Quest 3 Link App Issue

[firdimigdi]

20 minutes ago, Aurelius_IL2 said:

You (and I and everyone else) are now routinely failing the Turing Test

 

More like victims of enshittification.

[WWCraven]

Here is an interesting article that a squad mate posted on our forums. It doesn't go into any detail regarding the Link connection issue but does highlight another issue that has popped up for me in the last few days. The double image judder made me think I'd screwed up my settings somehow or I'd neglected to check the Occulus Debug to ensure that ASW was disabled (it was). I was actually getting ready to clear everything off and start all over again with a fresh install of Open Composite, Open XR and Link until I read the article and decided to wait a bit longer before going stupid.

 

I also noticed that Link app is at Ver 77.

 

https://www.uploadvr.com/meta-quest-developers-performance-issues-bugs-regressions/

 

Quote from the article;

Quote

 

Another issue emerged through an intentional but misguided decision from Meta.

Quest 3 & 3S offer a Battery Saver toggle in their settings. Since its debut at the launch of Quest 3, this feature capped the refresh rate to 72Hz, forced on fixed foveated rendering, and reduced the brightness to 50%.

But since v76, Meta silently changed the behavior of Battery Saver to instead force the refresh rate to 90Hz but cap the app frame rate to 45FPS. This, to be clear, happens regardless of whether or not the app uses Application SpaceWarp. In the case that an app doesn't use AppSW (most do not) this causes a distracting double-imaging judder that makes many people feel sick, and these customers have been blaming the developers of the apps.

This change to Battery Saver was not communicated to either developers or users, and Meta has now told developers that it's reverting the change, though as of the time of writing it still remains. What's most puzzling here is how anyone with decision-making power at Meta could possibly have considered this a good idea in the first place.

 

[TheSNAFU]

6 hours ago, Aurelius_IL2 said:

 

 

That is incorrect, ... You have been contacted by the exact same transformer neural network under eight separate guises most likely (this is the same LLM, .. large language model transformer AI). My guess is that no real human being ever "contacted" you. You (and I and everyone else) are now routinely failing the Turing Test

 

And yes, ... the PCVR "team" at Meta really has very little clue.

Yeah I wasn’t sure but suspected Alice, Eric and the other 5 or 6 were ai bots. They were dumb as rocks in any case and I stopped responding to them weeks ago. And Suckerburg wants meta to lead the world to ai bff’s lol. Those idiots can’t make a headset talk to a pc and they want to lead the world in ai. It’s laughable. 

[chiliwili69]

20 hours ago, Aapje said:

But over the cable you can transmit a higher bitrate with lower compression (H264). That provides higher quality

 

The H265 at 200Mbps vs H264 at 1200Mbps for VR is a well treated topic in the forums. You can find all kind of opinions.

My take on that is that the differences are really small and the placebo effect plays an important role.

 

I use H265 because in a reddit chat I had with Guy Goddin (the VD creator) recommended me for VR over the H264. But he can be wrong. I have not tested myself the quality of the H265 at 200Mbps vs. H264 at 1200Mbps with IL-2 in VR, which is what it matters for me.

 

Here a comparison I found in  reddit:

https://www.reddit.com/r/virtualreality/comments/14nye29/150mbps_vs_400mbps_vs_airlink_vs_link_cable_quest

 

I am not an expert in video codecs, it is a really complex matter where you can spend a whole life designing better algoritms.

 

I also asked this to my best friend, and this is what he said (advise he could be also very wrong as well), this is the answer:

 

At those bitrates, H.265 (HEVC) at 200 Mbps will almost certainly provide better image quality than H.264 at 1200 Mbps for the same content, assuming both are properly encoded.

Here’s why:

1. Codec Efficiency

• H.265 (HEVC) is 50–100% more efficient than H.264 (AVC).

• That means H.265 can provide the same quality at about half (or less) the bitrate.

2. Bitrate vs. Codec Efficiency

• You might think 1200 Mbps (H.264) is way more than enough. But that only helps if the encoder makes good use of that extra bandwidth.

• In practice, H.264 at high bitrates tends to hit diminishing returns in quality (especially with simple or noisy content).

