• Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.

Phil Spencer is teasing the rumored Xbox streaming stick/console codenamed Keystone in a new shelf photo

pasterpl

Member
Have you ever tried one of these "game streaming services" over wifi before?

"Acceptable lag" would be fine, the problem is "acceptable lag" doesn't exist over wifi if you are streaming games, if the console or device isnt connected via ethernet it simply does not work properly

And don't even get me started trying to stream games over a hotels wifi if you travel, you may as well just buy a series s and take the console with you
Actually yes, couple at home and couple when travelling (hotels, airb&b) and most games were beyond playable with long parts with almost no noticeable cars lag and not many iQ issues. Next week I am travelling to California/Palm Springs so will test it again I guess.
 

Three

Member
So let's just add on top of it where people get elite controllers so they can have every single edge they can get in these twitch games.

I know some's favorite brand is jumping into streaming more and more now, but let's not kid ourselves.
Forget back buttons and all that jazz even, just look at all the people who often call 30fps completely unplayable in other threads because of the input lag. They would kill eachother over that stuff. 30fps CoD would be a bloodbath thread. Could you imagine?

Yet here we all are happy about a whopping 175ms+ lag with jitter while streaming. That's acceptable for CoD somehow. At least with 30fps you are getting twice the computing power or higher fidelity. Not trying to go off topic, but it just goes to show.
 

THE DUCK

voted poster of the decade by bots
Actually yes, couple at home and couple when travelling (hotels, airb&b) and most games were beyond playable with long parts with almost no noticeable cars lag and not many iQ issues. Next week I am travelling to California/Palm Springs so will test it again I guess.

As ideal as it seems to use cloud gaming on the road, cloud streaming and hotel wifi is probably the worst combination possible. It's like taking a new mustang out for a drive and filling the gas tank with water and being surprised the car won't drive. Or buying fresh five guys burger, leaving it in the fridge for a week, then reheating it in the microwave, trying it, and saying five guys sucks.
The fact is hotel wifi is universally horrible and is no way to test anything.
 

Menzies

Banned
Forget back buttons and all that jazz even, just look at all the people who often call 30fps completely unplayable in other threads because of the input lag. They would kill eachother over that stuff. 30fps CoD would be a bloodbath thread. Could you imagine?

Yet here we all are happy about a whopping 175ms+ lag with jitter while streaming. That's acceptable for CoD somehow. At least with 30fps you are getting twice the computing power or higher fidelity. Not trying to go off topic, but it just goes to show.
Yes, here we ‘all’ are happy about “175ms+ lag with jitter”. As if to say the theoretical best performance can’t top this whilst living next door to an Azure data centre with gigabit fibre and latest direct capture API?

When was the time and date of your recorded “175ms”?
What is your location and ping to the data centre?
 

Three

Member
Yes, here we ‘all’ are happy about “175ms+ lag with jitter”. As if to say the theoretical best performance can’t top this whilst living next door to an Azure data centre with gigabit fibre and latest direct capture API?

When was the time and date of your recorded “175ms”?
What is your location and ping to the data centre?
Way to miss the point. I'm even being generous with the 175ms. A lot of games can go into the 300ms. Sure, if you live "next door to an an Azure data centre with gigabit ethernet" your milage may vary. Gigabit ethernet wouldn't do anything for your latency anyway only your image quality. Jitter would still exist too meaning input latency can vary while playing, you get packet loss.

The point being that latency is way higher and more unpredictable while streaming. Don't get hung up on the average figure I've provided by saying it's not the "theoretical best performance", sure it isn't but it's not unrealistic.
 
Last edited:

CeeJay

Member
the problem for cloud streaming is it'll always be an after thought. its like the worst way to play games, so why would anyone choose it as their primary way to play? microsoft will give it their best shot and may even have the best streaming quality, but reality has shown us already that their isnt really a big market for people wanting to play games like this. it will be just like ps now streaming. an option thats there but not used by many.
Historically games streaming services have not done great but isn't that as much to do with the model as it is the tech?

