38
votes
Google announces "Stadia", a new game-streaming platform with deep YouTube integration
Link information
This data is scraped automatically and may be incorrect.
- Title
- Google GDC 2019 Gaming Announcement
- Authors
- Google
- Duration
- 0:
- Published
- Mar 5 2019
Well, that was a fairly disappointing presentation, in my opinion. There are some neat ideas in there, but the fact that Stadia is a completely new platform that developers have to specifically build their games for seems like the biggest downside to me. If it could just run PC games that would be one thing, but making every developer build a Stadia version is unlikely to get a lot of adoption unless they're going to pay the devs to do it (and probably including exclusivity agreements, like Epic is doing).
Even if it did get widely adopted, it would probably also face a lot of opposition.
I have no doubt that Google has some of the most skilled networking engineers and software developers in the world working for them, but streaming games is a really hard task to get right.
Even if their implementation is flawless, gaming is a hobby that has people actively advocating for data ownership, with GOG and its supporters and their anti-DRM anti-online movement, and its a hobby that's pretty well known for causing a big PR storm if companies screw up.
This wasn't going to be an easy market to break into for Google, and definitely not in this way.
These are all reasons I have zero interest in Stadia, even if they do magically fix the latency issue. Additionally, I see no real benefit in renting someone else's hardware over buying my own hardware.
Very much agree with you on the reasoning for this service specifically, but more generally I think there's a definite price/usage threshold where renting makes sense.
GPUs are expensive, and might easily be sitting unused more than 90% of the time. If a platform can share that GPU between, say, 6 people over the course of a day then there's plenty of scope to cover overheads and make a profit while still charging each of those users less over time than they'd pay for the GPU alone. At least that's how it works in a genuinely efficient market - the suggestion above that Google is already working with Unity to push other competitors out of the game doesn't bode especially well for the
completioncompetition needed to get that pricing down across the space as a whole.Thinking back to this, I realise what I described above is pretty much identical to how mainframe timesharing worked back in the 1960s-70s when the idea of a "personal computer" was a laughably expensive luxury. Everything old is new again, indeed.
I mean a lot of games are built on unity and unreal engine so the porting should be rather quick, but I agree - all the "bonuses" of using their cloud will require specific coding.
I'm not sure the technology is there yet anyways. I'll have to give stadia a try with the new doom game, as I worry about the state of input lag for a first person shooter. It was fine in odyssey but that's because even a delay of 50-100ms wasn't really very noticeable as it wasn't a very twitchy game.
Unity games can not be ported to Stadia:
Emphasis mine.
Unity specifically announced support for it today: https://blogs.unity3d.com/2019/03/19/unity-support-for-stadia-heres-what-you-need-to-know/
I was worried there'd be something like this.
So their original EULA changes against improbable.io were set in place to give Google exclusive rights to stream Unity games.
That seems like a leap. One line you didn't emphasize: "...without a separate license or authorization from Unity."
It's more likely that such a license or authorization may be much more expensive, or be a custom deal. Unity said that improbable.io wasn't working with them in an adequate fashion, whatever the details of that may be. They claimed that they wanted to reach a partnership but Improbable did not. The matter is resolved now anyway.
Suggesting it was all a ploy to give Google exclusivity seems to be more of an accusation than deduction until there's more compelling evidence.
Both Unreal and Unity have support built in for it. Yes, it's a new platform, but it brings a whole bunch of new possibilities. You have to make some changes to your game, but to get the amount of cross-platform-ness that this provides, you'd have to put a lot more effort than you do with Stadia.
Yeah, it seems like a much worse option compared to something like Shadow, which just streams the whole Windows VM straight to you. I understand why manufacturers would want lock in (I've discussed it here before), but I can't see why a consumer would choose a locked in service with a limited selection when they could have their entire existing library instead.
Well that was fun, seems like they're doing their best to leverage YouTube to get people using the service. Here are a couple miscellaneous thoughts.
They say there's no upper limit to the processing power but then spec out a specific instance which seems odd.
Idk why they have to talk about 8K, that shit is so unnecessary.
The machine learning style-transfer was not impressive at all. If i wanted a crappy looking photo filter effect on everything I don't need machine learning.
