Pimax VR HMD review NDA embargo lifts: not great news for sim racers

[Mascot]

Testers are now allowed to release reviews of the final hardware, and while the Pimax units are undoubtedly a step forward for VR (higher resolution, wider FoV), the bad news seems to be that they're likely to struggle to provide adequate framerates in sim racing (even with powerful 8700k/1080ti combos). Doubts also remain about the effectiveness of Pimax's reprojection/ASW solution.

Assetto Corsa benchmarks at about 1hr 15 mins timestamp:

More reviews:

Foveated rendering can't come quickly enough, it seems.
The wider FoV for better peripheral vision in sim racing is was what got me excited about the Pimax in the first place, but it seems like PC hardware might need a couple more generations to adequately power the displays (or HMD tech needs to find efficiency solutions) before sim racing can truly benefit from this. Maybe one card per eye is required?

Anyone here backing the Pimax Kickstarter or getting ready to preorder?

 

[Michal Burisin]

It seems I'll be preordering once available, however it looks like 2080ti is a must if one wants to use large fov mode, so I'm saving up for new PC.

However I think it will be great also in normal fov mode or even small fov mode with lower performance graphic cards, as these FOV modes are still much better than what old gen HMDs offer. It has also many other benefits over the old gen HMDs - big sweet spot, less god rays etc.

It's really amazing that small unknown company could beat Oculus and Vive to market with second gen hardware. Especially Oculus with their unlimited funding.

I'm wondering when HTC will go bankrupt, because their Vive Pro seems like a joke hardware wise and especially price wise.

Oculus seems to go for standalone, let's see if Pimax make them release something in near future. But I guess it would take a long time anyway, so I'll probably preorder Pimax.

I'm still not sure whether to get 8k or 5k+. 8k seems better for sims but is more performance hungry. I lean to 5k+ now with future upgrade with 7nm graphic cards to 8kx or StarVR if they make something in consumer price levels.

 

[Mascot]

I'm still not sure whether to get 8k or 5k+. 8k seems better for sims but is more performance hungry. I lean to 5k+ now with future upgrade with 7nm graphic cards to 8kx or StarVR if they make something in consumer price levels.

Popular opinion seems to favour the 5k+ over the 8k. Sharper image, deeper blacks, less of a performance hog. The extra resolution of the 8k might be better for sims though.

 

[Mascot]

edit: DP

 

[Michal Burisin]

I'll wait for reviews as well, the numbers from nVidia shows 30 - 50 % performance boost which I think is very good, let's see if they are correct or not.

With DLSS there might be additional boost, especially for supersampling so it might be good.

I don't have VR right now, but I tried Rift and Vive and SDE was really terrible. It seems 8k is better in this regard than 5k+, so I'll wait for SweViver update after he gets the 2080ti what he has to say about performance.

Anyway it's still lot of time - Pimax said they will do update on October and possibly open pre-orders and then couple of months before they manufacture all Kickstarter backer units + possibly their commercial deals headsets, so even with early preorder, I don't expect it to arrive till the end of the year... Hopefully the prices of hardware will go down a bit but it may go the other way round due to the USA customs increase which could make hardware more expensive world wide...

Let's see, but I think this is really breakthrough in VR, at least for me

 

[TroyC2]

I´m an 8K backer. Thinking about "downgrading" to the 5K to be honest to maybe get better fps. Either way I´ll probably end up getting a RTX2080Ti to be on the safe side. I´m quite a late backer so there is still plenty of time to decide. Who knows, maybe Pimax will be able to tweak the software a bit to improve performance. That´s what I´m hoping for at least.

 

[RobertR1]

I got a 2080ti coming in so I'm very much interested in this. Sounds like 8k makes sense for sim racers as viewing distance and lack of SDE is more important than clearer text.

I'm also hopeful that since Kunos is working closely with nvidia for Ray Tracing that we get some special treat on the VR side with foveated rendering, DLSS specific to VR or other such techniques that are part of the RTX series.

