New headset : Pimax Crystal

[lokiss]

Gota wonder why they've taken to the road with the device in such a state. Pimax are a company riddled with reservations against them, not a good look.

It's unfortunate that Pimax are the current vanguard of the industry, on paper their hardware is extremely appealing.

 

[BigBANGtheory]

Gota wonder why they've taken to the road with the device in such a state. Pimax are a company riddled with reservations against them, not a good look.

It's unfortunate that Pimax are the current vanguard of the industry, on paper their hardware is extremely appealing.

I think that is true historically and it would be silly to ignore, personally I like that they show things early so long as they do something about the feedback.

It's really jarring when a software or hardware developer is told straight up something ain't right and needs to be fixed only for them to declare the alpha/beta/early view was great success and they changed the grand sum of eff all for the final release.

Im not saying Pimax will do that, but I am saying we need to check and see evidence.

 

[melmac]

It's due out in January 2023, isn't it? I think that was the plan anyway, End of November 2022 for China, January 2023 elsewhere.

Or do the issues mean delays?

 

[Tykey]

It's due out in January 2023, isn't it? I think that was the plan anyway, End of November 2022 for China, January 2023 elsewhere.

Or do the issues mean delays?

It was originally slated for release in Q3 2022 (so end of September), then they said shipping to the West by Xmas, that was before it was revealed to be in an unfinished state at the roadshows. So your guess is as good as mine. My prediction is it'll be being shipped as early as March or as late as June 2023.

 

[melmac]

My prediction is it'll be being shipped as early as March or as late as June 2023.

If it's June, then Ouch!! There will be newer headsets out by then, or at least there will be more concrete rumours about newer headsets. Will certainly cut into their sales.

 

[Zeeflyboy]

They said the best case is Jan next year, could however slip to after Chinese new year which would probably end up being pushed to later feb/March. Given it’s pimax I assume that probably means April - June!

Out of interest, what new headsets that are on the Crystal’s (theoretical) level are you expecting by March-June? Personally I don’t think they have a big rush and can take the time to get it sorted with the feedback - overall from the impressions I’ve read it sounds like it has the potential to be a really well rounded bit of kit IF they can execute correctly and other than rumoured Deckard specs (with an extremely uncertain timeline), I’m not aware of much in the pipeline that can directly compete with all the Crystal can offer.

I find the feedback I’ve been reading generally encouraging actually, despite the rough around the edges early showing. It seems that fundamentally at the core it is an astoundingly good headset… this early showing has allowed them to gather some pretty valuable feedback before ramping to mass production.

If you take the positive core impressions then it’s basically Index comparable FoV, aero level clarity but with better contrast and colours, zero distortions, decent built in audio… add in the wireless capabilities with hot swappable batteries and it sounds like it could be a dream headset.

In the big livestream with MRTV they also acknowledged their failings and seemed genuinely interested in trying to improve which is something at least.

Eye tracking state doesn’t concern me, it is basically all being handed to Tobii (who are the ones that decided it wasn’t ready to show apparently), and they know what they are doing so I have every confidence that will work great. Controllers are just reference Qualcomm so don’t expect any issues there either once it’s up and running, software stack is definitely somewhere they could work in but again it’s something they have at least acknowledged.

Only big elephant in the room for me is whether that lens diopter issue gets sorted. Again I’ve read mixed impressions where some headsets were apparently better than others so hopefully it’s mostly just pre-production quality control issues rather than anything inherent to the design.

 

[BigBANGtheory]

Eye tracking state doesn’t concern me, it is basically all being handed to Tobii (who are the ones that decided it wasn’t ready to show apparently), and they know what they are doing so I have every confidence that will work great.

It should concern you
- its new ground for Tobii
- Tobii isn't controlling the IPD adjustments they are simply measuring it
- Tobii isn't providing the Foveated rendering they are providing eye tracking data only

What I'm trying to say is Tobii could have the best eye tracking solution in the world but unless its properly integrated by Pimax then it's just a gimmick. All we know for sure right now is they weren't able to demonstrate it on a theoretical product launch which is 30-60days from now. Yeah I'm concerned but if Pimax take their time and get it right we could be singing their praises, track record say they will yeet it onto the market but hey people and companies can evolve... show me it working in a wide selection of VR games.

