John Carmack reviews VR app “Just Relax” | Hacker News Reader

John Carmack reviews VR app “Just Relax”

facebook.com

107 pointstsemple49 comments

The Oculus Home app (GearVR/S7 edge) recently received an update that has literally blown everything out of the water in terms of quality. I simply don't feel like launching an app now, just want to look at the scenery. I didn't see anything about it on HN, so here are a couple of links about the update:

https://www.vrfocus.com/2017/03/carmack-on-oculus-home-updat...

https://uploadvr.com/new-john-carmack-software-doubles-oculu...

> Static images with mip maps or proper prefiltering can go up to around 18 pixels per degree if you really want the most detail in the center.

I'm not very familiar with VR. What does he mean here by "center"?

Center just means center of the rendered scene. I dunno how to explain it clearly, but just think of wherever you are looking at any moment.

Mobile VR works by rendering both eyes to the screen. The result is then seen through lenses. As a result pixel density is highest in the center of each eye. So in this case center means the portion of the screen under the midpoint of each lens.

The title made me expect a well-worded post about the current state state of VR by Carmack. Maybe change the title?

"John Carmack critiques your VR bullshit he could code in 2 hours using less memory"

https://twitter.com/SuperAnimoGIF/status/853641374419341313

If you like forest scenes, there's Speedtree.[1] Speedtree is a commercial terrain and vegetation generator for games. It can procedurally generate huge landscapes in real time, and has level of detail handling so the world will fit in memory. There are some Speedtree demos where you can walk around in a forest. It's like a game with good visuals, except nothing happens. Very peaceful.

[1] https://www.youtube.com/watch?v=f9Z142aOsm4

Did I miss something? Should you give someone feedback on a public post? Writing generically about something you noticed would be fine but to specifically talk about a project, I don't know, I would feel super uncomfortable if it was done to me.

It is a series of public feedback posts; the devs know it will be public. It is supposed to be able to help other people.

Ok, didn't know that. It's upto him how he wants to do it. But really a generic post without naming anyone would also get the point across, just my opinion.

Imagine the exposure you get from such a post.. I think a little bruise on your ego would be worth it.

This is a thing that John Carmack, specifically, does routinely. I'd imagine most people who work with him know he does this, and he was probably in communication with the devs.

At least in the past the devs specifically ask for feedback from Carmack knowing that the review will be public.

Maybe in return for his time, he gets to post it publicly? It could prove useful to others, surely.

Here's what the game looks like: https://www.youtube.com/watch?v=Vfz6_V5Ldyw

I could see this being a great product when combined with those guide meditation apps. Like https://www.headspace.com

I remember one of these meditation tapes I found on PirateBay which combine binaural beats and guided meditation. It had you climbing a lush mountain, coming upon an old looking door in the side of the mountain, and being in this relaxing room with pillows on the ground and you look up and see stars. That was my first exposure to meditation and it surprisingly put me in a very hypnotic/meditative state after the 30 minutes was over. Very good state to get in before starting creative design work. It would be a great experience to translate to VR with 3d sound.

I mentioned Speedtree previously, and that gave me an idea. There's High Fidelity, which is a big, empty open source VR world, sort of like Second Life with few users but better resolution. High Fidelity starts with an empty world of a huge flat plane. Users are supposed to fill it up. Right now, it's very empty.

It should be initialized to be more like the real world. The terrain comes up as unexplored wilderness, with hills, mountains, valleys, lakes, trees, deserts, and vegetation. If you own land (which is a server-side thing), you can build it and clear it.

It would be a nice touch if moving around wore paths, so that paths and roads appeared naturally.

He makes the point about UE having default large packages. Not a huge problem for desktop VR apps for vive and rift, but is a limiting factor.

This is however a major problem for mobile, using both unity and UE and I know it is preventing more mobile VR and AR apps from being developed. Given that carmack is really focused on mobile I'd think oculus would want to create a light weight development/graphics engine.

I'm wondering why oculus hasn't made one especially since they have the skills and capability. Maybe I missed it and they have one.

Building an engine takes years and a large(~30+) team if you want to match UE4/Unity in terms of productivity.

30, or one Carmack.

I'm very familiar with what it takes. We built a mobile AR engine based on OpenGL starting back in 2013 that launched in 2015.

I'm sure FB has more than enough fantastic personnel and resources to build a great engine.

Game engine != Productivity(artist/designer) focused game engine.