As much as I love the rift, I'd be happy to replace it with something next gen and since Oculus is hush hush on any true successor, we'll have to rely elsewhere. This has the makings of a good self Christmas present!

 

[Dan Costa]

Go for the 8K to have better screen quality and just play at lower resolutions. You can always decrease resolutions, but you can never increase screen quality.

 

[RobertR1]

Go for the 8K to have better screen quality and just play at lower resolutions. You can always decrease resolutions, but you can never increase screen quality.

It’s upscaled. Both are still taking in 1440 but yeah for sim racing it seems like 8k is the easy option with lower sde and better distance viewing.

 

[Michal Burisin]

Go for the 8K to have better screen quality and just play at lower resolutions. You can always decrease resolutions, but you can never increase screen quality.

Well not that easy I would say, if I remember right, in SweViver review, PC2 had around 40 FPS on 5k+ on 50% in Steam and 1.00 quality in PiTool and that was with medium details in dry conditions during day.

So I wonder what FPS would you get with higher quality in PiTool as it's more needed on 8k to get the same look as 5k+. Also in evening / night and rain the FPS would drop further so even with 5k+ the 2080ti might struggle, I'll wait for reviews to make final decision.

 

[Dan Costa]

I dont understand that point of view.

You can always downgrade your in-game resolution to match the performance levels you need, regardless of the headset physical screen resolution. The screens may be 2x4K but you can render the game at the same 2x2K as the oculus. You still win massive improvements on SDE and have a future proof headset knowing the resolution could go higher on a 2080ti or 3080ti.

So If you have to buy one, i would choose the best screen available today knowing that next year I would only need to upgrade my graphics card instead of the graphics card + headset.

 

[Michal Burisin]

Well if 2080ti struggles with 8k, the 5k+ on native resolution might actually look better than heavily downsampled 8k.

I suppose when 3080ti will be available, I would want to upgrade to another headseat anyway (8kx etc.) so it might be better to run smooth with clear unblurred visuals for year or two than heavily downsampled 8k, which I could take advantage in couple of years, but by then, something better will be surely available...

I'm still not decided, maybe I'm overthinking it, but it's not as easy choice for me.

 

[Dan Costa]

you mean "upscaled to 8k" in the first sentence.

I dont think it could look worse than a native image on the 5K. The screens have the same physical dimensions. And the SDE improvement alone is a big quality jump regardless of aliasing.

Its a good point. We have to wait for more impressions.

 

[anton_Chez]

I would NOT bank on anything coming from game developers or Nvidia regarding features that improve performance in head units. I am still BURNING at those scumbags that advertised the 10 series cards as having dedicated VR rendering techniques that allowed the image of one eye to be used for both eyes, effectively halving the cost of performance. These are the same people that also advertised SMP, the technique that renders 1 screen across 3 in a triple screen resolution allowing for a much less cost to perfomance. I have had BOTH of these setups and have never seen it implemented in anything other than iRacing. It's developer dependant and absolutely should not be counted on when buying new cards.

I am not buying a 2080Ti. I am done with these ridiculous price hikes and empty promises from what I consider one of my most hated companies. I despise Nvidia's business practice, from the lies of the 10 series 'features' (oh sick, I can take screenshots in game, now where's my improved VR performance?) to the disgusting VRAM debable that the tech journo's said was 'no big deal'. I wish VR would support SLI so I can go and get another used 1080Ti and call it a day. 2 grande for 50% performance increase over last gen? That's what we got for the same price generations past. How else to they hope to sell new cards if the new generation isn't faster than the old one? Now we are expected to pay double? Over Titan prices now for a Ti variant? At launch now? Wow.

Anyway, back on topic. If my 1080Ti cannot at least power the 5K+ at native refresh with details and cars on screen similar to the Oculus, I'll sell my pledge and stick with Oculus. We were told hardware requirements at the kickstarter stage (which I FAR exceeded at the time) and now we need cards that aren't even released just to run these things? How are people not mad?