 

[Zeeflyboy]

The simple fact tobii decided it wasn’t ready to show, and that pimax listened, is enough to allay my concerns. If they weren’t happy with Pimax even showing it in a less than perfect state, they aren’t going to be happy with launching it either. (Joshua himself stated it was Tobii’s decision not to show it during his live stream with MRTV).

- Its new ground for Tobii:

It’s not new ground for them unless I’m misunderstanding you, they are already in the HP G2 omnicept, and also the Vive pro eye, the Pico NEO 3 pro eye and the NEO 2 eye.

- Tobii aren’t controlling the IPD adjustments, they are simply measuring it.

Not sure where you’ve seen this discussed? If so, would appreciate a link. Regardless, accurately measuring IPD is definitely the hardest part imo… accurately driving a lens module to a predefined position using a small motor and an encoder is literally something I could whip up in a couple of hours in the man cave and is very much a solved problem with a myriad of ways to achieve it.


- Tobii isn't providing the Foveated rendering they are providing eye tracking data only:

Again not sure where you’ve seen this, but would appreciate a point in the right direction if you have any info.

It’s pretty clear they aren’t just providing raw eye tracking data however, the foveated transport for example is a tobii technology that pimax are definitely using as it has been mentioned, and I don’t see any reason to assume they wouldn’t also be using tobii spotlight since they work hand in hand and I believe spotlight is a prerequisite. One of the main reasons to go with Tobii is their full software foveated integrations stack https://www.tobii.com/products/integration/xr-headsets/foveation-technology

As for the foveated rendering itself, pretty sure pimax are at least partly going to be planning on using Nvidia’s VRSS 2 framework (hence one assumes at least one reason why it’s Nvidia only). https://developer.nvidia.com/blog/delivering-dynamic-foveated-rendering-with-nvidia-vrss-2/ which if you have a read of you’ll see was developed in conjunction with tobii spotlight and works on the driver level which matches what Pimax have been saying (and is further evidence that this is not new ground for Tobii, and they are doing more than simply providing raw data).

On top of that, OpenXR toolkit can already work with the G2 omnicept (which is also a tobii eye tracking product) and is pretty much universal when using opencomposite to switch openVR games to OpenXR. I’m currently using fixed foveated rendering in OpenXR toolkit and it generally works pretty well, albeit perhaps with slightly more modest performance gains than could be achieved earlier in the pipeline.

Anyway, time will tell but honestly given all the above it’s the thing I’m least worried about in the whole headset

Also seen some encouraging stuff from Joshua on Reddit regarding the lens focal distance/diopter issue that was reported:

Yes we already identified the issue and have started the process of correcting it.

Yes we are actively fixing the issue. On this road-show I was able to identify the issue and feed it back to our engineers. They have confirmed my theory in testing and we are currently debating different solutions. We are debating 3 options, one of which would possibly case a minor delay in delivery time.

So at the very least they seem to be taking feedback seriously, which is encouraging to see.

We’ll find out if they have actually changed their spots when it launches next month… by which I mean March… I mean June.

 

[Unseul]

2024...

 

[drakulton]

 

[Tykey]

About the lens focal issue with the Crystal that Seb and some others found at the MRTV roadshow, it looks like Pimax have fixed it. At the roadshows after the MRTV one they used modified or new lenses and users are reporting the issue is fixed, this includes 1 person who initially seen the issue at the MRTV roadshow. To be absolutely sure Pimax should let Seb use the headset again, hopefully that will happen.

 

[BigBANGtheory]

As for the foveated rendering itself, pretty sure pimax are at least partly going to be planning on using Nvidia’s VRSS 2 framework (hence one assumes at least one reason why it’s Nvidia only). https://developer.nvidia.com/blog/delivering-dynamic-foveated-rendering-with-nvidia-vrss-2/ which if you have a read of you’ll see was developed in conjunction with tobii spotlight and works on the driver level which matches what Pimax have been saying (and is further evidence that this is not new ground for Tobii, and they are doing more than simply providing raw data).