I've worked on UE3 based titles, in-house engines and quite a few in-between. Building a new engine from the ground up is easily in the 10s of millions if not more which is why you rarely see new entries into the market.

Sure FB could do it, but it probably makes more sense to worth with UE4/Unity to augment them. You'll be able to draw from a large base of developers/designers/artists who already understand the workflow instead of having to build up that community from scratch.

[edit]

Also the default-large packages of UE titles is mostly a function of optimization, We'd have only ~200mb of system memory back on the PS3 which is way smaller then what modern mobile platforms have available to them. UE has a bunch of things you can do to pull them down, it just takes work.

I don't disagree with any of that. However the point is it would be huge for the development community but a relatively small project for Facebook.

Would solidify platform lock in even further.

Why do I want a platform lock? As a gamedev that's the worst thing that can happen unless there's substantial compensation on the other side of the equation.

Well we don't, but facebook certainly does, which is why I think it's in their interest to do so,

Sure you made an engine, I doubt it has 1/50th of the real feature of the top engines. Can it import layers for photoshop or do you have to save as .PNGs. Can you import Maya, 3DS Max, Blender, and others or only some exported format like Collada. Do you have support for Right to Left text, All of Unicode? Do you run on 15 platforms including things 3DS, Switch, ... Do you have an IDE non-programmers can use? Does it have undo? Is it localized into other languages? Does it have a plugin system? Can those plugins be loaded and unloaded while it's running to do you have to exit, recompile, and restart to use plugins. Do you have shader editors? Animation Edtiors? Can you apply different animations to portions of a character? Do you have physics? How about phyiscally based animations?

I'm sure you're going to tell me "yes" to all of those but sorry, as someone who's made several engines myself NES, Amiga, 3D0, DOS, Windows, PS1, PS2, PS3/XBOX 360, I know from experience that just "making an engine" is like 3% of the real work involved in getting to the level of Unreal and Unity.

It's very easy to "make an engine", it's a lot harder to fill it out with features and tools.

Right which is totally out of scope of Facebook's capabilities right?

Give me a break. You can pile up an infinite list of capabilities for any kind of ide, engine, editor etc... Point is, if anyone can do it its Facebook and it would be great for the dev community.

Take that chip of your shoulder for a minute.

There's no chip on his shoulder, all of those things listed are stuff that UE supports out of the box. There's also a ton of things not mentioned like the cinematic editor, scripting system, cooking process, etc.

Honestly even if Facebook could, I'm not sure I want them to. I've got enough privacy concerns around what they do on the web. The last thing I need is to have a game engine that's used to mine every piece of data then can from me. When you have a non-games company controlling a platform their concerns are going to be above what's best for the content creators(see also: publishers).

is it beyond facebook's capability? no. Is it beyond their will? Yes.

I was once naive, I joined Google to make a game engine. My belief at the time was like yours. It's Google, they're a giant company with infinite resources. There's no way this won't take over the game engine world.

it wasn't until about 14 months in I realized that competing with Unreal and Unity wasn't just a mater of 30 people building an engine. to actually succeed it would take more like 300-400 people and a mandate for those people to form an everlasting division within the company of support. people to make an IDE, people to continually update the engine and IDE, people to hand hold devs. people to make 100s of videos. people to run classes. people to organize conferences. people to organize game jams. people to do bizdev. people to woo the top devs. people in all the major game dev hubs in the world, people to localize docs, etc...

this works for unity and unreal because selling their game engine and services is their core mission. for Google (and for Facebook) all it would ever be is some open source code of some under featured engine they ship once and forget. Unity has about 1000 employees. Unreal has at least 250. Their entire missions are to make, sell, and promote those game engines. That will never be facebook's mission which is why it would never succeed.

All of what you say is totally reasonable and I don't disagree with any of the thrust. That said, I'm arguing that it's in Facebook's best interest to build the engine and further I think it would make it drastically easier for the dev community if something like it existed.

I talked with the Unity founders at GDC and they have very little interest in putting resources around mobile right now , so I think there is a hole in the market that will be really hard to fill by another startup. Hence, why I think these platform owners would be best to tackle it.

As an aside, I think it would be more optimal if there were a third party focused on a mobile VR engine - but again, as we agree it's such a huge project, with really difficult distribution pathways, that I think it's too big of a mountain right now for an independent group. So we're gonna be waiting around a while for that.

Ha, well it looks like I was right, except instead of an engine for VR they made it for AR.

He makes the point about UE having default large packages. Not a huge problem for desktop VR apps for vive and rift, but is a limiting factor.

