Google Teaser? #Stadia

Yes it will allow games made with unity to run on stadia. It has nothing to do with unity, its whether stadia itself allows non-gaming applications and currently it has not been announced anywhere that they would. So your better off assuming it will be games only until officially it is stated otherwise. Stadia has not been marketed as anything but a gaming platform, and from researching there does not seem to be an offering or an appetite for it to be for other things. This wont be a platform where anyone can release a game, it will be vetted heavily like consoles were back in the day.

 

Thanks

 

A big load of "not very much yet".
https://blogs.unity3d.com/2019/03/19/unity-support-for-stadia-heres-what-you-need-to-know/

"just allow"... it's not that simple.

 

Stadia needs one of the following two to happen:

1. If I am expected to buy full priced games on the service, they should also have a "normal" launcher, where I can download the games normally. Then I might be tempted to buy something there.

2. If they insist of being streaming only, they need to partner with either Epic or Steam and for a monthly fee you can stream your Epic or Steam games.

 

This will never happen because it's not what the platform is designed for in any capacity.

This will also never happen because it will impact their bottom line too much.

 

Yeah, I wasn't saying it was possible, I'm just saying it's what they need to do if they want adoption.

Although I'm not sure the 2nd option isn't viable, it could be a subscription only option and they could ask for something from the storefront. (doesn't Geforce Now work like that these days?) I'm not saying Google is willing to do it though.

 

If they allowed streaming from Steam or Epic, they couldn't sell those games on their own platform, which means they won't do it at all. There's no way Stadia isn't already a loss leader in its current state, so making less money isn't a thing google will do.

 

Eh, I mean they bought Youtube before they properly figured out how to monetize it. And if they aren't in it for the long game with Stadia, it's dead. (and it looks like it's headed that way).

A subscription only add on to another storefront would have far more adoption than their current business model.

 

They did that ages ago and the entire reason was to get more user data, which is worth way more money than anything youtube does otherwise. That same data isn't as robust if they just add Steam and Epic libraries into the mix.

 

It's not, but it's still better than no one using Stadia, which is where they're heading towards at the moment.

 

They're headed towards that more because Stadia is a bad service that works as well as everyone expected it to. It's not a library issue.

 

Eh, if it was tied to a library I already own, and if I had a compatible Android phone and if I was living in a country that had more accessible high speed internet, I would consider paying the 10$ / month to be able to stream my library at random places.

I realize that's waaaaaay too many ifs in a row, but I think more people fulfill those requirements than the current target audience for Stadia which is... no one?

 

The current target audience for Stadia is people who want to play games, have reliable high speed internet, but do not own a console or PC, and even then Stadia isn't any good because it suffers a lot of stuttering and dropped frames. If you drop below 20mbps, the games become even more unplayable.

Stadia's real issues are tech based, and despite what google says, you can't solve the speed of light.

 

So this is basically the same as OnLive which was active around 2011? Before it went bankrupt in 2012? (Maybe they are using the same technology!)

I was signed up to this service. It was quite good. I played games like Batman Arkam Asylum and BioShock. And also, once the service went bankrupt I lost access to all those games. :'(

One problem was the lag. Unless Google have found a way to combat this I don't know how it will be any better.
I think one difference is more Americans have Broadband now than in 2011.

I doubt very much this will be any competition for Steam. Although personally I would sign up for it if it was about $5 a month and you get lots of free games AAA with that.

BTW I don't think the speed of light is too much of an issue. You could send a beam of light to the opposite side of the world in about 1/15 second. And assuming there will be say 4-6 servers spaced out round the world we could bring that down to 1/60 second. So it would take about 1 frame to send a signal to a server and 1 frame to send it back again. Human reaction time is in the region of about 8 frames at 60fps. So it's almost as if the world is coincidentally the perfect size. (There's no doubt a lot of other electronics that slows the signals down in between but I don't know too much about that!)

 

Lag is one thing, how is image quality, it cant be lossless compressed. I would hate to get artifacts in low lit scenes because of bitrate reasons.

 

The brain is very good at detecting lag. We had a 4 frame (90fps, 11ms per frame) lag between the VR conroller and the gun before we fixed it with throwing some hacks on the physics engine. It was unplayable basicly.

