Bigscreen Beyond PCVR Headset

[Ravenger]

Ironically, the one recent headset that could use eye tracking to solve PC performance issues, and which also offers one of the clearest and sharpest PC VR images, doesn't support eye tracking for PCVR... at least not yet.

 

[Toekiller]

Do you know roughly what percentage of the gpu is used up by streaming? I wonder if I'll be able to run 90hz in all conditions (sim racing) with my 4090. There are times, mainly night races in the rain where my gpu is hitting 95% and that's at 72hz. It's in extreme conditions, and very rare, but still no chance of 90hz.

I've seen 25% mentioned a couple of times. For example... https://forums.flightsimulator.com/t/pico-4-what-do-you-think/546978/42?page=3.

 

[glenimp617]

I've seen 25% mentioned a couple of times. For example... https://forums.flightsimulator.com/t/pico-4-what-do-you-think/546978/42?page=3.

Yeah, I can believe that. My Pico 3 was amazing to be honest, it was only the constant dead pixels which sealed its fate.

 

[Tykey]

Features/specs/capabilities > form factor/size/weight
Everytime for me.

 

[BigBANGtheory]

Features/specs/capabilities > form factor/size/weight
Everytime for me.

I tend to agree however social media is flooded with people that to them VR = VRChat so as long as it ticks that box its nextgen VR

 

[melmac]

I've seen 25% mentioned a couple of times. For example... https://forums.flightsimulator.com/t/pico-4-what-do-you-think/546978/42?page=3.

Wow, the Pico 4 has an awful lot of overhead. I hope updates since October have reduced that.

 

[glenimp617]

I have a wired PSVR headset with me at the minute so going to carry out a few tests this weekend.

1. Performance Overhead (Virtual Desktop vs Displayport)
2. GPU requirements for Bigscreen Beyond

Pico 4 90hz (2160x2160)
SteamVR (3120x3120) = 9.73m Pixels (Native)
SteamVR (2116x2116) = 4.47m Pixels (to match HTC Vive Pro)

Bigscreen Beyond 90hz (2560x2560)
SteamVR (3696x3696) = 13.66m Pixels (40% more pixels than Pico4)

HTC Vive Pro 90hz (1440x1600)
SteamVR (2016x2240) = 4.51m Pixels (Native)
SteamVR (2960x3292) = 9.74m Pixels (to match Pico)
SteamVR (3512x3904) = 13.71m Pixels (to match Bigscreen)

Automobilista settings during test
AF 16x
MSAA High (highest it will go)
Everything Ultra (or the highest it will go)
Shadows Medium
Reflections Medium
Enhanced Mirrors Off

Test 1 (Virtual Desktop vs Displayport / HTC Vive Pro Res)

VD Pico 4 90hz (2116x2116) = 4.47m Pixels
DP HTC Vive Pro 90hz (2016x2240) = 4.51m Pixels

Automobilista 2 - 5 Laps,25 Cars,Silverstone,Clear,Day
Pico4 GPU Load 41%-[46%-48%]-55% | Avg FPS 90
HTC Vive Pro GPU Load 38%-[40%-45%]-48% | Avg FPS 90

Automobilista 2 - 5 Laps,25 Cars,Silverstone,Wet,Night
Pico4 GPU Load | Avg FPS
HTC Vive Pro GPU Load | Avg FPS

Test 2 (Virtual Desktop vs Displayport / Pico 4 Res)

VD Pico 4 90hz (3120x3120) = 9.73m Pixels
DP HTC Vive Pro 90hz (2960x3292) = 9.74m Pixels

Automobilista 2 - 5 Laps,25 Cars,Silverstone,Clear,Day
Pico4 GPU Load 75%-85% | Avg FPS 90
HTC Vive Pro GPU Load | Avg FPS

Automobilista 2 - 5 Laps,25 Cars,Silverstone,Wet,Night
Pico4 GPU Load | Avg FPS
HTC Vive Pro GPU Load | Avg FPS

Test 3 (Bigscreen)

DP HTC Vive Pro (3512x3904) = 13.71m Pixels

Automobilista 2 - 5 Laps,25 Cars,Silverstone,Clear,Day
GPU Load | Avg FPS

Automobilista 2 - 5 Laps,25 Cars,Silverstone,Wet,Night
GPU Load | Avg FPS

 

[glenimp617]

Having to expand my test a bit as I wonder if the resolutions make a difference to VD overhead.

Currently looking like only 3-6% overhead at the minute, although displayport doesn't jump around as much, it's more constant load whereas VD keeps changing making it hard for me to get a feel for what's going on. If it scales in a linear fashion then at the Pico res were looking at a 6.5-13% overhead, which fits quite well with reverb g2 performance figures (pico 75% load / g2 65% load).