Xcloud is a feature of a service, it's not a service itself.

Most streaming services in the past have been something where you pay specifically for the cloud service. Xcloud is part of a much wider holistic approach. You are subscribing to GamePass as a whole which has many different concurrent methods of consuming the content. In many ways Xcloud cannot fail as long as GamePass is successful. You can attach a percentage to it to convey how popular the method it is for accessing the content but whether that is a high or low number doesn't in any way show how successful GamePass is.
 

Menzies

Banned
Way to miss the point. I'm even being generous with the 175ms. A lot of games can go into the 300ms. Sure, if you live "next door to an an Azure data centre with gigabit ethernet" your milage may vary. Gigabit ethernet wouldn't do anything for your latency anyway only your image quality. Jitter would still exist too meaning input latency can vary, you get packet loss.

The point being that latency is way higher and more unpredictable while streaming. Don't get hung up on the average figure I've provided.
Where’s your “average figure” from exactly? A 12-month old video from Digital Foundry tested from 1 location? Have you personally tried it?
 

Three

Member
Where’s your “average figure” from exactly? A 12-month old video from Digital Foundry tested from 1 location? Have you personally tried it?
I have tried several streaming platforms including GFN, xCloud and PS Now personally and the lag is noticeable, even local streaming it's noticeable but if you are asking if I've set up an experiment of my own then no. I haven't personally set up a latency test measurement with circuits and a high speed camera. I know others who have though like DF, Nvidia and reddit posters. 175ms is not some unusually high stated value from these sources. It's average, as I said it can be as high as 300ms depending on the game.
 
Last edited:

CeeJay

Member
Way to miss the point. I'm even being generous with the 175ms. A lot of games can go into the 300ms. Sure, if you live "next door to an an Azure data centre with gigabit ethernet" your milage may vary. Gigabit ethernet wouldn't do anything for your latency anyway only your image quality. Jitter would still exist too meaning input latency can vary while playing, you get packet loss.

The point being that latency is way higher and more unpredictable while streaming. Don't get hung up on the average figure I've provided by saying it's not the "theoretical best performance", sure it isn't but it's not unrealistic.
https://www.azurespeed.com/Azure/Latency

Currently I am getting 35ms to UK South which is around 200 miles away. Thats on Wifi over a 70Mb/s internet connection.

Sure, mileage will vary but you are sure trying to convey that unless you live next door and have massive internet you are not going to have a good experience. I would say that (for the country i live in) i'm an average user (Wifi using the free router i got with my internet package) living in an average city with an average internet connection that is quite a long way geographically from the data centre yet I am still getting a perfectly acceptable Xcloud experience.
 

Three

Member
https://www.azurespeed.com/Azure/Latency

Currently I am getting 35ms to UK South which is around 200 miles away. Thats on Wifi over a 70Mb/s internet connection.

Sure, mileage will vary but you are sure trying to convey that unless you live next door and have massive internet you are not going to have a good experience. I would say that (for the country i live in) i'm an average user (Wifi using the free router i got with my internet package) living in an average city with an average internet connection that is quite a long way geographically from the data centre yet I am still getting a perfectly acceptable Xcloud experience.
And how much jitter were you getting with that average ping alone? look at your graph. Azurespeed doesn't show you latency with xcloud tech. It's just pinging the servers. Measurements in games where you look at input on the controller vs movement on screen gives you the input latency for xcloud.
 

CeeJay

Member
And how much jitter were you getting with that average ping alone? look at your graph. Azurespeed doesn't show you latency with xcloud tech. It's just pinging the servers. Measurements in games where you look at input on the controller vs movement on screen gives you the input latency for xcloud.
Maybe 15ms jitter. Peaks go upto about 50ms and down as low as 30ms.
 