 

[RobertR1]

Anyway, back on topic. If my 1080Ti cannot at least power the 5K+ at native refresh with details and cars on screen similar to the Oculus, I'll sell my pledge and stick with Oculus. We were told hardware requirements at the kickstarter stage (which I FAR exceeded at the time) and now we need cards that aren't even released just to run these things? How are people not mad?

If you want to sell your pledge, please let me know.

 

[anton_Chez]

I'll probably sell it once I've actually used it and have it here. I mean the choice for me is simple. I need the headset to run at the hz it's meant to. If that means 5K+ then that's what I'll change to. It's going to be miles ahead of the Oculus for screen quality and FOV even at it's lowest levels. I simply cannot justify upgrading my GPU and feeding this machine that is Nvidia. Not for the small amount of games that I actually play these days.

If the head unit is chugging along at 50hz, then it's a paperweight to me. Only way to know is to get it in hand. I still feel completely ripped off on the whole thing. Recommended cards were WAY off, the advertised refresh rate was not able to be met when backers were pledging. Now, just as new cards are becoming available and might be able to relieve some of the pressure the headsets find themselves creating, we get this insane push to make anything even remotely related to gaming priced out of reach of most people.

I can afford whatever Nvida charges for their graphics cards. What I won't do, is be part of the group that made the transition to $1K+ consumer grade GPU's for Nvidia a reality.

 

[Michal Burisin]

I would NOT bank on anything coming from game developers or Nvidia regarding features that improve performance in head units. I am still BURNING at those scumbags that advertised the 10 series cards as having dedicated VR rendering techniques that allowed the image of one eye to be used for both eyes, effectively halving the cost of performance. These are the same people that also advertised SMP, the technique that renders 1 screen across 3 in a triple screen resolution allowing for a much less cost to perfomance. I have had BOTH of these setups and have never seen it implemented in anything other than iRacing. It's developer dependant and absolutely should not be counted on when buying new cards.

I am not buying a 2080Ti. I am done with these ridiculous price hikes and empty promises from what I consider one of my most hated companies. I despise Nvidia's business practice, from the lies of the 10 series 'features' (oh sick, I can take screenshots in game, now where's my improved VR performance?) to the disgusting VRAM debable that the tech journo's said was 'no big deal'. I wish VR would support SLI so I can go and get another used 1080Ti and call it a day. 2 grande for 50% performance increase over last gen? That's what we got for the same price generations past. How else to they hope to sell new cards if the new generation isn't faster than the old one? Now we are expected to pay double? Over Titan prices now for a Ti variant? At launch now? Wow.

Anyway, back on topic. If my 1080Ti cannot at least power the 5K+ at native refresh with details and cars on screen similar to the Oculus, I'll sell my pledge and stick with Oculus. We were told hardware requirements at the kickstarter stage (which I FAR exceeded at the time) and now we need cards that aren't even released just to run these things? How are people not mad?

Well, it's up to developers and that's also dependent on VR adoption. Hopefully people like me who waited for VR to be usable will jump in. The current headsets SDE + other issues was no go for me.

I think now VR won't be only for early adopters from now on, maybe Vive Pro was more usable than first gen, but I would never consider paying this kind of money to HTC and I think many others feels the same.

Once there is more people in VR, hopefully companies will start to put in more effort in VR implementation. Also with next gen cards the performance goes up.

I have no love for nVidia, but what other choice one have nowdays. Hopefully AMD will start competing in coming years again to push the market forward. This happened in CPU market with Ryzens. So hopefully same thing happens with graphic cards and now with VR headsets when Pimax push 8k and 5k+ to the market, where big companies are just waiting with their prototypes in the labs as much as they can.

I would wait for prices to settle down, these preorder prices are just crazy, but as you can see, lot of people are willing to go for it so the demand is there so no reason for nVidia to not take advantage of that... It's just business ultimately.

If you would want to sell the pledge, let me know

 

[anton_Chez]

It's not the support for VR that I am talking about. It's the 'features' that Nvidia use to sell their new generation of products that they have abslotely no intention of implementing at a hardware or driver level themselves. That is to say, that they say "these cards are going to be 100% faster in VR because of clever rendering techniques that reduce the processing power required", without even putting in the fine print "IF the developers spend thousands of dollars on manpower to implement these features on a game by game basis".

