Google jumps into gaming with Google Stadia streaming service
- Thread starter JournalBot
- Start date
One good thing about mothballing so many products: You can just lop off a piece of an old logo and voilà, instant branding...
(Google Wallet, circa 2015)
(Google Stadia, 2019)
(Google Wallet, circa 2015)
Upvote
41
(41
/
0)
If they want to really pull this off, they would probably need to install the servers in ISP's premises (to be as close to end users) and have some special low-latency agreement with them. And then ISP would get the cut directly from Google (or whoever tries to apply this model).
Upvote
0
(0
/
0)
You can't do h.265 and achieve latency requirements (or h.264 or V8 or v9). You have to generate raw frames (or something close to it) ready for local rendering.
Even compressing (and then decompressing) these raw frames is problematic for latency. Basically any kind of local processing is a latency killer. Even something as routine as double-buffering eats into your latency.
Decoders are total poison for latency.
Upvote
-1
(2
/
-3)
BR: teams. It doesn't need to be 1-player wins. Think Helms Deep. We could have war-games that involve any number of factions with more players in each than we've had in any single-instance MMO or BR to date.
MMO: You're thinking existing games with existing limitations. I'm not thinking another WoW. I'm thinking an actual living, thriving, world on a scale we've yet to see.
it's all scale, and from the sounds of it - this can scale beyond anything we've considered to this point.
Upvote
0
(2
/
-2)
It is Google so my bet is on VP9 (since it can do HDR 4k) and h.264 for compatibility.
Upvote
-2
(0
/
-2)
I don't believe that's true. I use h265 for in home streaming all the time with Moonlight and nvidia gamestream from my gaming PC and to my iphone, android tablet, and macbook pro. Hardware encoding and decompression makes for reasonably low lag.
Upvote
1
(2
/
-1)
Every single one of these services uses video encoding. Some of them used an internally developed codec but a lot of them right now are h.264. The encoding delay is literally one frame at most. Google is using custom hardware so they may be able to have the native output of the hardware be a encoded streams of data which would have the minimum possible latency.
Upvote
-1
(1
/
-2)
I love the weird optimism of people thinking that all of these games will be available under Vulcan and Linux in the greater PC landscape.
This thing is basically a console, but in the cloud (they said as much) - "We run Linux and it supports Vulcan" doesn't necessarily mean what you think it does. Does the Xbox running a version of DirectX mean anything about a PC port? (ask Square and others with their wonky PC support, where Direct X exists). I doubt the spec/API is literally Vulcan-as-it-exists-on-PC and a version of Linux that you run.
Maybe it happens, but it is not in Google's interest for their special sauce to benefit your local stuff.
This thing is basically a console, but in the cloud (they said as much) - "We run Linux and it supports Vulcan" doesn't necessarily mean what you think it does. Does the Xbox running a version of DirectX mean anything about a PC port? (ask Square and others with their wonky PC support, where Direct X exists). I doubt the spec/API is literally Vulcan-as-it-exists-on-PC and a version of Linux that you run.
Maybe it happens, but it is not in Google's interest for their special sauce to benefit your local stuff.
Upvote
5
(6
/
-1)
Upvote
-6
(1
/
-7)
I don't see any comments in the article regarding the availability of this service outside of the US? Games suffer much less from the "different publisher in different regions" syndrome than the films/TV but there is still a non-negligible quantity of smaller games where this is definitely an issue. So "this game is not available in your country" is definitely going to be a thing at least partially.
Personally, this isn't really a thing for me, to be honest, for multiple reasons which have already been mentioned: the main issue is of course the massive privacy issues around Google as a company, and then of course the fact that nobody knows if they'll shut this down within a couple of years or if by that point they'll have three different competing services, akin to their message apps situation right now.
Personally, this isn't really a thing for me, to be honest, for multiple reasons which have already been mentioned: the main issue is of course the massive privacy issues around Google as a company, and then of course the fact that nobody knows if they'll shut this down within a couple of years or if by that point they'll have three different competing services, akin to their message apps situation right now.
Upvote
2
(2
/
0)
I am interested in the economics of this. If a game is a hit, then publishers will LOVE this... Why just get a $60 per user onetime payday, when they will get paid every time someone plays their game. When I boot up Rage on my Xbox One using the 360 compatibility, id does not get any more money from me. However now, anyone that plays an older game, the publisher will get a piece of that action.
However there is a flip side, how many people buy games and never finish them. Each copy sold even if never played is money in the bank, and if a game does not have a dedicated fan base that will keep on playing it, the economics might not work. I would be interested to know how many times people buy games the day they are released and then never play them. I have done it a few times myself. So if I play say 5 hours of RDR2 on Stadia, that might not net Rockstar as much as if I bought it full price and then only played 5 hours of it.
