Sure, livestreaming videos to VR headsets would be great, and the desirable 2x8k@120hz would require 5G. But if that's the use case then there's no need to hurry with 5G since most applications would be just fine with local wifi and haven't proven significant demand.

You can't necessarily prove demand for even retroactively obvious products without the products sometimes. I don't think VR is something you are extremely likely to do on the go.

AR on the other hand has potential.

360 video is not going to take off because it’s a sucky experience. “Real” VR with positional tracking makes you feel like you’re in another space, and has lots of potential for that reason. 360 video does not.

I'm not sure what 360 video even is.

AR stands for Augmented reality, the overlay of information and artificial constructs over the real world in such a way as as one can both perceive the artificial constructs and interact with the real world simultaneously.

Nobody will view this by holding their phone up and looking at the world through their phone screen. Real products in development use glasses. Science fiction depicts contacts, currently impossible, or some sort of implant, even less realistic.

Remember we wrote about going to the moon prior to being able to actually go there.

A 360° video is one which has been recorded with a 360° field of view. Add in software + a head mounted display, and you can turn your head to view the video in different directions.

It's a fun novelty, but it should not be conflated with VR. If you move your head forwards or backwards (as opposed to rotating your head), the environment does not adjust itself to match your new position, because it's just a prerecorded video. Without this ability, you can't trick your brain into believing it's in a real place.

I only brought up 360° video because the GP mentioned "livestreaming videos to VR headsets [at] 2x8k@120hz", and there's no other type of video you can stream to a VR headset†. Real VR, with positional tracking, needs a 3D modeled environment, like a video game. We may be on the cusp of being able to stream traditional video games, but VR needs much lower latency than that—around 30ms for absolutely everything in the chain.

I don't think AR is any different in this regard. If the processing is being done on-device the assets are small and you don't need 5G. If the processing is being done remotely, you've got that same latency problem.

How does 5G benefit AR?

---

† Well, okay, there have been some weird experiments, but they preclude standard resolution measurements. They also require an order of magnitude more data—5G won't cut it.

I'm certainly open to a future with a Google Glass that works well (or even Google Contacts?) and its users aren't widely viewed as "Glassholes." But, in the meantime, a lot of use cases can be explored with devices that many people already own and that aren't considered obtrusive in many contexts the way that computer eyewear is.

AR for businesses(like AR repair manuals, etc) seems like a good idea.

But in that case , downloading the data locally seems like a real option.

And if the AR processing is being done on device—which I think it has to be due to latency, even under 5G—the data to be downloaded really isn’t that large. We’re talking images and 3D models.

> data to be downloaded really isn’t that large. We’re talking images and 3D models.

I think you underestimate image sizes for modern AR/VR applications. For example, good quality light-field can take few gigabytes for one image. Also there is a big UX difference between downloading full image ahead and downloading parts on-demand (similar to modern video streaming - today we never download full video ahead).

At the type of latencies needed for good VR, I’m not convinced any type of streaming from a remote server is going to work. You can’t beat the speed of light.

We’re only just starting to be able to do this with traditional video games (Stadia), where significantly higher latency is acceptible, and even that’s largely unproven.

If I’m wrong, well, I still think we should proove this tech over traditional cable + ethernet connections first, which already exist today.