They mentioned buying games? I don't understand how something like this works without it being a subscription model. Either that or we're going to see a rise in demos.
Is it ironic that the stream itself had 3-4 issues for me?
I don't disagree, but they're talking about the potential for the future. I for one am excited for that eventual possibility of streaming games in real-time at 120fps.
Mine did too, but I have to say during the Project Stream beta, I experienced zero issues.
Did you notice any sort of input latency when you tried project stream?
The subjective experience of people is going to be different, even in the same networking conditions. Some people can play a game with 80ms of lag and not notice a thing, others tear their hair out over 16ms of lag.
I guess what I'm saying is, don't just take some random persons experience as gospel, what feels good to them could feel like crap to you (or vice-versa).
I think it largely depends on the game, too. On strategy multiplayer games I usually don't notice high latency, low framerate or even input lag... but on twitch based games like shooters and ones like Rocket League, anything over 80 ping, under 60fps with any input lag whatsoever feels virtually unplayable and winds up being an exercise in frustration.
I didn't, but I also have really good internet. I've heard it works well on slower network connections, but I can't attest to anything but my own experience.
I'll be more focused on the future when US ISPs shape up their networks enough for 720p game streaming to be decent. That was always the main reason why I never bothered with PlayStation Now after my free time was up. Resolution doesn't mean anything if there's not enough bandwidth to make the pictures look reliably sharp.
I've ditched cable long ago. Are they still wasting their bandwidth on analog "basic cable" channels?
I have to agree here, cant't trust buying digital games to a known company who discards projects without notice (Still mad about Google Reader).
Like an idiot I didn't migrate or backup my feeds in time and still haven't fully recovered/reconstructed/found them all again. The loss of all my obscure webcomic feeds especially stings. :(
Google's announcement at GDC. Confidently expected to be related to video game streaming, after the beta of project stream last year. Could be a massive sea change in the gaming world if they manage to pull it off. Very curious to see what's announced and how soon it'll be available.
Speculative article from Kotaku is predicting all kinds of wild things, like Twitch & YouTube integration, OC games, play anywhere functionality... Huge bucket of salt required naturally, but it should be an interesting conference!
Also, bets that this will be US-only for at least a year?
Good lord I hope not. When publishers start releasing games exclusively on streaming services I will permanently become a "retro gamer".
I'll oppose that US-only bet. There are plenty of other countries with much better ultra high speed internet penetration, which can only help with performance.
Apparently Amazon is working on cloud streaming games as well, can't wait to see what it means for competition and mobile gaming capabilities with both of them going at it.
In the case of Nintendo i can see they will still follow their play offline together model with your friends and family.
I'm both glad Nintendo is the one to fill that niche and sad that it's considered a niche these days :(
I was a beta tester for this back when it was project stream. It actually works really well when your internet is good! The problem is that if your internet isn't consistently at least 50mbps, it will lag a lot and look really bad.
When it works though, it feels pretty seamless. The game I was playing, Assassins Creed, doesn't require super precise inputs, so it's hard to say how much input lag there really is.
Uh, wow, thats a tall order for anyone. It eliminates anyone using wifi right off the back and even if you have a wired connection, I dont think I've ever topped out above 20 or 30. And thats saying nothing about something like just having to pay for the amount of internet you actually use.
This (the isp problem) more than anything is why i think games as streaming is doomed to failure.
Yeah I had to use ethernet if I wanted a good experience. I'm fortunate enough to have google fiber so my internet is really good. I'm personally quite excited by this.
I recall them saying 25mbps was the minimum needed, 50 was just the plan I used to have when I tested it. And datacaps are also a big problem. I managed to used over 1TB by myself in a month when I was playing a lot.
Whoah, over 1TB from Project Stream usage alone? Can you estimate about how much playtime that was?
1TB was how much I personally used that month. So it's not just from project stream, though it primarily is.
I would estimate about 60 hours that month. Some quick back of the envelope math suggests that at 25 mbps (the minimum recommended) this would equate to .675TB, so the number feels reasonable to me.
