-
Ever wanted an RSS feed of all your favorite gaming news sites? Go check out our new Gaming Headlines feed! Read more about it here.
-
We have made minor adjustments to how the search bar works on ResetEra. You can read about the changes here.
Ars Technica - Halo 5 on XCloud tested, 67ms of latency (Halo 5 locally - 63ms)
- Thread starter JackEtc
- Start date
First you claim he's a fanboy, now you're quoting Jez in order to discount the XCloud service. You change your alliances pretty conveniently huh?
1. Don't buy into the hype. We need more testing by hundreds of real users, when open beta comes, before having useful, real-world info.
2. Azure > Google Cloud so I wouldn't be surprised if Microsoft beats google at this.
3. Amazon could've been in the picture. Their cloud is still the biggest, followed by Azure.
2. Azure > Google Cloud so I wouldn't be surprised if Microsoft beats google at this.
3. Amazon could've been in the picture. Their cloud is still the biggest, followed by Azure.
considering how shitty many console games feel in the input lag department, 63ms is pretty damn good for a network based gaming service
Could you do the same test with stadia next time you go hands on with it?
This needs to be thread marked.
We have someone who was their doing the test no need to speculate for them.
We can just ask them
That's like saying the PS5 and next gen Xbox are DOA in America since a whole lot of American households can't afford to pay $399+ for a game console.
Yeah, if the lag is steady and predictable, you can adjust. If it's all over the place, then it gets tricky. Space Channel 5 on DC was fucked for beat timing, but it could still be completed.
Even assuming they do, which they wont, internet infrastructure in america is trash. And that applies to landline internet and cellular services. Its slow and unreliable for over 90% of the country. I think spectrum currently has more customers than any other company in america and it probably spends almost as much time down as it does up. And the vast majority of us stuck with it have no better option
Correct, all of the latencies add up. That is why increasing the framerate of the game running in the cloud, can compensate for the network latency.
Last edited:
You have no idea what you are talking about. Perfect conditions my ass. Then again looking at your post history with anything MS related it makes sense.
You would think that is obvious to anyone that says that cloud services will fail in the US because not everyone has access to the service.
Based on Digital Foundry data.
the fact that halo 5 over xCloud would have less input lag than titanfall 2 local is... remarkable
Based on Digital Foundry data.
Man, didn't realize Doom and Killzone were THAT bad. I remember complaints with Killzone, but Doom is rough too.
That's the thing, people need to look at the actual numbers, not all games have the same input lag. What we should start to see in the future are games that are optimized for lower latency. See Call of Duty Infinity warfare for example. A game like Killzone Shadowfall would actually have lower latency than the local version if it would be optimized to run with a 39.3 ms latency like call of duty.
It sounds really ace, but I need to try it myself before I'm a true believer. It'll be crazy if they get it right though.
Lol people want this to fail so bad. It's kinda sad to read stuff like this
The local version would still be lower if they could optimize it to CoD levels.
You may be confusing a stage show with the show floor. The show floor tends to be a nightmare for wireless networking. They try to keep things as simple/smooth as possible, but the density of people and devices pulling from the same pool of networks isn't ideal.
The US is "only" ~2500 miles wide, and the current density of data centers puts most potential users within a 400 mile (or less) range, because of how population centers work. It's more practical than you might think, even for a geographically challenging territory like NA. Microsoft having based their service business off Azure infrastructure for so long puts them in a good place to deliver on this in the current landscape, and a lot of the local infrastructure development pushed by companies like Amazon for their own purposes have made a big difference.
Last edited:
Lmao at the amount of people drinking the kool aid.
I'm ready to eat crow whenever you want (spoilers: i won't) but all this downplaying of controlled environment is remarkable.
I'm ready to eat crow whenever you want (spoilers: i won't) but all this downplaying of controlled environment is remarkable.
You can see examples of games running at 120fps like Fortnite and or PUPG on Geforce Now to reduce latency, that is not possible on consoles. Even if it were the case that the same optimization could be applied locally, at that point we would be talking about a game with a more than acceptable latency for millions. You can imagine the number of people that say that they would never play a cloud game because of the latency and right after that they go and play/enjoy an online match of Killzone Shadowfall. That is the main point of my example.
Last edited:
How every thousand miles is 7.5ms, the latency is well within reason considering that or other barriers to network latency.
I meant the whole service is going to feel better than RDR2. Not any particular game.
Though I shudder to think what RDR2 will be like on it.
This sounds incredible! Wonder where Jez picked up on that lag from. I'm sure that while there will be user experience in a beta that shows a degree of variability, the fact that they could also offer documented experiences like what Ars reports - has to mean that they've figured some critical parts of the streaming landscape out.
I got to try this out last night with Halo and was very impressed. I mean i was also thinking well this was a controlled environment so it's not like they are going to show it off broken lol. I did overhear that the same tech used by their azure servers is the same that will run on our consoles to stream. So I imagine technically you can set up your own controlled environment and make it as fast as you want with your own internet. I know this is old news with PS4 and others but if there really is some secret sauce behind this that's real it could be the leap it needs.
I can't say I would consider an E3 showfloor really a controlled environment. A controlled environment is in a conference room/lab in Redmond, Washington. At E3 they're subject to network traffic that they can't account for along with having a pretty high amount of devices connecting through the same network.
Absolutely what was shown off should not be brushed off by anyone just because it was at e3, but I also can't blame anyone for being skeptical until they see it work on their own network and device. The big impression here is that when I played Halo I didn't get the feeling I was playing at a disadvantage and that's an impression to build upon for me.
This. Worked at many a conferences and for every one of them connection speed is always an issue.
Absolutely. The October launch of the public preview of both the in home console streaming and xCloud should really be insightful to see how various network setups handle it.
I have like 100 down at home on WiFi. I really don't care what the connection is like when there's like 20 journalists connected to this on the floor. People keep saying it's not normal conditions and I'm like....yeah no shit if anything it should work way better in the comfort of your own home.
Tested this yesterday playing halo 5 and forza. It felt amazing and didn't really notice any latency problems. Really excited for more testing in October and curious about the pricing details. Its really good
What America do you live in? Because there are already cloud servers all up and down both coasts, where most people live, in the U.S. Thinking this is DOA is some wishful thinking, this stuff is going to be pretty popular next gen and perfectly accessible to millions and millions of people.
Can you provide his original video file? Because There is approximately zero chance the game was actually streamed from SF while only adding 4ms. xCloud doesn't use any code changes. That 4ms would have to include input travelling 400m to the server, frame rendering on a One S, video compression, a 400m trip back to LA, and video decompression.
It is far, far more likely the video was actually 120fps, and he did the math assuming 240fps. That or some other error in deteermining which button press cause which action.