It should be, how do you say it.... "It just works".

Remember that one? Well, what works is that Nvidia lie and continue to sell products that never live up to their advertised advantages over the current generation. I've bought all of the their products over the past few generations with the specific intention of utilising the 'features' they touted. None of them have come to pass. The cards should do these things natively, without any intervention or implementation needed from software developers. Can't do that? Don't sell it on the pretense that it can then.

But hey, at least the board partners listen to the RRP Nvidia markets their new cards at for reviewers, when the WORST card you can buy is the reference (sorry, I mean 'founders edition') card for $100 more than what the AIB's are supposed to sell their custom cards for. So genuine.

I'd not be surprised if by the new year I've made room for a set of Freesync triples and a pair of Vega 56's. So done.

 

[Michal Burisin]

It's not the support for VR that I am talking about. It's the 'features' that Nvidia use to sell their new generation of products that they have abslotely no intention of implementing at a hardware or driver level themselves. That is to say, that they say "these cards are going to be 100% faster in VR because of clever rendering techniques that reduce the processing power required", without even putting in the fine print "IF the developers spend thousands of dollars on manpower to implement these features on a game by game basis".

It should be, how do you say it.... "It just works".

Remember that one? Well, what works is that Nvidia lie and continue to sell products that never live up to their advertised advantages over the current generation. I've bought all of the their products over the past few generations with the specific intention of utilising the 'features' they touted. None of them have come to pass. The cards should do these things natively, without any intervention or implementation needed from software developers. Can't do that? Don't sell it on the pretense that it can then.

But hey, at least the board partners listen to the RRP Nvidia markets their new cards at for reviewers, when the WORST card you can buy is the reference (sorry, I mean 'founders edition') card for $100 more than what the AIB's are supposed to sell their custom cards for. So genuine.

I'd not be surprised if by the new year I've made room for a set of Freesync triples and a pair of Vega 56's. So done.

I don't watch too much of marketing mumbo jumbo, but I'm quite close to computers and regarding implementation of new features I don't expect it to work out of the box without developers, because it's just not possible.

If nVidia could get even more advantage over AMD, they would certainly do it, but this is something that's not technically doable and it's up to developers. So as I don't watch marketing too much I'm not really that surprised it didn't get implemented and if I would want to blame someone, it would be developers. But as the market isn't big enough, I can't blame them either... It's about the money again, it's just not worth it to them yet.

 

[RobertR1]

I can afford whatever Nvida charges for their graphics cards. What I won't do, is be part of the group that made the transition to $1K+ consumer grade GPU's for Nvidia a reality.

I can completely see where you're coming from.

For me, a 980ti even with rift is a compromise for AC and I expect even more for ACC so I can either go 1080ti or just jump up. Outside of simracing for a few hours a week, I don't play anything else so the upgrade is just for that that use case plus a Pimax or some other next gen VR in the coming months.

Ultimately, it's a combination of consumer drive, lack of choice and lack of competition that gets us here but there is a legitimate technology jump here for a change rather than the normal speed bump.

From a technology standpoint and having been in the tech industry my whole time along with 20yrs of building my own PC's, there's always a promise of the "next great thing" with each new gen of tech. At times, some of it does pan out. Great examples being AA/AF/FXAA/HDR all started out of a promise that turned into something real and fruitful. Other things such as hair works, not so much. It's important to note that ray tracing is also part of the DX pipeline so this is more of an industry effort than just nvidia doing it's own thing.

The RTX series does have real things in there that could move the ball forward. I'm already happy to see 20+ games having DLSS support before the card even launches. Ray Tracing is certainly the future and hopefully the RTX series is the beginning of it like 6800 series was of FP32 HDR. You also have to take into account the massive die size of the newer chips which is a gamble for nvidia.

They could have just done clock bumps with a similar price and pocketed the margins but instead they saw an opportunity put a massive chip out there with all sorts of new tech. I'm more inclined to support this attempt to move the industry forward than just getting a faster iPhone.