Plus lets talk about bandwidth here, and lets forgo the rural areas in the US that are just screwed. I would wager that the average urban household can get a 75Mbit connection, and there are 4 people in the family. Dad is steaming Netflix at 4K, Mom watching a different show off of Amazon at 4K, say the kids are both playing on Stadia, all at the same time, is 75Mbit going to be enough anymore? What about data caps?
As a consumer I am really getting sick of "yet another subscription" Netflix/Amazon/Hulu/Spotify/etc... It's a death by a thousand cuts of your wallet. I would love to hear how much an average family pays in subscription services these days. Even if you cut the cord, you quickly rack up just as much in all of these other services. If I buy a PC/PS/Xbox, it's a sunk cost that offers me years of entertainment.
Finally, there is something visceral, for myself at least about physically having the hardware that could render what's on screen... be that a PC or Console...
However there is a flip side, how many people buy games and never finish them. Each copy sold even if never played is money in the bank, and if a game does not have a dedicated fan base that will keep on playing it, the economics might not work. I would be interested to know how many times people buy games the day they are released and then never play them. I have done it a few times myself. So if I play say 5 hours of RDR2 on Stadia, that might not net Rockstar as much as if I bought it full price and then only played 5 hours of it.
Plus lets talk about bandwidth here, and lets forgo the rural areas in the US that are just screwed. I would wager that the average urban household can get a 75Mbit connection, and there are 4 people in the family. Dad is steaming Netflix at 4K, Mom watching a different show off of Amazon at 4K, say the kids are both playing on Stadia, all at the same time, is 75Mbit going to be enough anymore? What about data caps?
As a consumer I am really getting sick of "yet another subscription" Netflix/Amazon/Hulu/Spotify/etc... It's a death by a thousand cuts of your wallet. I would love to hear how much an average family pays in subscription services these days. Even if you cut the cord, you quickly rack up just as much in all of these other services. If I buy a PC/PS/Xbox, it's a sunk cost that offers me years of entertainment.
Finally, there is something visceral, for myself at least about physically having the hardware that could render what's on screen... be that a PC or Console...
Upvote
9
(9
/
0)
It's not in their interest to screw developers with lock-in either. They explicitly mentioned using open source technology.
Upvote
-1
(0
/
-1)
Initial reaction: yeah that's cool.
Actual gameplay: *waiting* ... OH MY GOD WHY DID I CHOOSE ORC THERE'S LIKE 1,000 PEOPLE WAITING IN FRONT OF ME TO CLIMB THIS LADDER AND THEN I JUST GET ARROWED DEAD ANYWAY AND SET BACK 3,000 SPOTS!
Methinks helms deep wasn't there best example.
Upvote
2
(2
/
0)
Well, for example local button-to-pixel input delay while playing PUBG capped at 60 fps w/ V-Sync off is 30ms. It goes all the way up to 100ms if you turn V-Sync on. Note, that's the local delay, in the game client, before sending anything to the server.
Streaming a game from a server with 25ms ping you get about 90-100ms input delay. So streaming is about as bad as V-Sync.
Upvote
0
(0
/
0)
Like all of Googles dead services, depends entirely if they can get a user base. If not. It dies and the company focuses elsewhere. This is a GOOD thing though most think it is bad for some reason....
Upvote
-2
(0
/
-2)
Whereas I spent part of a year working with a company trying to do what Google Stadia is doing (working with, not for, as I was on the device software side)
Decoders kill latency.
Ideally, the server software wants to generate a raw frame and have the (low level) software on the local system render this frame directly from the network stream into the frame buffer or surface. That's not really possible, so the best we could do was to render from the network stream into a privileged buffer, and then copy into the frame buffer. Maximum of one copy, even though the guys working on the media pipeline said they needed a couple at a minimum (those were pretty "fun" meetings) because they didn't want to copy directly into their protected internal buffers from the network stream.
Upvote
0
(0
/
0)
I'm of two minds about this.
1. Its disappointing in that it will likely be too expensive for its own good, just like every other game streaming service.
2. Its exciting because its Linux and Vulkan based ... which could finally put some real support behind Linux as a native gaming platform.
1. Its disappointing in that it will likely be too expensive for its own good, just like every other game streaming service.
2. Its exciting because its Linux and Vulkan based ... which could finally put some real support behind Linux as a native gaming platform.
Upvote
1
(1
/
0)
Third-party games aren't going to be "trapped" anywhere. Nobody will develop exclusively for this.
Upvote
0
(2
/
-2)
Initial reaction: yeah that's cool.
Actual gameplay: *waiting* ... OH MY GOD WHY DID I CHOOSE ORC THERE'S LIKE 1,000 PEOPLE WAITING IN FRONT OF ME TO CLIMB THIS LADDER AND THEN I JUST GET ARROWED DEAD ANYWAY AND SET BACK 3,000 SPOTS!
Methinks helms deep wasn't there best example.
