Microsoft Project xCloud (Game Streaming), now offering Fortnite free without GPU membership

The free remote play snippet was interesting. I wonder how the performance will compare to xCloud proper.

Edit: I'm sure they implied it was also running via Azure. It made me think that they're already broadcasting / recording to Azure from the console and they're probably just building on that infrastructure.

No but it would not surprise me if the console did register itself into the Xcloud or Live, that way you can connect via the central server then that drops off giving you peer to peer access. This would help with firewall issues and is how many remote pc support applications work.
 
No but it would not surprise me if the console did register itself into the Xcloud or Live, that way you can connect via the central server then that drops off giving you peer to peer access. This would help with firewall issues and is how many remote pc support applications work.

That's how a lot of things in azure work. It was shown that way at the 2008 Microsoft PDC (Professional Developers Conference) in Los Angeles. The initiall discovery is done via known connectable endpoints in Azure and then the two clients attempt to connect directly to one another. They only keep the connection to Azure in the event both sides are under extremely strict connection rules.
 
It was to be expected even in a controlled environment. Similar to what people have experienced with Stadia and PSNow.

The lag was noticeable, but totally and completely playable, and more than adequate for an offline shooter. There were occasional instances of artifacting here and there, and you probably won't want to get competitive on it, but the audio delivery, the responsiveness of the controls, and the visuals were all incredibly impressive, vastly exceeding what I would ever have expected

https://www.windowscentral.com/i-tried-project-xcloud-xboxs-e3-2019-show-it-real-it-insane
 
ArsTechnica has some latency numbers from their quick & dirty test...

We didn't have a high-end, custom-built latency testing rig to measure things precisely. But we did have an iPhone with a slow-motion camera to do some quick visual testing.

In our video tests, the time between tapping the A button and seeing a response on the smartphone screen took 16 frames of a 240fps video (or 67 milliseconds) across three subsequent tests. That's almost imperceptibly slower than the 63ms (milliseconds) input latency Digital Foundry measured on the Xbox One version of Halo 5 in 2017 tests.

Testing latency of a wired Google Stadia demonstration at March's Game Developers Conference, Digital Foundry found total latency of 166ms, compared to a low of 100ms on a 60fps PC.

https://arstechnica.com/gaming/2019...o-5-nearly-indistinguishable-from-local-play/

Tommy McClain
 
Yeah. I also had a perfect experience with psnow. But I'm not a dumbass thinking it's representative of the logical limitations. I live in a large city with under 10ms to tier 1 from most cable and dsl providers. Psnow is perfect here but it can't be that good where there is no low latency paths to edge nodes, i.e. most rural areas. Providers congestion algorithms are often shit and buffer bloat in most places, and there is nothing to be done at the server side.

Google claim 7500 edge nodes for sativa and it probably reach only the big cities. MS is probably competitive. Sony uses amazon apparently, not sure what their plans are for edge nodes.
 
Last edited:
Google claim 7500 edge nodes for sativa and it probably reach only the big cities. MS is probably competitive. Sony uses amazon apparently, not sure what their plans are for edge nodes.
Sonys plan is to use MS azure/xcloud
 
Sonys plan is to use MS azure/xcloud
Yes, but is it just to expand edge nodes and use all providers in parallel? They said they don't even have a partnership yet, just exploring possibilities. MS definitely have similar needs, so sharing edge nodes would be profitable for both. Sony could be their first non-internal client for that sort of deployment.
 
You think an 800 mile round trip, video compression and decompression could all happen in 4ms?

My point was you didn't explain why you thought that. Second, there's more than a zero chance it could be correct. If everything was equal then yeah you might have a point, but we don't know what's changed to Halo 5 since Digital Foundry did their latency tests in 2017. We also don't know what changes Microsoft have made on the xCloud side. We've heard reports of about 10ms latency added for xCloud. So if they improved the native latency of the game by just 6ms, then it would still be in the realm of plausibility. But what do I know I'm just an average Joe.

Tommy McClain
 
My point was you didn't explain why you thought that.

I thought it was self evident that an addition of only 4ms was not physically possible. Especially not after all the reports from people who said the lag was noticeable in the demo. An extra 4ms is not a noticeable difference.
 
Guys its not an apple to apple comparison they are testing now vs some test in 2017
how much has the codebase changed in 2 years?
 
Using a different monitor is enough to offset these lag tests.

The monitor and times used by DF are that configuration.

The monitor and times used for XCloud was a different one.

Apples 2 apples requires the same recording and display equipment.

Nothing wrong with the report, just keep those considerations in mind when doing cross compare.
 
Yeah. I also had a perfect experience with psnow. But I'm not a dumbass thinking it's representative of the logical limitations. I live in a large city with under 10ms to tier 1 from most cable and dsl providers. Psnow is perfect here but it can't be that good where there is no low latency paths to edge nodes, i.e. most rural areas. Providers congestion algorithms are often shit and buffer bloat in most places, and there is nothing to be done at the server side.

Google claim 7500 edge nodes for sativa and it probably reach only the big cities. MS is probably competitive. Sony uses amazon apparently, not sure what their plans are for edge nodes.
what does sativa means? Dont you mean stevia?
 
Back
Top