 

[glenimp617]

Well as tests go, that was awful

Trying to run the HTC at Pico res resulted in reprojection kicking in at 45fps and the brief moments it came out of it, GPU 97% at around 80fps

 

[melmac]

Well as tests go, that was awful

Trying to run the HTC at Pico res resulted in reprojection kicking in at 45fps and the brief moments it came out of it, GPU 97% at around 80fps

Keep going with your testing. It's been interesting reading so far.

I did something like this with my Quest 2 during lockdown. I got roughly the same overhead as you somewhere around 5-6% Which would make sense as they are both using the same Xr2 processor.

 

[glenimp617]

Keep going with your testing. It's been interesting reading so far.

I did something like this with my Quest 2 during lockdown. I got roughly the same overhead as you somewhere around 5-6% Which would make sense as they are both using the same Xr2 processor.

I'm struggling. I don't understand why the pico performs so much better than the HTC when running at the same resolution.

 

[melmac]

I'm struggling. I don't understand why the pico performs so much better than the HTC when running at the same resolution.

You got results from running the Pico 4 at the same res as the HTC Vive Pro.

But you are having trouble with the other way around? How are you changing the resolution of the Vive? Through Steam VR or the in game supersampling?

 

[glenimp617]

You got results from running the Pico 4 at the same res as the HTC Vive Pro.

But you are having trouble with the other way around? How are you changing the resolution of the Vive? Through Steam VR or the in game supersampling?

Yeah

I'm changing it in Steam as the in-game setting does bugger all

To be honest, I mainly just wanted to see if I could run bigscreen beyond at decent settings, and it's not looking good. Dry tracks are just about ok, 85% - 90% but forget about rain, or any night racing. That's at 72fps though, not 90. Hard to tell really as I still don't know the overhead of VD, or if SteamVR SS has any kind of overhead. I'm just going with about 10% for VD, so should just be about ok.

Biggest thing I've learnt is that the OLED screens on the Vive are sooooo much better.

 

[Mr_Sukebe]

Any see Sad Brads recent video about the Bigscreen headset?
Clearly he was very impressed with "some" aspects of it. I thought the final 2 mins were interesting, where he pretty much implied that he knew the "big boys" were bringing something similar out in the near future, i.e. with mini-OLED hires screens. Interesting.

 

[glenimp617]

Any see Sad Brads recent video about the Bigscreen headset?
Clearly he was very impressed with "some" aspects of it. I thought the final 2 mins were interesting, where he pretty much implied that he knew the "big boys" were bringing something similar out in the near future, i.e. with mini-OLED hires screens. Interesting.

Problem is, yet again these new headsets are a GPU generation too early.

 

[Unseul]

Problem is, yet again these new headsets are a GPU generation too early.

I don't think so, not if they have eye tracking. Ravenger has been saying that his quest pro is really clear, despite the fact it's not really much different to a q2 for pixels.

With good lenses and screens, with dlss/fsr and eye tracked foveated rendering, I reckon the requirements may not be as high as they were.

 

[Ravenger]

Yeah, pancake lenses are the gamechanger for sure. And there's something going on with the Quest Pro's complete optical stack that seems to make it more than the sum of its parts. Whenever I go back to Quest 2 it just seems like the entire image is washed out and smeared with vaseline by comparison, and cut off at the edges due to the single panel.

The resolution on the Bigscreen Beyond does trouble me, at least for gaming, as it's going to require a hefty GPU to drive. Of course for its use case of media consumption that may not be an issue, but it would be for things like PC sims.

Ironically the Beyond's biggest strength - custom fitted faceplates and IPD, are a big weakness for demoing it to other users. If I wanted to see how good it is I can't unless I buy my own faceplate and my IPD matches the person who bought it.

 

[Ravenger]

I think the Quest Pro shows that resolution isn't everything, and even OLED isn't everything. With local dimming the MicroLED displays can approach OLED colour reproduction. Even without local dimming the blacks on the Pro are closer to black than LCD headsets.

The issue with pancake lenses is they require really bright panels as the lenses block a lot of light. The Beyond is using really bright OLEDs to compensate for the pancake lenses but a lot of that brightness is being absorbed by the lenses so the display isn't that bright, according to reports.

 

[melmac]

Yeah

I'm changing it in Steam as the in-game setting does bugger all

I don't know much about the game, sorry. But, Changing the Resolution in Steam is super sampling. It adds a lot of overhead and if you have the setting in game changed as well, that's double super sampling.

I am guessing that's where your performance issues are.

 

[glenimp617]

I don't know much about the game, sorry. But, Changing the Resolution in Steam is super sampling. It adds a lot of overhead and if you have the setting in game changed as well, that's double super sampling.

I am guessing that's where your performance issues are.

No, not at the same time. I tried one, then the other just to see. Reading online I think the ingame one is broken as it does nothing at all.