Stadia, Google Game Streaming platform [2019-2021]

Could be tweeking the encoding algorithms etc. Might give a worse IQ until its done.
Probably lots to do before it's release still.

I'm just assuming everything isn't ready yet, because for them 1080p60 should be the minimum. Unless the game was hardcoded for 30fps
 
Yeah, Linux and Vulkan are really bad choices, maybe not Vulkan, but lower level API in general demand much more involvement from the developer. Windows and DX11 should result in a faster porting process.

And much higher run time cost due to the cost of the Windows license and more system administration needed. Also, Google will have the capabilities to optimize the whole stack (everything from the drivers to the deployment of updates) as well.
 
Game development is a lot different now than in the past.
Having tools and engines notably (unreal & unity) on your platform goes a long way.
I'm not saying work still isn't needed, but the amount of work needed is drastically cut.

As for bespoke engines, most are developed with multiplatform in mind. So even those aren't as bad as they used to be, especially given the underlying hardware isn't lightyears different. Especially if they target the low level api's.

I'm not saying it's a 2 man day job. The ubisoft about porting AC should give a good indication.
 
The ubisoft about porting AC should give a good indication...
...of how much Google may be willing to pay. No amount of launch titles is indicative of anything, because they are often paid for. It's after the honeymoon period, when devs are choosing to take on the cots of porting themselves, that you see if a platform is viable or not.
 
...of how much Google may be willing to pay. No amount of launch titles is indicative of anything, because they are often paid for. It's after the honeymoon period, when devs are choosing to take on the cots of porting themselves, that you see if a platform is viable or not.
Their presentation would give us more information than we have now.
May even tell us what tools they currently have etc
Although I'm pretty sure a few devs could tell us how hard it is to port from dx12 to Vulkan for example.
The fact that the hardware is so much more powerful than console will mean you could get away with a lot less optimization for this gen of games.

Imo porting isn't the biggest issue they have to the platform being viable.
 
It's not like their teams don't do anything but DirectX, they do have PS4 versions after all so I'm sure their engine is designed to be fairly portable.
 
Ars story on recent Google history of shutting down or abandoning features, https://arstechnica.com/gadgets/2019/04/googles-constant-product-shutdowns-are-damaging-its-brand/ . Those who wonder just how long will Google support Stadia game streaming have legit concerns.

... snippet intro below ...

It's only April, and 2019 has already been an absolutely brutal year for Google's product portfolio. The Chromecast Audio was discontinued January 11. YouTube annotations were removed and deleted January 15. Google Fiber packed up and left a Fiber city on February 8. Android Things dropped IoT support on February 13. Google's laptop and tablet division was reportedly slashed on March 12. Google Allo shut down on March 13. The "Spotlight Stories" VR studio closed its doors on March 14. The goo.gl URL shortener was cut off from new users on March 30. Gmail's IFTTT support stopped working March 31.

And today, April 2, we're having a Google Funeral double-header: both Google+ (for consumers) andGoogle Inbox are being laid to rest. Later this year, Google Hangouts "Classic" will start to wind down, and somehow also scheduled for 2019 is Google Music's "migration" to YouTube Music, with the Google service being put on death row sometime afterward.

We are 91 days into the year, and so far, Google is racking up an unprecedented body count. If we just take the official shutdown dates that have already occurred in 2019, a Google-branded product, feature, or service has died, on average, about every nine days.
 
...of how much Google may be willing to pay. No amount of launch titles is indicative of anything, because they are often paid for. It's after the honeymoon period, when devs are choosing to take on the cots of porting themselves, that you see if a platform is viable or not.
Just realised I posted about this in the general streaming thread, so going to link the post with the vid here also.

https://forum.beyond3d.com/posts/2063715/
 
someone named Andrew McGregor, ex-google employee, wrote a very good response on the issue of lag in this quora thread.

https://www.quora.com/What-is-the-f...reach-until-its-limited-by-the-speed-of-light

Yup I had this discussion here somewhere can't remember which thread or when exactly. But latency is all about how close you are to the data centers, for something like Stadia and for a game like Doom you would want to be within about 500km from the data center.
 
Even if they get the min latency to some "acceptable" level it will be irregular which I imagine would be even far more annoying for games which can live with a higher latency.
 
If nothing else, having google work on a game streaming themselves is helping bring to light the technical necessities of this type of workload. Nice.

Wi-Fi performance modes

In Android Q apps can now request adaptive Wi-Fi by enabling high performance and low latency modes. These will be of great benefit where low latency is important to the user experience, such as real-time gaming, active voice calls, and similar use-cases. The platform works with the device firmware to meet the requirement with the lowest power consumption. To use the new performance modes, call

https://android-developers.googleblog.com/2019/05/whats-new-in-android-q-beta-3-more.html?m=1
 
We know that stream games often uses a lot of resource to eliminate input latency. For example, 30 fps games on client side may run at 60 fps on server side.If this is also the case for stadia (AC odyssey only have 1080p 30fps), that means stadia may use 2x resource of a local PC to run a 30fps game. In other words, a local PC only needs 50% resource of stadia, that is, 5.35 Tflops and 1.35 GHz CPU, to have the same performance as stadia. If we consider the additional resource on server side, statdia may not deliver next-gen games yet. Stadia can only have games similar to running on PS4 pro or xbox one x.
 
I think the game changer won't necessarily be games, but rather having the infrastructure to offload any traditional computing environment to the cloud. A dirt cheap watch, smartphone, tablet or TV that's a blend of local GUI and remote terminal session that leverages the resources of your home desktop/server or nearest datacenter. Purpose built silicon for handling a terminal session like that would probably give orders of magnitude more battery life than needing a beefy high clock SoC, and effective performance scaling would be decoupled entirely from the form factor.

We're kind of already there in the case of certain isolated web apps and services, but if it were at the level of the operating system and seamless to the user, it seems like the most obvious way forward in a post-Moore's Law world where new devices are both increasingly costly to manufacture and increasingly difficult to justify the cost for marginal differences. That sort of corporate controlled fully-vertical ubiquitous computing environment sounds extremely dystopian, but the benefits would be so significant that you'd have little choice but to join the collective.
 
Or news that xCloud has higher hardware specifications ?
 
Back
Top