Three

Member
Maybe 15ms jitter. Peaks go upto about 50ms and down as low as 30ms.
I'm in London and I get 80ms average, 100ms jitter on the UKs fastest broadband but again azurespeed.com isn't xCloud this is just pinging Azure data centres and getting a rtt. To measure input latency for xcloud you need input to pixel latency tests with a high speed cam. You need to measure the time it takes between a button press and the corresponding movement on your display. It isn't a simple matter of measuring ping. My 175ms isn't some preposterous number, it's relatively normal:




Suggesting it's 32ms based on your azurespeed.com ping times is though.
 
Last edited:

CeeJay

Member
I'm in London and I get 80ms average, 100ms jitter on the UKs fastest broadband but again azurespeed.com isn't xCloud this is just pinging Azure data centres and getting a rtt. To measure input latency for xcloud you need input to pixel latency tests with a high speed cam. You need to measure the time it takes between a button press and the corresponding movement on your display. It isn't a simple matter of measuring ping. My 175ms isn't some preposterous number, it's relatively normal:




Suggesting it's 32ms based on your azurespeed.com ping times is though.

I didn't try to suggest anything of the sort I just gave my general circumstances including a ping response time to Azure, I don't think there was even a hint that I was implying my total latency was 35ms. Why would I need to work out my total latency when my eyes and brain can tell instantly whether the latency i'm playing at is acceptable and that is way more important than trying to put a number on it?

It's strange how I live 200 miles away from London (where UK South is located) yet get less than half the ping that you do living close whilst also claiming to be on the UKs fastest broadband. Maybe you need to get your service checked, 80avg with 100ms jitters sounds borderline broken.
 

Three

Member
I didn't try to suggest anything of the sort I just gave my general circumstances including a ping response time to Azure, I don't think there was even a hint that I was implying my total latency was 35ms. Why would I need to work out my total latency when my eyes and brain can tell instantly whether the latency i'm playing at is acceptable and that is way more important than trying to put a number on it?

It's strange how I live 200 miles away from London (where UK South is located) yet get less than half the ping that you do living close whilst also claiming to be on the UKs fastest broadband. Maybe you need to get your service checked, 80avg with 100ms jitters sounds borderline broken.
So what was your intention of posting your 32ms ping time to Azure data centres if not to dispute the measured 175ms input latency for xcloud? What were you trying to show with it?

It's Virgin Media's gigabit broadband. What's yours? though I hear VM don't have the best latency and got in trouble with ofcom for saying best broadband for gaming or something. My service is fine though. Much better than your average provider in the UK in terms of speed.
 

CeeJay

Member
So what was your intention of posting your 32ms ping time to Azure data centres if not to dispute the measured 175ms input latency for xcloud? What were you trying to show with it?

It's Virgin Media's gigabit broadband. What's yours? though I hear VM don't have the best latency and got in trouble with ofcom for saying best broadband for gaming or something. My service is fine though. Much better than your average provider in the UK in terms of speed.
If you read my post you would see that I was merely stating that even with an average setup I can get a reasonable experience, if you read more into it than that then that's on you. I'm not getting into a pseudo-science war with you about total latency values, like I said, it's just a number that I don't really care about. Your eyes and brain are more important to determine whether streaming is acceptable in your own chosen location on your own chosen device not some random quoted number from some random video on Youtube.

By the way I Googled What is a good ping UK?

Result...

Perfectly average ping: 20ms to 50ms (0.02 to 0.05 seconds)

If you are getting 80avg with 100ms jitters then there is something wrong with your internet connection especially when you are in the capital.
 

Three

Member
If you read my post you would see that I was merely stating that even with an average setup I can get a reasonable experience, if you read more into it than that then that's on you. I'm not getting into a pseudo-science war with you about total latency values, like I said, it's just a number that I don't really care about. Your eyes and brain are more important to determine whether streaming is acceptable in your own chosen location on your own chosen device not some random quoted number from some random video on Youtube.

By the way I Googled What is a good ping UK?

Result...



If you are getting 80avg with 100ms jitters then there is something wrong with your internet connection especially when you are in the capital.
Scientific evidence is more important than some random guys "feelings" too. I'd put more weight on the former personally. The lag is there both in anecdotal experience and scientific measurements.

