Do not listen to the eye-tracking/dynamic foveated rendering crowd, the most important thing about the quest 3 is the price, accessibility and availability in different countries. Apart from eye and face tracking raising privacy eye-brows, they're expensive in battery life, price, performance and weight. Lot's of people still haven't tried modern VR yet, but showing people VR is still hard (hint: work on casting).
You will be undercut by a company saying look what we can do for cheaper.
The thing is, eye-tracking the pros from a $299 Quest 3 without eye-tracking are 10x higher than that of a $899 Quest 3 with eye-tracking. Don't for get this is not a pro device and will be targeting mainstream adoption like quest 2 did, Eye-Tracking will be a suicidal PR move.
Only reason Quest 2 was 299 was because they were selling the device without profits. Quest 3 will be at least 399 and very likely 499.
Assume they can do eye tracking at 500-600$, it is definitely something they should go for instead of 400$ without eye tracking. If they can't do eye tracking without going all the way up to 700$ then fine, they can skip eye tracking
I think they want to seem the Quest 3 price at or below $500. If eye tracking would make it cost even $525 I doubt they would go for it, and I also do not think we should expect the Quest 3 to be subsidized. Carmack talked about how the Quest and Quest 2 on average only got about 3 apps bought for them. The most popular apps on Quest 1 and 2 seems to be free apps like Rec Room and VRChat. Many probably also buy the Quest 2 only to be used for PCVR.
With that being the case, it ruins the whole idea behind subsidizing the headsets, which is part of why the prices got raised 100 USD for the Quest 2 devices. So I think with the Quest 3 we will also see a 400 and a 500 USD version, but Mark said we should expect its price to be 300-500. So I guess they might still aim for a 300 USD Quest 3.
As I already mentioned in several posts, I doubt they will be able to include eye tracking and keep its price below 500 USD. Especially without subsidizing it. But they are trying to find a way to include eye-tracking, because the benefits of foveated rendering can be pretty big, and I think especially if they can end up being able to apply dynamic foveated rendering to everything because of it being a standard in the Quest 3. If it remains a "high end only" thing, I do not think it will get nearly the same focus to add dynamic foveated rendering to everything, because they might as well instead focus their time on optimizing for everything without foveated rendering. That might be the case either way as some will decide not to enable it, and it should maybe not be something Meta should expect users to enable by default.
Snapdragon made the announcement for AR2 recently, which is a new version of their AR chipset designed for ultra-compact AR glasses. And guess what? Even that little 1 watt power processor is now coming with eye tracking sub-processor included! This all but guarantees they will feature dedicated eye tracking circuit in XR2 Gen2.
All that remains now is for facebook to include appropriately placed camera's for eyes and they will have eye tracking without having to worry about putting another chip into the headset.
Even if Facebook doesn't put eye tracking in Q3, another company such as Pico will very likely do which would put them at really bad position. It would be huge mistake not to include eye tracking given the processor already supports it
Qualcomm Launches Snapdragon AR2 Designed to Revolutionize AR Glasses | Qualcomm
But isn't there some uncertainty about whether the AR2 will have anything to do with the XR2 Gen2? I hope it will, but it seems to be an assumption that anything from the AR2 might be included in the XR2 Gen 2 chipset. As I understand it, they already completed XR2 Gen2 and the AR2 probably used tech from XR2 gen 1 and 2 to be possible. The concept of the AR2 seems strange to me, because the whole idea of XR is that its supposed to encompass AR, VR and MR. So doing an AR specific chipset seems like a step back? But I guess it might make sense to go that way because an AR headset might have different priorities, like probably having a even smaller form factor than the XR2.
I hope you are right though. If XR2 Gen2 has built in dedicated eye-tracking support, it sure makes it that much more likely for the Quest 3 to have eye-tracking.
There were some news a while ago about a partnership between Meta and Qualcomm about Meta helping to develop parts of the new chipset and that it would eventually be released for everyone, making it possible for Pico and others to use it as well in their headsets... but of course Meta will have the advantage of being first with the chipset and having helped design it they will probably have a better chance of utilizing it better than the competition.
Basically, AR2 is meant to be ultra-compact, ultra-low power. It is the kind of chip you put inside AR glasses which are only bigger than traditional sun glasses.
XR2G2 is what they use for actual headsets like Quest, Pico etc. It has more relaxed dimensions and power requirements. AR2 is meant for glasses only.
So far, we only know that they will both by made in TSMC 4N process. But given XR2G2 has less restrictive requirements, I expect it to support everything AR2 can do.
We will know for sure next month when XR2G2 gets official announcement from Qualcomm
Mark my words, Meta will regret shipping Quest 3 with eye-tracking. It will cost more than 400 dollars, battery life will be shortened to 1 hour. Even if they increase the battery cells (which will make it even heavier). The weight will obviously affect comfort just like the Quest Pro.
All they have to do is provide people a cheap and more powerful Quest 2 improvement with pancake lenses and the growth we've seen with the Quest 2 will continue. The rest of the investments are much more needed in software. After seeing failures of other VR companies and even the Quest Pro, suggesting anything otherwise is moronic. And will see even more billions go down the drain.
Even PCVR/SteamVR is only surviving on the success of Meta's Quest 2. Other PCVR headsets struggle to even hold 40% of SteamVR's user base against a Quest 2 (made to be standalone).
Battery life will be shortened to 1 hour? The Quest Pro even with eye and facetracking last more than 1 hour. But its the face-tracking that is battery intensive. Eye-tracking might help increase battery life with foveated rendering, and it can be disabled.
PCVR is not only surviving on the success of the Meta Quest 2. Its just that the Quest 2 is highly successful, and the amount of sales it has had, overshadows the success of PCVR. Many buy a Quest 2 just to use it for PCVR because of how cheap it is while still being good quality.
When it comes to hardware vs software improvements its not like its one or the other. No reason to not just focus on both.
Foveated rendering + Eye tracking overall results in %10 increase in battery life from research that I did previously. Numbers won't exactly hold same for different vr headsets, but eye tracking won't have negative impact on battery. Worst case you can disable it as Cain said.
And not to mention previously eye tracking was done with help of another chip that needed to be powered. If it comes as a built-in co processor in XR2G2 then energy consumption will be ultra-low
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.