• H.265 at 200 Mbps is already a very high bitrate — it’s more than enough for most 4K content with excellent quality.

3. Visual Quality (subjective and objective)

• H.265 uses more advanced techniques: better motion compensation, larger CTUs (coding tree units), improved entropy coding, etc.

• So even at a lower bitrate, it can produce crisper edges, fewer artifacts, and better preservation of detail compared to H.264.

---

TL;DR:

✅ H.265 at 200 Mbps will provide better image quality than H.264 at 1200 Mbps, assuming you're using high-quality encoders.

[chiliwili69]

1 hour ago, chiliwili69 said:

I have not tested myself the quality of the H265 at 200Mbps vs. H264 at 1200Mbps with IL-2 in VR, which is what it matters for me

 

LOL! I have just realized that when I was doing this comparison of Questlink (cable) and VD (wifi) I was doing in fact a H264 vs H265 comparison!!  LOL!

 

 

15 hours ago, Aurelius_IL2 said:

It is my understanding that the Snapdragon XR2 Gen 2 chip in the Quest 3

 

I don´t know exactly what are the true limits of the chip, but what it is certain is that for HVEC (and AV1) in the Quest3 hardware is 200Mbps. In the first pico4 is was 150Mbps. In the Pico4 ultra it si 200Mbps. I don´t know if this is a limit chosen by Meta/Pico (for performance reasons) or it is a true limit of the chip.

[Aapje]

@chiliwili69

 

I don't understand how 50-100% more efficient leads to the conclusion that 200 Mbps H265 beats 1200 Mbps H264. That is way more than a 100% difference in bit rate.

 

I don't disagree that H265 beats out or matches H264 at higher bit rates, but I think that the limit of that is around 600 Mbps in my experience.

[1PL-Husar-1Esk]

What about H264+ at 500Mbps , how this compare to H265@200Mbps? I need to test it myself,I think.

On 5/10/2025 at 10:37 AM, chiliwili69 said:

 

This is not my case.

I use Airlink (I created a dedicated 5GHz network in the Wifi6 router just for the Quest3) using the H265 10-bit codec (Airlink default codec) at 72Hz and removing the Auto bandwidth and setting it at 180Mbps. Rendering resolution at 1.0 in Meta app and then I use SteamVR at 150%SS. I have zero stutters.

I didn´t find any advantage in using VD (which I bought) so I don´t use VD since it is simpler.

Also, I didn´t need to touch OTT at all.

 

Setting the frequency to 72Hz not only give you more CPU and GPU margin, but it produce a better image quality. For a given bandwidth, lower frequencies allows to put more info in the image.

You don't see any stutters when frames in  game drops below 72Hz (no ASW enabled)? 

Why I don't use air/cable link. Because my experience and this arguments made by Mbucchia 

 

And new futures added recently 

 

Edited May 16 by 1PL-Husar-1Esk

[Aurelius_IL2]

Hey Chili, one thing to be aware of is that Virtual Desktop often reports erroneous information in their banner. For example, if one connects a Quest 3 or Quest Pro (both use 2x2, 160 MHz Wifi 6E chips), VD will report a bitrate of 2400 Mbit/sec. This is incorrect, ... rather VD is reporting the theoretical maximum of the 2x2, 160 MHz capabilities of the chip. In real life, although VD is not displaying it, you will be seeing 1100 Mbit/sec up to 1800 Mbit/sec. 

 

Also, while hardware decoders (meaning the HVEC pathway (IC) built into the SnapDragon) do have true limits that cannot be surpassed (the limit at which overclocking voltage increases produce no more viable gain), I would be very surprised if the XR2 Gen2 chip has it set exactly at 200 Mbit/sec. I will check it with an oscilloscope later this week and see what I find. The Qualcomm, SnapDragon XR2 Gen2 can definitely decode H264 faster than 1300 Mbit/sec and my suspicion is that HVEC can reach at least 600 Mbit/sec or more. I'll check it if I get time early next week.