You didn't say, who's your provider?
 
Last edited:

CeeJay

Member
Scientific evidence is more important than some random guys "feelings" too. I'd put more weight on the former personally. The lag is there both in anecdotal experience and scientific measurements.

You didn't say, who's your provider?
Why i put mileage will vary on my first post, cloud streaming is extremely subjective. I know i'm some random guy (any of us guys can say that about ourselves or others). But, for this random guy it works fine despite the secondhand science you are quoting (That incidentally was also carried out by some random guy in a random location on a random day using random hardware). If they carried out the same tests from a different location on different hardware on a different day they would have likely got different results. Just because they carried out some quite technical tests and got accurate results from them doesn't mean to say that I am (Or any other random person) governed by them.

You are saying that I personally should put more weight in the science you are peddling even if my own personal experience flies in the face of it?

Surely this is the point. Our own personal experience is way more important than the numbers from some isolated test. Hence why I posted my experience to counter your "science says no bruh" mentality.

My provider is also irrelevant as well. Currently EE although my connection was no different when I had Plusnet or BT, it's all using the same copper to the same exchange.
 

DaGwaphics

Member
Obviously, GFN can have a latency edge on Xcloud just because you can run at a higher FPS. That's why I say, for best results MS probably should have gone with 3 consoles. Even if the 3rd wasn't available to consumers and was just for Xcloud. Create a box that can run the same visual settings as the X model at twice the FPS and use it in the cloud. Would really reduce the latency of the streaming experience and make it more comparable to the console experience.

But, with that said, Xcloud works great for me as it is in most instances.
 
Last edited:

Three

Member
Why i put mileage will vary on my first post, cloud streaming is extremely subjective. I know i'm some random guy (any of us guys can say that about ourselves or others). But, for this random guy it works fine despite the secondhand science you are quoting (That incidentally was also carried out by some random guy in a random location on a random day using random hardware). If they carried out the same tests from a different location on different hardware on a different day they would have likely got different results. Just because they carried out some quite technical tests and got accurate results from them doesn't mean to say that I am (Or any other random person) governed by them.

You are saying that I personally should put more weight in the science you are peddling even if my own personal experience flies in the face of it?

Surely this is the point. Our own personal experience is way more important than the numbers from some isolated test. Hence why I posted my experience to counter your "science says no bruh" mentality.

My provider is also irrelevant as well. Currently EE although my connection was no different when I had Plusnet or BT, it's all using the same copper to the same exchange.
As do people who stand by 30fps being fine and that is subjective too. The point wasn't "science says no bruh" it was to show that you get 175ms or more latency with jitter which is worse and some of the same people being cool with it. I don't know what your actual latency is that you are OK with and tolerate, and it's fine to be OK with it, but it certainly isn't the 32ms ping to the Azure servers you posted. I'm not here trying to tell you not to like things.
 

Swift_Star

Banned
Again, Playstation needs to be ready, if this thing can play Call of Duty and Madden with acceptable lag, its gonna sell like cookies.

The entry point for consoles is too extreme, as evident by the Series S success. People want cheaper stuff like this.
Lol acceptable lag. This thing will flop. A séries S is better value.
And if they really wanted cheaper stuff like this, series s would be selling like hotcakes and it isn’t.
 
Last edited:

Bragr

Banned
Lol acceptable lag. This thing will flop. A séries S is better value.
And if they really wanted cheaper stuff like this, series s would be selling like hotcakes and it isn’t.
What are you talking about? The Series S is a success, and casuals don't give a shit about minor cloud lag, something that could stream Minecraft or Stardew Valley on the TV for a small sum has clear potential.
 

Kacho

Member
The timing of this device is absolutely hilarious. Phil said how energized he was about competition in the streaming space a couple years back and Stadia is already dead in the water. The other streaming services exist with basically no fanfare. They could give this thing away in a Doritos bag and it still wouldn't take off IMO. How it performs will be fun to watch.
 
Top Bottom