For who?

That's a serious question.

I have fast wifi connectivity at home, at the local coffee shop, at my local bar, at the local libarary, etc.

I have a limited cell phone data plan, and I could download that app 16 times in a month for free.

For who is it a limiting factor to download a 191 megabyte app?

Old-timers who calculate how many multiples that is of the RAM of their first 8-bit home computer? The under- or non-developed world?

Or am I so spoiled that I don't even realie that 191 megabytes is a thing?

Or am I so spoiled that I don't even realie that 191 megabytes is a thing

I think that's the most likely option honestly. I don't mean that to be snarky, by the way. Most of the dedicated VR users now are still largely in the early adopter space and are heavily high end PC users - so they generally have higher levels of technology ubiquity.

To reach the vast majority of people, the costs, in both money and accessibility, need to drastically come down. Practically, for facebook and other developers, that means reducing the overhead on these systems dramatically.

Mobile is the nearest term approach to this - though still unsolved. Again, this is why Carmack is focusing heavily on GearVR development and says so himself [1]: "There is no way that a PC peripheral or a game console winds up being the path that billions of people interact with."

[1] http://www.reaxxion.com/7793/carmack-announces-that-oculus-r...

I tried to take a frame snapshot with Snapdragon Profiler to be able to make specific comments, but that didn’t work because the app doesn’t have the INTERNET permission listed in the manifest, which is required to communicate with the profiler.

Now that's a problem I would have never even imagined being possible.

It looks like Illusion's Sexy Beach, only without the sex.

I expected this to be an insightful post by Carmack about the current trends and concerns in the VR industry. It turned out to be what sounds like an entirely predictable review that could have been written by anyone with a bit of technical knowledge of a poorly designed VR app. Not very interesting. Don't really understand why this is getting so much attention.

I think it is awesome Carmack takes the time to do this. He clearly cares about VR and the experience users have with VR, even on seemingly trivial apps such as this one. I imagine most developers in this space would love to be peer reviewed by Carmack. Some of his comments seem to be nuanced rendering suggestions, I'm not sure anyone with a 'bit of technical knowledge' could provide this level of feedback.

It would be nice if Facebook's register an account pop-up didn't occupy 1/4 of the screen.

Agreed. Unfortunately, when you take your social network public, investors demand growth. When user growth plateaus, it seems FB has chosen to be more aggressive in forcing you into their network ^.^

I don't understand why growth is measured as a percentage?

Obviously the growth rate for Facebook can NEVER be greater than the growth rate of the population of people born into households with internet. And that can be nowhere near 20% YOY even.

It's what is responsible for the stupid fallacy that PC is dying. It's not dying - it's has more longevity. A machine I built in 2008 is still going strong. Only had to upgrade the CPU once and switch out a RAM module. Many other people are the same.

Also consider this scenario:

1. You have a world of 1 bn people.

2. A company has 0.5 bn of them as customers.

3. The population increases to 1.08% (https://en.wikipedia.org/wiki/Population_growth) = 1.0108 bn

4. The company grows to 1 bn people which is a 100% growth.

5. The population increases to 1.02171664 bn.

6. The investors say that growth has slipped from 100% to 0%.

That's stupid. The growth has stagnated because there IS NO ROOM to grow!!!

Am I missing something?

New debt issuance comes with an interest rate expressed as a percentage. The system has insanity as a core feature.

If Facebook went away tomorrow, no one on the planet would be adversely harmed. IMO, investors price in the fact that most social networks are ephemeral. Grow or have stock price hammered into dust.

I would argue that productivity would increase and social stress would decrease. I can't think of any harmful effect of Facebook disappearing except for a few lost jobs.

> the growth rate for Facebook can NEVER be greater than the growth rate of the population of people born into households with internet

> Am I missing something?

People who are already born but not a member of Facebook?

Households who newly have access to the Internet?

If in doubt, archive the facebook post. http://archive.is/mdDuI

The mobile view is slightly more readable https://m.facebook.com/permalink.php?story_fbid=195070274849...

You can use uBlock Origin to hide the pop-up frame. You won't be able to click through it, but it won't interfere with reading.

You shouldn't have to, Facebook is a poor medium esp. in the context of HN as it is a pain to browse/use on purpose if you don't make an account.

I got a huge popup blocking everything asking me to register before I could even start reading it. Stopped bothering at that point

I don't see that on Firefox w/ uBlock Origin and uMatrix

Reply on news.ycombinator.com