 

Pretty much everyone has reported that the compression artifacts even under ideal conditions are pretty noticeable, made even worse by low light or standard "this ruins compression" situations like loads of particles and the like.

 

This is pretty much assuming that everything always works at the speed of light and also works end to end at that speed, which is 100% wrong.

 

When we have quantum communication this will be a viable solution. Until then not so much

edit: though it seems it can never be faster than ligth anyway

 

That is a very light hearted response to a very serious matter.

 

quantum matter?

 

A photon checks into a hotel and is asked if he needs any help with his luggage.

Spoiler

He says, "No, I'm traveling light."

 

True.
But I mean, actual electronic calculations and stuff, if they've got their own microchips to do compression and stuff presumably work much faster than 1/60 of a second. Considering CPU's work at 1/3000000000th of a second. It wouldn't take many CPU ticks to push something to the port. And once it got to the modem a few more CPU ticks etc. etc. So if you had fibre optics all the way up to your house, I guess the only places where a jam would occur would be in the data packets getting queued up somewhere (not sure how that works!) Or inefficient code in one of the servers.

I'm not sure how packets are sent and how many routers and servers they go through.

So if everything worked perfectly, it really is the speed of light that is the main bottleneck. But for single-player games that might be solved just with more servers in each country. For multi-player games across continents, yeah nothing much you can do! (Maybe build a wormhole?)

 

Or, get this. I have a wild idea.

See, what we do is we send the actual game data to the user and then, instead of having them rely on navigating things like download caps from 7+gb per hour streaming, the inability to play games while offline, or the potential issues with compression and dropped frames due to network stability and switching speeds, they use a local machine to interpret that data.

This could be done on some sort of "device for computing things personally," or a "console for videos game," which they would pay a nominal fee for, allowing them the security of knowing that their games would, generally speaking, just work. I find that this is a much better idea than inventing a problem to be solved.

 

Modern game console are already streaming machine, they just happen to have a very big gigabyte size cache for decompressing the visual from the streamed data

Also stadia won't cover your arbitrary streaming need, it's an old tech they haven't upgrade (yet? where is my NN compression google!) and it's already spread around thanks to onlive splashing the idea on te world:
https://moonlight-stream.org/

Create your own stadia kit exist if you look hard enough. And it seems unity jumped on the bandwagon recently ...


I would play a stadia version that is 640p on browser anyway, I don't need high end, as long as it cost WAY less.

 

Think is, the lowest it goes is 720p and it still eats up your bandwidth.

It's just not worth it, and now we have concrete evidence in the form of Stadia's existence.

 

Looks like the problem was latency after-all. I dont know why they would release it so early although they did the same thing with google glass... almost as if they are intentionally trying to poison the well for emerging tech that might break there monopoly

 

It's more that Google is just very bad at releasing products that they didn't buy from somebody else.

 

I was watching a video where they were saying the latency was anywhere from 150 ms, to 250 which I thought would be unacceptable to gamers because it will depend on how far you are from the server. Google could have waited 5 years until 5g was rolled out

 

It's not just that. There's also severe performance differences between wired and wifi (the one people will actually use). Also, if they waited that long they'd be entering the market up against both Sony and Microsoft's offerings.

 

I'm severely limited where I am by a lot of thing, fortunately we dodge the datacap bullshot ... I would rather be allowed to sell on the global market like google apps instead ...

 

I disagree, even if Stadia fulfilled every promise Google made, it would still be a subpar service.

 

Well, all I can say is that I was one of the people who had subscribed to OnLive and I loved it. And I might be one of the people to subscribe if there are good games on Stadia.

I could play lots of the latest games on my laptop. The ones that worked best were things like Batman Arkam Asylum, and BioShock. The one that worked least was Braid as the you needed precise key presses.

So the games that will work are the ones with nice stories, cutscenes and environments. Because you can get an AAAA quality experience (if it streams properly!). The games that won't work good are those that are more about complicated button bashing and precise timing. Probably even Minecraft would be ideal as you could have the server run billions of cubes and play it on a little laptop.