"Although eye-tracking headsets themselves appear to require per-headset integrations to support dynamic foveated rendering with VRSS 2, Nvidia says that applications don’t need to be modified in any way to get the benefits of VRSS 2, provided they support DirectX 11, forward rendering, and MSAA. That’s a good thing because it means developers don’t need to rely on any technology that’s specific to Nvidia GPUs in order to benefit from VRSS 2.

Developers with compatible titles need only to submit their application to Nvidia for consideration. If the application benefits from VRSS 2, Nvidia will whitelist the app to use VRSS 2 in a future driver update."

Right so what that means to gamers in plain English (at least how I see it unless things have changed) is you may or may not be able to take advantage of eye-tracked Foveated rendering in your games. That's why we need to see it and understand what 'we fully support eye-tracked Foveated rendering' actually means in reality to gamers and consumers.

On the other topics I think we are talking cross purposes, my point there relates to Pimax proving that they've integrated the eye tracking properly, it's not that they can't it's more a case of them having a reputation of rushing products out and then solving problems by iterating the products which kind of sucks at the consumer/customer level. They need to show stuff working to earn back reputation and if you're going to charge big $$$ then its got to meet high quality expectations from customers.

 

[Zeeflyboy]

Right so what that means to gamers in plain English (at least how I see it unless things have changed) is you may or may not be able to take advantage of eye-tracked Foveated rendering in your games. That's why we need to see it and understand what 'we fully support eye-tracked Foveated rendering' actually means in reality to gamers and consumers.

Sure, but honestly it should work in almost everything one way or another:

VRSS should be able to work on most games that meet that criteria - even if not whitelisted by Nvidia you can actually turn it on in Nvidia inspector yourself.

For DX12 based games it works differently as DX12 natively supports VRS and doesn’t require Nvidia’s framework in the same way as DX11…

As I said though, most games can work with opencomposite to run in OpenXR and then you can use OpenXR toolkit. That works with foveated rendering and is for all intents and purposes universal (with Nvidia anyway, unfortunately for AMD the OpenXR toolkit foveation only works with DX12 since AMD don’t have a way to implement VRS in DX11 like Nvidia do).

Anyway, my point is that as I showed Tobii isn’t actually new to it at all, and is doing much more than simply providing raw eye tracking data. The impression from the interviews is very much that they are working collaboratively on implementation, and given all my points in the previous post I really have very few concerns about it. Do I expect it to work in absolutely every game out there? No, probably not. But it should be able to work in the vast majority to varying degrees… as more and more headsets come out with eye tracking (PSVR 2 will clearly be a big one) native developer support is only going to get better too and the benefits will continue to improve as it hopefully gets implemented earlier in the pipeline. That’s where having a more widespread implementation like Tobii will probably be better than bespoke implementations too as it’s more likely to get developer support.

My actual concerns at this point revolve mostly around obviously delivery dates slipping, comfort due to the larger form factor vs something like a pico for example… and also their pimax home software features and stability. Without OpenXR for the standalone we won’t get virtual desktop yet, so I fully expect their native Wi-Fi streaming to be a complete mess… wigig just seems to plug into displayport so should be plug and play, but I don’t think we know when that is expected to release yet.

I also can’t believe they still don’t have a proper native PC OpenXR runtime at this point, but you can fortunately get around that thanks to the work of mbucchia and his unofficial PimaxXR runtime.

As tyke mentioned it seems like they’ve already found a temporary fix for the lens issue which obviously gives confidence that it’ll no longer be a problem once the proper fix is implemented, they’ve got local dimming working well which is something even Varjo hasn’t managed to enable yet, and all round impressions seem generally favourable… it seems the number one complaint is that the field of view is small compared to the 8KX but that’s what the 12k is meant for. For anyone that’s been using more typical headsets like G2, Quest 2 etc it sounds like the FoV will be more than adequate.

Anyway like I say we’ll see. Hopefully we won’t be waiting too long to see

 

[BigBANGtheory]

absolutely I hope they build the market leading VR enthusiast headset and in doing so push the boundaries on a level, if they can do that then they can justify the price point and the rest is down to reliability, volumes and availability

 

[Zeeflyboy]

Just saw pictures of sadlyitsbradley trying out a crystal at the roadshow. Will be interesting to hear his thoughts as he’s been pretty openly critical with Pimax in the past.