Edited May 17 by Aurelius_IL2

[Aurelius_IL2]

It has already been mentioned in passing, but just to help new guys out: When you set up your Quest 3 or Quest Pro with Air Link or Virtual Desktop, you want to make sure that one of the bands of your wireless router (one of its radios) is ENTIRELY dedicated to only your VR headset. That means if you have the typical 2024 tri-band router with one 2.4 GHz band, one 5 GHz band, and one 6 GHz band, ... you want to use the 5 GHz or 6 GHz band but you don't want any other clients (cell phones, baby monitors, security cameras, etc ....) connecting to that band. If you live in a family of, ... say, ... four, ... and you get your Quest 3 rolling along smoothly as you fly your FW-190 or F-18 Hornet only to have stutters show up as soon as Junior's phone connects to the same band and begins downloading a 4K movie, ... you're going to be less than a happy camper.

 

(It should be noted for those who are more tech savvy that some higher-end wireless routers allow for very specific frequency and band settings (usually done through accessing their firmware and then heading to the "wireless" settings). It is possible to set your VR headset to use its own channel and if you live in a house with multiple family members all hogging the various bands of the wireless router, you may want to investigate this.  )

Edited May 17 by Aurelius_IL2

[1PL-Husar-1Esk]

7 hours ago, Aurelius_IL2 said:

Hey Chili, one thing to be aware of is that Virtual Desktop often reports erroneous information in their banner. For example, if one connects a Quest 3 or Quest Pro (both use 2x2, 160 MHz Wifi 6E chips), VD will report a bitrate of 2400 Mbit/sec. This is incorrect, ... rather VD is reporting the theoretical maximum of the 2x2, 160 MHz capabilities of the chip. In real life, although VD is not displaying it, you will be seeing 1100 Mbit/sec up to 1800 Mbit/sec. 

 

Also, while hardware decoders (meaning an HVEC pathway (IC) built into the SnapDragon) do have true limits that cannot be surpassed (the limit at which overclocking voltage increases produce no more viable gain), I would be very surprised if the XR2 Gen2 chip has it set exactly at 200 Mbit/sec. I will check it with an oscilloscope later this week and see what I find. H264 across the SnapDragon can definitely reach over 1300 Mbit/sec and my suspicion is that HVEC can reach at least 600 Mbit/sec or more. I'll check it if I get time early next week.

Hi . That very interesting. In my setup VD and information about wifi in quest 3 do report 2400 Mbit/s . Latency is very low - 2 Ms . I'm using 6ghz band for quest only, nobody else is using this router. Computer is connected to router via Ethernet cable and  internet also from fiber optics modem.

Edited May 17 by 1PL-Husar-1Esk

[Aurelius_IL2]

Yes, Husar, the exact chip in the Quest 3 is the Qualcomm SnapDragon XR2 Gen2, which is limited to one receiver with two antennae (that is a MIMO config rated at 2x2 at 160 MHz). That 2x2, 160 MHz Wifi 6E hardware ability is theoretically rated at 2400 Mbit/sec, ... but that is, ... theoretically, ... on paper. In real life, you will if you choose H264 see somewhere around 1100 up to possibly 1800 Mbit/sec depending upon the decibels relative to a milliwatt power of your wireless router and how close you are sitting to it. For maximum wireless transmission bit rates, you want to be about two to three meters from it with its power at -30 dBm to -40 dBm. (You can use NetSpot or another app on your phone to check this and receive detailed output.) Again, it does not matter whether you use the 5 GHz or 6 GHz band (radio, in this case), ... it only matters that your headset is the only client on that band.

 

I cover all this stuff in detail in my newly released wireless guide, which can be found here:

 

The Wireless VR Guide to Flight Simming 1.1

Edited May 17 by Aurelius_IL2

[DD_Crash]

Thanks for the guide it is really good.

[Aurelius_IL2]

Thanks Crash. Hopefully, it can be helpful to a whole bunch of guys!

Edited May 17 by Aurelius_IL2

[chiliwili69]

On 5/16/2025 at 8:39 AM, 1PL-Husar-1Esk said:

You don't see any stutters when frames in  game drops below 72Hz (no ASW enabled)? 

The fps drops below 72fps only when I am CPU limited, basically in very dense scenarios with many AI objects, then yes, I can recognize very well when I am below 72 without looking at the fps counter because moving objects has like double edges

On 5/17/2025 at 1:46 AM, Aurelius_IL2 said:

Hey Chili, one thing to be aware of is that Virtual Desktop often reports erroneous information in their banner.