Also, I had loads of games on Steam and had them stored on an external hard drive. Which of course broke. So I'd be happy with a cloud solution.

I wonder if there could be a third solution. Which would be a way to do game-logic on the client side and graphics processing on the server side. That would be best of both worlds.

Yes not all games would work well. But imagine this as an example: A chess game where the pieces were raytraced and animated! Try running that on your laptop.

 

So... I got mine this weekend (ordered in July). I wasn't super excited about it, but it was cheap, and I like to keep up things. I have to admit, I am fairly impressed. After some initial weirdness signing on (see image), it kicked in and performed very well. At least way better than I expected. No stuttering at all, very, very smooth. The artifacts aren't invisible, but you do have to look for them. Definitely not distracting and not really noticeable when playing.

I am honestly not sure about lag... I haven't played Destiny2 before or on any other platform, I does feel a bit "soft" but I don't know if that is how it is designed. (more often than not I play games with a mouse). And there is a bit of audio delay when firing, but that may be a design thing rather than an actual delay. The click sound is instant, and other things like menus, environment sfx and such are all instant, so I don't know.

For an initial outing, it is pretty solid. Totally playable. For many games the lag would be completely unnoticeable. For a core mp experience it may be sub-optimal (but tons of folks play fortnight on their phones.. so...) . But overall, really impressive. (admittedly, I had pretty low expectations). For environment, I was using the chromcast thingy, on a 75inch UHD tv, and have fat pipes, though I have a ton of stuff running on my network, I did prioritize the connection to the chromecast. I played around with it for about an hour and not a single frame of stutter.

 

I had a go with Stadia on my PC and was very impressed indeed. The FPS was better than anything I could possibly get on my PC ($20 GPU!) and although the compression was noticeable, it doesn't bother me in any way. This is perfect for someone like me who might want to play the occasional PC game.

 

I guess you could say it was

Spoiler

destined to be this way Scott

 

Also, it doesn't look like they're gonna let me put my indie games on it... well not yet anyway.

More importantly I keep forgetting to get the free games from the Epic Game Store.

 

There is tv with neural correction of compression artifact that work very well, why it's not in stadia yet given they are on top of that game?

 

That's done locally though, right? There are also great upsampling algorithms (like the one in the new nvidia Shield tv).

But somehow I highly doubt a Chromecast Ultra can handle either of those.

 

That's kinda the point, with cheap coral TPU they could have deliver a chromecast that handles it, why they didn't baffle me

 

Eh, there are more baffling things than this. Most games are running at medium/high settings anyway. I don't see compression artifacts as the #1 sticking point.

And they could still do that down the line if Stadia develops any kind of base.

At this point though, it seems like it launched as good as OnLive, at best.

 

For Google's sake I'm glad to hear that someone out there has had a positive experience.

 

It's def not the main draw, but it's like they don't bother polishing their product, it's a low hanging fruits since a lot of people have pointed to just that. Especially at the pricing point they go with (subscription + full game price), also if that upscale too they could send less data and consume less bandwidth ... and even interpolate frame on the fly to fake higher framerate.

 

I will never understand why they didn't call it beta.

 

They could keep iterating on it forever, and regardless of when they deliver and in what (realistic) state people are still going to point at its flaws and act as if they're fatal.

Note that Google are in the habit of testing changes in the wild. It could be that they considered it "good enough" to change to that mode of product development.

 

The entire plan was to get out ahead of the offerings that would be coming with the PS5 and Xbox... Twoooooo? That's it. They wanted to be the first major iteration of this sort of product.

Keyword "major." PSNow and OnLive were not major.

Because it needs to work across a range of devices, including older Chromecasts.

 

I wonder how much these new console will future proof the neural aspect (essentially just having equivalent to tensor care, aka MAD register pipeline), because they could very be the last traditional console at the rate at which neural image synthesis is flying. Deepfake can already get rid of the uncanny valley for free.

In fact if google is smart, they should be experimenting with neural augmented game production because that will gave them some huge visual advantage that competitor will struggle to close, as they have huge infrastructure and team competent on the matter. So far I only know of ubi soft taking that part seriously, but right now it seems they just augment traditional pipeline.