 

[andybird123]

Only big elephant in the room for me is whether that lens diopter issue gets sorted. Again I’ve read mixed impressions where some headsets were apparently better than others so hopefully it’s mostly just pre-production quality control issues rather than anything inherent to the design.

Given that I had very similar issues with the lenses on the 5k and 8kX (not specifically diopter, but ipd related) which suddenly appeared half way through development of the 5k/8k and they never addressed it, I don't hold out much hope of them getting the lenses right on any of their headsets.

Did they ever get their previous eye tracking working right? That was broken for at least 12 months after release and I stopped following it.

 

[Zeeflyboy]

Given that I had very similar issues with the lenses on the 5k and 8kX (not specifically diopter, but ipd related) which suddenly appeared half way through development of the 5k/8k and they never addressed it, I don't hold out much hope of them getting the lenses right on any of their headsets.

Did they ever get their previous eye tracking working right? That was broken for at least 12 months after release and I stopped following it.

The feedback from the later roadshow with the "temporary fix" had no complaints about the issue, including from one guy that went to the earlier roadshow and had the problem initially... so, who knows. There were zero complaints I saw about the lenses/visuals from the more recent showings, so perhaps it won't be an issue unless it gets re-introduced with the more permanent fix.

From what I saw Brad post on youtube (just comments, no video yet) he said in his opinion it's visually better than an aero, but he also said he feels it's very much still a WIP headset overall which seems like a fair assessment. I think the biggest concern is that either the delivery date slips substantially, or they ship out a generally unfinished rough around the edges product.

I don't have any experience with the pimax eye tracking module, but I know there were definitely some issues - no idea if it got sorted, but if I had to guess I'd probably go with not! That said I don't really think the two are necessarily directly comparable between an add on module by pimax and an integrated from day one solution from Tobii, their troubles with the eye tracking module are probably a big part of what led to the partnership with Tobii for the new headsets I assume.

 

[Tykey]

I don't have any experience with the pimax eye tracking module, but I know there were definitely some issues - no idea if it got sorted, but if I had to guess I'd probably go with not! That said I don't really think the two are necessarily directly comparable between an add on module by pimax and an integrated from day one solution from Tobii, their troubles with the eye tracking module are probably a big part of what led to the partnership with Tobii for the new headsets I assume.

Did they ever get their previous eye tracking working right? That was broken for at least 12 months after release and I stopped following it.

the ET worked eventually with the help of someone not connected to Pimax in an official way, "guppyexpress", who wrote better firmware for it. ive not used it myself but it's said to be usable with the improved firmware.
i agree that this is a different scenario now with someone like tobii being involved.

I don't hold out much hope of them getting the lenses right on any of their headsets.

i can only relay to you the experience ive had with 3 (lol) 8kx's, all have been ok lens and display wise. no ipd issues etc. getting that bit right was as simple as other hmd's for me. but i understand what you're saying as i've read about numerous reports similar to what you say. i don't understand why there is such variance, unless there is variance between pimax hmd's, but as i mentioned the 3 ive had were ok.

 

[Zeeflyboy]

I think VR headsets are a nightmare for consistency generally speaking because we all have completely different faces, head sizes, vision, brains, preferences in terms of which trade offs you are willing to make etc. that’s before you even get into manufacturing variance or changes made through a product’s life cycle.

It’s definitely a common theme that for one person a VR headset has X or Y problem and for another it doesn’t.

Must make reviewing them a nightmare…

 

[melmac]

i can only relay to you the experience ive had with 3 (lol) 8kx's, all have been ok lens and display wise. no ipd issues etc. getting that bit right was as simple as other hmd's for me. but i understand what you're saying as i've read about numerous reports similar to what you say. i don't understand why there is such variance, unless there is variance between pimax hmd's, but as i mentioned the 3 ive had were ok.

I wonder is it something to do with the price of the headset? That they put more effort into the production and quality control of the higher end headsets. And most of the variance in quality happens in their cheaper line of headsets like the 5k models.