Thanks, but I am not using anymore VD.

On 5/17/2025 at 1:46 AM, Aurelius_IL2 said:

my suspicion is that HVEC can reach at least 600 Mbit/sec or more

I could not fine any info about that in the web, If that chip can handle H265 at 600Mbps it will be amazing if this is unlocked in other devices which uses the same chip.

[1PL-Husar-1Esk]

1 hour ago, chiliwili69 said:

moving objects has like double edges

That means ASW is active. You can have it disabled but despite it it kicks in . On the fly only keyboard shortcut can turn it off temporarily ( to next FPS drop). My point is when AWS is disabled on link cable/air game do stutter but not in VD.

Edited May 18 by 1PL-Husar-1Esk

[JMax]

In regards of the original posting,has meta released any update to solve the problem of being unable to launch Link as before. And not having to carry out a repair each session?

 

[1PL-Husar-1Esk]

@Aurelius_IL2 I read your paper it's excellent.

BTW look how fast one aspect of wireless VR technology is changing but other not that fast (codecs bandwidth) , we have now wireless HMD ( Play for dream MR) with resolution of 3840x3552 but run what looks like same in terms of performance XR chip inside quest 3.  I wander what implication would it have and if full  potential can be reached.

[WWCraven]

On 5/18/2025 at 4:17 PM, JMax said:

In regards of the original posting,has meta released any update to solve the problem of being unable to launch Link as before. And not having to carry out a repair each session?

 

Not so far, it's still irritatingly there as of yesterday.

[Aurelius_IL2]

19 hours ago, 1PL-Husar-1Esk said:

@Aurelius_IL2 I read your paper it's excellent.

BTW look how fast one aspect of wireless VR technology is changing but other not that fast (codecs bandwidth) , we have now wireless HMD ( Play for dream MR) with resolution of 3840x3552 but run what looks like same in terms of performance XR chip inside quest 3.  I wander what implication would it have and if full  potential can be reached.

Thank you for the kind compliment Husar. Hopefully the guide can bring good information to VR simmers looking to go wireless in the years ahead.

 

As to limitations you mention, ... principally you have the work (in wattage) that any XR chip (like the Qualcomm Snapdragon XR2 Gen2) has to do. The technology exists right at this moment to make a 4x4, 320 MHz, Wifi 7, fully 4K Snapdragon chip, ... the only problem is that you would only get something like 50 minutes before you would need to recharge your headset (unless you use bigger batteries than the Quest 3 is using or go with hot-swappable Lithium-ion).

[1PL-Husar-1Esk]

1 hour ago, Aurelius_IL2 said:

Thank you for the kind compliment Husar. Hopefully the guide can bring good information to VR simmers looking to go wireless in the years ahead.

 

As to limitations you mention, ... principally you have the work (in wattage) that any XR chip (like the Qualcomm Snapdragon XR2 Gen2) has to do. The technology exists right at this moment to make a 4x4, 320 MHz, Wifi 7, fully 4K Snapdragon chip, ... the only problem is that you would only get something like 50 minutes before you would need to recharge your headset (unless you use bigger batteries than the Quest 3 is using or go with hot-swappable Lithium-ion).

Thanks. Btw I read that Windows 10 is not good for Wifi 6 and above, but you mean for using it as hot-spot or wireless connection from PC to wifi router?

Another thing I'm interesting about adding AI assisted package from GitHub to improve the hand tracking in DCS, but I think link you provided is about Unity not that one you mentioned?
https://github.com/handzlikchris/Unity.QuestRemoteHandTracking

 

Edited May 29 by 1PL-Husar-1Esk

[Aurelius_IL2]

Wifi 6/6E or Wifi 7 have not been officially added to Windows 10. There are some hacks and word-around methods where you can "forcefully" add it in but often with poor results. I recommend that anyone wishing to go wireless, make sure that they have Windows 11, 24 H2. Microsoft spent some quality time to make sure that WIndows now supports Wifi 7 correctly, so that is what I am recommending. It will just work better with the latest Wifi 6E and 7 capable motherboards and/or add-in PCIe cards. It also gels better with Virtual Desktop.

