12th August Update - Discussion


#1

Continuing the discussion from [Updated 0812]Pimax 5K/8K FAQ:

Interesting, so tracking was an issue after all. There was quite a bit of speculation if that was the single BS demo or another issue.

Good to know distortion’s are reducing and the idea of using eye tracking to reduce it dynamically is a neat idea! I assume that could also work if foveated rendering was also implemented on the eye trackers?


#2

Yeah, that update (while it’s always nice to get some news) doesn’t fill me with confidence.
It starts at point one which indicates in a roundabout way that the M2 may not be the final hardware yet.
The 90% possibility of shipping ‘most’ backers their headset before Christmas will mean that retail will be easily into 2019… jeez :confused: And the accessories, even the base stations, may not even be ready yet either.
Controllers definitely don’t sound like they’re anywhere close to being ready yet. So if you have a “full package” pledge and you don’t have steamvr lighthouse / controllers yet, well …
At least it seems that the M2 might be shipped to the testers this week or the next. Hopefully those will be without NDA…

Overall, more delays, it seems. That’s unfortunate :frowning:


#3

I think it is more honest this time, even if its vague it is better than “all great, we finished yaaa!”

If somebody gave me a 90% chance of anything, i would be in the 10%, it seems to be my challenge in life haha. The M2 reveal will be the biggest reveal they have done to date. A lot rides on that. We want “OMFG this is so worth the wait” kinda responses then the world will be right again. :wink:

As to the peripheral delay, maybe you can get the HMD before the base stations are available, then you can use Inside-Out while you wait. So hopefully they will work on hand-controllers as the priority or at least do an Oculus and enable XBox controller support while they are being worked on. Not ideal but a compromise that allows you to experience the HMD with a few apps that support controller.


#4

Pimax 90% chance is more 25% chance in the real world.

I’ve learnt to not believe 90% of what pimax says


#5

That’s kind of my thinking as well. 90%chance to deliver to “most” is not very confident.
Especially since hardware still isn’t for sure final right now.


#6

There is a 90% chance, but there is a 10% chance of that happening they should of but with front size 000000000000.2


#7

If you don’t trust 90% of PI MAX’s 90% claim then your estimated real world chance would be 9% not 25% :smile:


#8

True. But with cup half full, even if most (some) people get it, and those are really happy with it, then the wait would/will be worth it, the ball would be rolling by then. A good working HMD has to be a priority for backers vs when they get it?


#9

Accepting fault as they have done is better than more excuses. Think how difficult it must have been in board meetings to secure the extra time and expense that comes with the extended development. We have not paid Pimax any more money, they are footing the bill here, probably no small amount either.


#10

Hi @xunshu do we have any update about the headstrap? will it be showed at the Meetup?


#11

Of course. After all, you can only launch once. Once reviews are out for the 8K, then that is it, it’s done. Even if afterwards big changes happen, the damage will already be done.
Nevertheless, even though I’m usually an optimistic and positive guy, I’m not holding my breath.
I had planned to preorder and get it when retail opens up but that’s gonna be easily into 2019 now and I’m not gonna wait that long. Who knows what will already exist by then from other manufacturers.

So yes, waiting with release even longer has its benefits but we’re getting to a point where it’s gonna do more bad than good.


#12

@xunshu - some things missing from the update.

  1. Can you provide some brief technical details how BrainWarp now works and what state of progress it is in?
  2. Are you sticking to 80Hz, is that locked in now?
  3. Is there a chance the new updated Bridge Chip could be used in the consumer version?
  4. Why is the 5K at 90Hz less demanding than the 8K at 80Hz with the same input resolution?
  5. Can the FOV reduction code (Fixed 170 or 200) work dynamically? so users can input a value between 110 and 200.
  6. Have you arranged any M2 meetup yet? What date, location?

Thanks


#13

Wait for next update in 6 weeks time


#14

An update just as SIGGRAPH 2018 kicks off.


#15

I would be very careful to draw any conclusions with regards to foveated rendering. For the distortion algorithm choice you measure the eye-movement once in some way; which is not really clear to me but I don’t need to understand. Possibly linked to a screen telling you to click the mouse/controller when you are looking exactly at a target displayed on the screen, and then doing this over and over again asking you to evaluate which one looks better to discriminate the better from the worse algorithms. And rinse repeat for other positions on the screen.

However, this would not need to be performing near-realtime, if the eye-tracker needs a second to settle in on where you are gazing, fine, you can do so for the algorith testing purpose. For use of foveated rendering for games you would go nuts if the area you quickly are gazing at, e.g. rear mirror in a racing sim, is blurry at first and you’d have to wait for the eye-tracker to have caught up to your sudden move of gaze to a completely different part of the screen. For foveated rendering you need it to work with very low latency.


#16

investors and business partners are probably angrier than most backers. they invested way more than us and they are the one losing time and money.


#17

That is an interesting way. I did not consider it a initial setup step as you describe.

The distortion happens in real-time anyway, and eye tracking cameras can operate at very fast rates, like 240Hz+ so I have no idea if the overhead from a dynamic distortion algorithm would add a lot more to what is already going on?

Actually, we have IPD that should be enough info for the distortion to work right based on your idea, unless that is what they really mean, they will use the eye tracking to automate the IPD which in turn gives the lens distortion algorithm an optimum value to work on.


#18

I’am very dissapointed, things only move very slowly and this year as a baker # 5074 I will not get a pimax anyhow. I would buy at once Star VR if it is consumer HMD.


#19

@xunshu , thank you for an update.

Some questions I would like to ask:

  1. Any information on the comparison of 8K with 170FOV and 5K with 200FOV - performance and visual. I’m not sure backers meetups will be sufficient for most people to make a decision.

  2. If 1070 is not suitable for 8K - can the people have an option to order 5K as an extra unit, maybe like who ordered 8K, may have 5K in their KS backer order and 8K after? Because I don’t think 10 mins demo will be good enough to judge the difference between 5K and 8K and you may have a problem with people complaining after receiving 5K with lower visual quality.

  3. Can we have more information on the 5K model, mainly the hardware/software differences between 5K and 8K?


#20

I think most are wondering the same thing. Tomorrow we will know after they announce whatever it is. Pimax already have your money so it is not like they will lose out but the investors will want to make sure Pimax has a place in the market no matter what, even if they have to delay again based on a potential StarVR competitor.

One thing that may pop up is Acer’s involvement, they are a giant in display technology and probably have this new bridge chip already for their upcoming monitors. If StarVR uses it too, that would confirm it is a solution to the Pimax Hz bottleneck too.