Okay, well, think Helms Deep, but with any scenario you can come up with on that kind of scale. ;-)
Upvote
0
(0
/
0)
I also wonder if they are using Wine. I'd totally expect them to, in order to boost their catalog beyond native titles.
Though I didn't see CodeWeavers in the list of partners.
Upvote
0
(0
/
0)
In two years Google will also offer Aidats, which will bring a new font and improved pull-down menus to the gaming experience, and which will deprecate Stadia. Users will be encouraged to move to the new platform. Then both will be killed one year later in favor of a new messaging app.
Upvote
2
(3
/
-1)
Initial reaction: yeah that's cool.
Actual gameplay: *waiting* ... OH MY GOD WHY DID I CHOOSE ORC THERE'S LIKE 1,000 PEOPLE WAITING IN FRONT OF ME TO CLIMB THIS LADDER AND THEN I JUST GET ARROWED DEAD ANYWAY AND SET BACK 3,000 SPOTS!
Methinks helms deep wasn't there best example.
Okay, well, think Helms Deep, but with any scenario you can come up with on that kind of scale. ;-)
Something like Total War skirmishes where every unit is actually controlled by a person (like 1000 on 1000 with cavalry, archers, ballista crew, infantry, etc) would be interesting. Although you'd need a way to keep dead people engaged either by a ratio of bots or something else.
Upvote
1
(1
/
0)
Upvote
0
(0
/
0)
As of right now Google appears to be pitching this to game devs as an actual platform the way console manufacturers do. Wine would be of little use for this there are actual developer toolkits that can ease Windows porting.
Upvote
0
(0
/
0)
Upvote
-2
(4
/
-6)
Upvote
4
(4
/
0)
Why? Actually "porting" a game using Wine can be quite trivial. Many run already out of the box through it on Linux (especially considering advancements of dxvk and the like), which makes such "porting" almost a zero effort (besides testing). Given Google have powerful backend, Wine can be more than adequate to run games that have no native Linux ports yet, at acceptable performance. It basically gives Google an easy option to expand their catalog with thousands of titles, until more native ones will arrive.
It can beat any existing console by a sheer number of compatible games, so I see no obvious reason for Google not to do it.
Upvote
1
(1
/
0)
I'm on board. Can't wait to try it out. Bring on the future.
Just need to know the pricing.
And everyone, plz stop. 1/3rd of the comments here are lol wHeN wiLL goOgLe KiLL iT?
Just need to know the pricing.
And everyone, plz stop. 1/3rd of the comments here are lol wHeN wiLL goOgLe KiLL iT?
Upvote
-2
(3
/
-5)
Upvote
0
(0
/
0)
Almost nobody ports games using Wine (I won't say nobody because I'm pretty sure a few Linux game releases actually literally have been the Windows version packaged with Wine). Wine isn't for porting - it is for end users to run existing software in Linux. These aren't the same thing. And Stadia isn't a Linux PC in the cloud from the user's standpoint - Google isn't interested in users running existing PC games in the cloud.
Upvote
-1
(1
/
-2)
Let's be realistic here (and speculate a bit
) - the datacenter HW is probably EPYC CPU (possibly with some custom flair on data encoder) + variation on Vega 56 (multi-GPU'd) that runs in some kind of virtual machine. This server needs to churn out full frame of graphics data to player each 16.67ms. In optimistic scenario, it could serve some 8 players, more likely 2-4. Therefore, in order to support 1000vs1000 battle, it would need one central server for running the game state, and some 500 servers to run the graphics. The only advantage to current setup (with everyone having the rendering machine at home) would be very low latency between the game server and rendering servers. But that probably would not be that much for people who are on quality broadband already (others cannot use game streaming anyway).
And still PC/console have latency advantage (even if identical rendering HW).
Upvote
2
(2
/
0)
That powerful backend is going to have to be updated constantly if they ever expect to deliver 4k:60fps. Its hard enough to do locally and the WINE overhead is certainly not free.
Upvote
0
(0
/
0)
Upvote
3
(3
/
0)
I'd modify that to lossy 4k HDR@60hz - at which point, 2k lossless likely looks a lot better.
Upvote
0
(0
/
0)
What "almost nobody does" is irrelevant. Google are making something new and can propose such option in the short term, especially when it is a lot cheaper than a proper native port.
Valve did just that with Proton (Wine's variant), providing a lot of games for Linux out of the box successfully. So you can't even claim almost nobody does it anymore.
Of course Google should pitch native releases to developers, but as solution for those who can't afford them or want to release already now - Wine is perfect.
Upvote
-2
(3
/
-5)
That's called 'The Firefly Effect', and is a growing problem for Google. https://tvtropes.org/pmwiki/pmwiki.php/ ... eflyEffect
Or:
https://allthetropes.fandom.com/wiki/The_Firefly_Effect
Upvote
1
(1
/
0)