Apple Introduces ‘Metal for VR’ in MacOS High Sierra | Hacker News Reader

Apple Introduces ‘Metal for VR’ in MacOS High Sierra

techcrunch.com

73 pointssalimmadjd22 comments

I wonder how many wheels they'll reinvent before they realise they should have embraced Vulkan.

Feels like such a missed opportunity to really stick it to Direct3D.

Metal was released before Vulcan, IIRC.

It was, but they knew even then that the vendors were working towards pulling their compute platforms into better graphics APIs. Mantle (which became the foundation for Vulkan) was being tested a good year before Metal was out the gate and was very much designed for bridging console and PC and devices.

Since then, Vulkan has been leading the way, both in API development and broad hardware support. The later is important if you're a developer considering your market.

I think at some point being stubborn here is going to hurt Apple. But drawing it out is hurting all of us. A bit of rare unification behind an open API would help its quick adoption and make cross-platform 3D development a little easier in the future.

>it has partnered with Valve to bring SteamVR SDK support to the Mac

Interesting to see Oculus drop the ball like this. Perhaps further evidence that Oculus has little interest in PC VR in the long run and a mobile Facebook VR is its endgame. Rumors claim all of Carmack's talents have been used on their upcoming mobile product, so it sounds plausible.

That said, its wonderful SteamVR is now the defacto standard for PC VR.

Given the R&D firepower of Oculus and Facebook, they are in a better position to execute. I'd bet on them winning the VR/AR/MR race. I think with continuing hardware developments, PC VR will plateau and become a commodity and the real battle will be on mobile VR where Oculus will win. If I had to bet, Valve may never venture into that space.

>where Oculus will win

Daydream is going to be hard to beat. Google is poised to release standalone Daydream headsets soon. They control Android. How many people bought the "Facebook phone" compared to the many other Android phones available? I think Facebook is in far more trouble in mobile than you assume. There are no givens right now, but I do agree thats the arena they should be competing in.

My gut feeling is that Facebook VR will just end up being a Daydream app and a Facebook branded Daydream headset.

Last set of numbers I saw, Samsung Gear VR sold 5x Daydream and PlayStation VR sold 6x Daydream. The new HTC U11 isn't even compatible due to lack of OLED. So, unfortunately, it turned out to be a bit of a flop. The biggest players, Samsung being the largest Android manufacturer, have their own headsets and the smaller players like HTC aren't even going all in. The hardware wasn't even well designed. Covering the back of phones with cloth led to overheating.

I have Daydream, and I don't think it's going anywhere.

Google is ruining it in stupid ways. It has been deliberately made incompatible with all Cardboard apps. I get that Daydream is technically much better, but it has very little content, and there are tons of VR apps made for Cardboard.

Now my Daydream is gathering dust. I try to use it so rarely, that every time it I try, it refuses to run, because I have to install yet another controller firmware update. It's is a slow process and has about 75% chance of getting stuck and never finishing.

Worth dropping this article here, too. https://9to5mac.com/2017/04/11/hands-on-powering-the-macbook...

I thought they would introduce their one VR platform, since they have enough brand/market power and relationship with developers to do so.

They also showed off AR working with iOS 11, and I was a little surprised that they didn't have any head-mount options.

This is just a dev kit. There's not really any point for Apple to get into consumer VR this generation. A dedicated VR space with room sensors and a beefy PC is still a hard sell to consumers. And mobile VR can't do high quality positional tracking yet, which is a requirement.

Once they can put high quality positional tracking in an iPhone, the whole problem is basically solved, they can throw it in a GearVR-like chassis, add VRKit to iOS and instantly be the biggest VR platform in the world. They'll simultaneously be the biggest AR platform in the world with hand-held iPhone-based AR. They're technically already there with Pokemon Go.

One can quibble about the handful of areas where iPhone can't compete with other more specialized AR/VR hardware (head mounted AR, high performance realtime graphics, etc) but content creators are adept at overcoming limitations like that. An iPhone 8 or 9 with positional tracking will be basically Good Enough for any <1hr experience you could think up. Content creators will adapt to the device market, and the best content will run on iPhone, probably iPhone first, with lots of indie experimentation happening on PC but big brand name experiences generally starting on mobile.

I'm hard pressed to think of a seven figure audience who would need head-mounted AR or high performance realtime graphics. Airplane mechanics, gamers, etc, will stick to Windows/PlayStation, but all of the mass market stuff will be iPhone first. Even if the iPhone is mostly limited to composed prerendered light fields, I don't think that's going to hurt them. PC gamers overestimate the importance of GPU flops because current VR engines are new and unoptimized. Once we get the next generation of engines, hardware won't be nearly as big of a limitation.

As for competition, maybe Google somehow gets in their first, but I'm just not sure they can package it as slickly as Apple and get the content creators on board. We'll see, Google is trying to pick up those skills. I don't see Oculus or Valve really competing in the mass market space because allegience to gamers is in those companies DNA.

I'm super excited about Vive and general VR support for Macs. My interest in VR is not airplane mechanics, gaming, or anything of that sort. I'm just interested in the VR experience. Conceivably, there'll be many VR documentaries, VR puzzles, etc. as there already are. To see that I can buy a Mac and not be locked out of the VR revolution is a relief.

> Once they can put high quality positional tracking in an iPhone, the whole problem is basically solved

Not a big fan of Apple, but AFAIK they have a decent camera on the back of that thing capable of high-FPS video capture (google tells me on iPhone 7 it’s 240 fps for 720p video).

You fix e.g. 3 balls in your room near you, from that image it should be possible to track position by looking at the video stream. In robotics, balls like this are sometimes used for similar application: http://www.precisionballs.com/Satin_Finished_Balls.php

You can even go without special objects in the view, if the background is contrast enough, ain’t too far and ain’t too close.

Not sure about latency, though.

This camera tracking is already being used for the ARKit object placement and plane detection announced today.

https://developer.apple.com/documentation/arkit

Right, the idea is kinda obvious.

What's most surprising is Apple's endorsement of eGPUs via Thunderbolt 3 - given only a few months ago Apple seemed to be simply aghast at the concept of external GPUs: https://techcrunch.com/2017/04/06/transcript-phil-schiller-c...

> Matthew Panzarino (TechCrunch): What’s your philosophy on external GPUs?

> Craig Federighi: I think they have a place.

How do you get 'aghast' from 'I think they have a place'??

I meant to link to another piece that described how it looked like Apple explicitly disabled eGPU support in TB3 on Macs, instead of merely not-supporting it.

Kind of seems like maybe you were adding your own narrative to this, because Federighi's comment doesn't seem anywhere remotely aghast. It's a neutral response, and in hindsight we can see that he just didn't want to give too much away.

Historically speaking, an Apple executive acknowledging that a product or concept they haven't yet embraced as "having a place" is actually very nearly an endorsement of it.

Aghast? They built the Mac Pro entirely around the concept. That's why it shipped with a bunch of Thunderbolt ports but non-replaceable GPUs: because they thought external GPUs would become widely adopted.

If they're aghast at anything, it's how wrong they got the pro market. Everyone still wants user-upgradeable internal GPUs.

Reply on news.ycombinator.com