 

As to hand tracking, the link is to the guy (Chris) who has provided an AI hand tracking package in the past. At the moment he does not have his AI package up but he has in the past (I know bc I am using it :)). I am beginning work on an AI hand tracking package myself. It will of course use OpenXR but it will be more specific to the Quest Pro and Quest 3 at least initially. Check at the MSFS 2024 VR sub-forum to see the updates as I release them.

Edited May 29 by Aurelius_IL2

[1PL-Husar-1Esk]

2 hours ago, Aurelius_IL2 said:

gels better with Virtual Desktop

That part I don't understand because I'm connected to router throw Ethernet cable and my wireless headset are connected to my router. There is no wireless connection used in windows, how this can make any difference between windows 10 and 11?

2 hours ago, Aurelius_IL2 said:

As to hand tracking, the link is to the guy (Chris) who has provided an AI hand tracking package in the past. At the moment he does not have his AI package up but he has in the past (I know bc I am using it :)). I am beginning work on an AI hand tracking package myself. It will of course use OpenXR but it will be more specific to the Quest Pro and Quest 3 at least initially. Check at the MSFS 2024 VR sub-forum to see the updates as I release them.

Would you kindly provide a link for this?

BTW could it be in the future implemented in VD ? (Those guys constantly work to make whole wireless expirence better to average Joe). You could collaborate with them to benefit all VR community - just thinking).

Edited May 29 by 1PL-Husar-1Esk

[Aurelius_IL2]

The reason it makes a difference Husar, is because Windows 10 is not natively coded for anything above Wifi 5. The processing and passing of the codec (say, H.264 or H.265) is done through a complex combination of the graphics card (hardware encoder), the graphics driver, the CPU, the operating system, and the VR API stack (say, Meta Air Link or Virtual Desktop). You can certainly add drivers into a Windows 10 rig which will give it said capabilities but you will often find that there are small, hand-off problems and other various small non-linearities.

 

As to my hand tracking package, I am only in the first few days of work. I won't be using one of the more popular AI packages but instead will be basing  the work upon a custom tensorial model that requires less training and that incorporates some elements of topology and differential geometry. When I release a more mature Beta, I will post a link. I can certainly reach out to the guys at Virtual Desktop but that is step for later this summer or in the fall.

[firdimigdi]

On 5/29/2025 at 11:35 PM, 1PL-Husar-1Esk said:

That part I don't understand because I'm connected to router throw Ethernet cable and my wireless headset are connected to my router. There is no wireless connection used in windows, how this can make any difference between windows 10 and 11?

 

It makes absolutely no difference in that setup. It would only matter if you were trying to connect to a wifi 6e or 7 access point on 6GHz wirelessly from your PC or trying to set it up as an access point using 6GHz. AFAIK the official reason being that Windows 10 lacks the telemetry output required for a 6e or higher wifi module to be enabled at 6GHz in accordance to regulations and actively prevents any such attempt any more; however, mostly as an oversight and through older drivers, it did allow this for a while and possibly with some not-worth-the-time tinkering still does.

[1PL-Husar-1Esk]

On 5/14/2025 at 7:22 PM, Aurelius_IL2 said:

Chili, ... It is my understanding that the Snapdragon XR2 Gen 2 chip in the Quest 3 can decode at 8K at 60 Hz (see attached spec sheet here: https://docs.qualcomm.com/bundle/publicresource/87-73689-1_REV_A_Snapdragon_XR2_Gen_2_Platform_Product_Brief.pdf).

 

The rate limiting factor of the Quest 3 headset is not the decode ability but the 2x2, 160 MHz Wifi 6E abilities of the XR2 Gen2 chip. It can go far higher than 200 Mbit/sec. Provided you are dedicating one 5 GHz or 6 GHz band from your wireless router to the Quest 3 with no other devices accessing it, ... you can and should be hitting up to 1300 Mbit/sec regularly. You will occasionally see spikes up to 2 Gbit/sec if everything is tuned correctly.

Correct me if I'm wrong , since decoding and encoding specifications for XR2+ are the same as for XR2 but XR2+ does support WiFi 7, it means that  decoding rate can  go higher than 200Mbit/s ? 