It's interesting to me that 50Mbps strikes you as an extreme. Are you in a relatively sparsely populated area, by any chance?
Modern WiFi will comfortably do several hundred Mbps, so there's a good amount of headroom there, although a good old fashioned ethernet cable does still win on both speed and reliability if you're near the margins.
As for the ISP, it depends heavily on where you live - in the UK I get 80Mbps on FTTC as standard, and that's not especially unusual. There are relatively widely available 300Mbps cable plans, and a few FTTP networks rolling out gigabit connections (albeit in much more limited areas for now). Peak time contention might well still pose a problem (something I can attest to with my own game streaming usage), but the baseline capacity doesn't strike me as too onerous.
Christ, I've never even used internet that fast (you know, for an individual), let alone had it in my home. I currently get a really inconsistent 5 Mbps for where I am part of the year, and less than that where I am the rest of the time (which was a huge upgrade over what we had until very recently, the top of what the area offers!).
I was raving earlier today to my girlfriend about going to a public area around here during low activity hours and getting around 11 and downloading games in a few minutes. Stop making me feel like a hick! :P
So that's about what I get at home on wifi (5 mbps), but I just checked on my universitity's wifi and apparently here I get 190! Maybe I'm a hick too...
That's a lot more in line with what I'd expect - your high end connection is roughly double Stadia's upper bound and 4x its lower, which means that dropping back to a decent mid-range connection right now should still be well within those bounds (hopefully even around the upper end), and a few years of general progress should bring even basic connections into the fold, at least in terms of bandwidth.
How and if latency, stability and contention issues can be solved or worked around is likely the more complex problem, I suppose.
Right, my apologies, I just did a speedtest and my numbers got crossed somewhere. I've lived in a few different places, but I think I was misremembering my results.
I use Geforce Now to play Magic Arena almost every day. I've also previously tried parsec.tv (runs on Amazon), Shadow and PS Now as well as home streaming when I've owned a gaming PC.
In my opinion, the technology is still not there to compete with a proper gaming PC, however if targeted at the "casual" demographics it can be good enough. I wouldn't recommend streaming twitch shooters (Doom Eternal feels like a bad idea), but Hearhstone, Magic and Civilization or slow-paced single player stuff like the Last Of Us or Witcher are perfectly suitable.
Even though I'm still not convinced, I'm rooting for the solution. I believe that GPU-sharing would be very beneficial for PC gaming, since it lowers the entry cost to basically nothing, and once we get proper 5G network coverage, it could be very well extended to mobile clients. But right now connectivity is just not good enough and it can take years before we're there. Also knowing how Google treats services, i'm worried that Stadia might disappear after awhile.
Here's their website: https://stadia.dev/
They're using Debian on the backend.
This will be huge. Normies will be all over this. People were saying the same things about Netflix that everyone is saying about Stadia now. I'm looking forward to trying it out.
There's a big difference between Video Streaming and Game Streaming.
Video Streaming doesn't care much about latency, doesn't require as high of bandwidth, and you have a piece of content that can be reused for others.
Gaming streaming limits the amount of processing you can do to the outgoing frames, and the latency has a big influence on your immersion into the media. A site like Netflix can have lot of latency and work fine if you have have the bandwidth. If your network's latency increases due to external outages, you might find your Stadia with noticeable lag.
Am normie, can confirm. I want to be aaaaaaaaaaall over this.
So we're just using the word "normies", now? Like, really?
We don't know yet, but we should get more details in May.
I hope the controller that leaked isn't the real one as it doesn't look great.
It wasn't. It was a fan-made render based on the patent.
Can't even really be called a leak. It will look different.
Edit: The controller introduction.
I think it came out very similar.
Here are the renders based on the patent: https://www.yankodesign.com/2019/03/07/a-first-look-at-googles-project-stream-gaming-controller/
And the actual controller revealed today: https://www.polygon.com/2019/3/19/18272817/google-stadia-gdc-2019-controller
A couple of different buttons and slightly different arrangement, but quite similar (@R0b).
Not a huge fan of it, especially that d-pad but luckily you can use other controllers on most platforms so isn't really an issue.
Seems you can use Xbox One controller, that's great controller in my opinion.