall. If it ends up being many years than you may be the wiser man. It's the end of March and all we've got is crickets.

Steam Next Fest is the next scheduled set of announcements and that is June 22nd. "Assuming" they announce something then it is still likely 6 months out.
 
Last edited:
Back
Top