 

[Aurelius_IL2]

There is some debate as to whether XR2+ supports Wifi 7. Some say that it does while others say that it can be enabled with a firmware update. I have not investigated the chip much electrically to see what it can do. My suspicion is that a firmware update will not give the chip Wifi7 capabilities. What is more likely is that Meta simply chose the less expensive variant that included Wifi 6E capabilities.

 

Also, from an electrical engineering perspective, Wifi capabilities and Codec capabilities are handled separately on a given chip. The hardware codecs on a chip are separate from the circuitry that governs Wifi performance. XR2+ is an upgrade from XR2 but I would not expect a massive difference with respect to H.264, H.265, or AV1 decoding abilities. 

[chiliwili69]

I think the Wifi technology is not too important for the PCVR streaming as far as it is Wifi6 or higher, and everything is setting correctly.

The bottleneck would be really in the decoding performance in the chip of the VR device.

The wifi bandwidth/latencies is quite ahead versus decoding bandwidth of the chip o fthe VR device.

BTW, in your great Bible of wireless PCVR you said this:

Just to correct that:

Pico4 Ultra (which uses XR2 Gen2) has Wifi7: https://www.picoxr.com/global/products/pico4-ultra/specs

Play for Dream (which uses XR2+ Gen2 ) has Wifi7: https://pfdm.ai/products/mr-headset

[1PL-Husar-1Esk]

1 hour ago, chiliwili69 said:

The bottleneck would be really in the decoding performance in the chip of the VR device.

Yes it's true, this about PfD

 

 

[Aurelius_IL2]

On 6/13/2025 at 2:02 AM, chiliwili69 said:

I think the Wifi technology is not too important for the PCVR streaming as far as it is Wifi6 or higher, and everything is setting correctly.

The bottleneck would be really in the decoding performance in the chip of the VR device.

The wifi bandwidth/latencies is quite ahead versus decoding bandwidth of the chip o fthe VR device.

BTW, in your great Bible of wireless PCVR you said this:

Just to correct that:

Pico4 Ultra (which uses XR2 Gen2) has Wifi7: https://www.picoxr.com/global/products/pico4-ultra/specs

Play for Dream (which uses XR2+ Gen2 ) has Wifi7: https://pfdm.ai/products/mr-headset

 Yes, provided the 5 GHz or 6 GHz band is not being utilized by other clients, the wireless router will be able to supply more information than the headset can decode (due to the Snapdragon XR2 limitations).

 

Nice eye for detail. I did a cursory look at all the major headsets before I finished off the first edition of the Wireless Guide and did not see that info for the Pico4 Ultra. I was not even aware there was a Play for Dream headset  Thanks! I will update this in the next version of the guide (coming out in July most probably)!

 

 

[Roach-]

Hey guys I to seem to be having this issue and cant seem to resolve it. oculus Link works for a bit but then just randomly crashes and freezes, I can close down  oculus Link and then reopen it and it might work for a minute or two and then it just crashes again. Honestly I'm so tired of this meta POS I'm glad I purchased an extended BestBuy warranty so I can get my money back and buy a better headset that isn't so incompatible. I tend to think Meta doesn't really care about pcvr gamers and is really just focused on stand alone games.

[Aapje]

Try a different USB cable, try a different GPU driver, try reinstalling the Meta Quest Link app, etc.

 

As far as I know there isn't a global issue with Link at the moment, so it must be something on your end.

[dgiatr]

1 hour ago, Roach- said:

Hey guys I to seem to be having this issue and cant seem to resolve it. oculus Link works for a bit but then just randomly crashes and freezes, I can close down  oculus Link and then reopen it and it might work for a minute or two and then it just crashes again. Honestly I'm so tired of this meta POS I'm glad I purchased an extended BestBuy warranty so I can get my money back and buy a better headset that isn't so incompatible. I tend to think Meta doesn't really care about pcvr gamers and is really just focused on stand alone games.

It might be your cable. I had freezing issues too and changed the cable and now it works.

[Internethetzer]

I had cable issues as well

Headset had either very unstable connection to PC or none at all

I simply switched off Wifi when attempting to use Link & somehow it works