PC Gamer test drives Google's Stadia

Blade-Runner

Supreme [H]ardness
Joined
Feb 25, 2013
Messages
4,330
https://www.pcgamer.com/au/i-tried-googles-stadia-and-latency-got-me-killed/

So how does Stadia fare? Both games are playable, and if that's all you're after, Stadia will suffice. But latency is clearly present. During an intense fight in Doom, moving and aiming and shooting with the mouse and keyboard just feels sluggish, especially compared to playing on a high-end PC.

Playing the first level of Doom, for example—the training level—I died five times. Five! I never die on the first level these days when playing at home, at least on the default difficulty. The problem is that the latency, or simulated latency in this case (more on that in a second), caused me to repeatedly aim past possessed engineers, soldiers, and imps. Aim, click, boom! I missed. Then I'd overcorrect. Click-boom-miss again. That's a great way to waste shotgun ammo.

As expected, the latency makes it crap for any twitch based game which requires decent reflexes.
 
Google can't fix the laws of physics. There's a reason why Stadia was demoed with a gamepad and not a keyboard and mouse cause you'll feel the latency with something precise like a mouse, where a gamepad was never precise to begin with. Thankfully the people at Level1Techs agrees that Cloud Gaming is dumb cause so far a lot of big YouTubers seem to back Stadia including AdoredTV.

 
Until someone magically solves the internet latency problems, streaming will be a sub-par experience. It may look and test great under pristine test situations... but the internet used by the masses is anything but pristine.

I think it is solvable, but it will require some tech secret sauce that is currently missing.
 
Until someone magically solves the internet latency problems, streaming will be a sub-par experience. It may look and test great under pristine test situations... but the internet used by the masses is anything but pristine.

I think it is solvable, but it will require some tech secret sauce that is currently missing.
I think the secret tech sauce will be to get AAA developers to stop making regular games so in the future it's your only choice.
 
I think the secret tech sauce will be to get AAA developers to stop making regular games so in the future it's your only choice.

That's probably true. Rather than porting existing games, it might end up working better if everything is made for it from the ground up.
I see this tech eventually working, but probably not in the next few years.
 
This is going to flop at least until the internet access the average gamer in the US has is up to snuff. Newsflash: that's going to take a long fucking time.
 
cloud gaming wont be as fast. it just does not seem possible. but fps and stuff will be higher if they are not limiting it
 
Until someone magically solves the internet latency problems, streaming will be a sub-par experience. It may look and test great under pristine test situations... but the internet used by the masses is anything but pristine.

I think it is solvable, but it will require some tech secret sauce that is currently missing.
Quantum entanglement :cool:.
 
I didn't know they allowed mouse and keyboard. Much harder to see slight latency with a game controller than mouse.
 
I'd say legit, full-on game streaming is still at least 5 years away..

Before then, it'll work for some genres but not all.
 
I've tinkered with Geforce Now, and while somewhat impressed with it, this sort of gaming is still a ways off.
What I've experienced with Geforce Now:
Either good visuals with a bit of a latency, or better latency with lower visuals when I set it to run at 120fps mode.

I also run an ideal setup for this, a wired connection with a 300/300 fibre internet. And if my results were not always ideal I imagine someone with a slower setup will not be able to use this at all.

I'd be curious to give this a try but I won't spend any money on it.
 
I do think there are plenty of games that could benefit from this and not be affected by lag. The thing is it is not ALL games. We really need to get out of this all or nothing thought process.
 
Turn-based RPG? Golden Sun funded by Google? I haven't played Darkest Dungeon but always have it on my wishlist.

There was a game called Stratosphere where you build a floating fortress and ram it into opponents, or fire some cannons. You build up momentum and can't turn quickly if they wanted to reboot some games like that where the delay isn't going to ruin it.

Mass Effect 2 was also worth it if the combat and entire game suffered a delay from streaming. If someone had a new IP with a good story to carry it, the protagonist wouldn't have to be a soldier/vanguard/adept. Or they could at least have streaming-friendly engineer classes with drones. That's more passive but it could be fun. And you could pause so easily. Andromeda also had a rifle with enemy-seeking projectiles so you just had to barely aim and it worked.
 
Last edited:
The only way I can see the latency issue being addressed is some kind of predictive AI. I have no idea how such a thing would work or if it's even possible, but certainly there is no overcoming the basic laws of the universe.
 
One thing to consider is that input latency for your controller and even the mouse and keyboard are higher than in the past.

http://danluu.com/input-lag/

This is one major reason that while I can pretty much one shot Super Mario Bros. on a CRT with a real NES, I struggle with the NES Mini and the Virtual Console on my Wii.

Then you have the fact that lag compensation has changed dramatically over the years. I played a lot of Unreal Tournament Instagib. You always had to prefire, even with a T1 connection and 20 ms ping, because if you shot where they were, you missed because the opponent was no longer there. FPSs have changed now so that they calculate where you were and if your shot fired, rather than having to guess where they are.

One major thing that's changed with FPSs though, is that they've slowed down the game, even with twitch shooters. Modern FPSs, there's almost always some sort of delay firing. It's not much, but it's something along the line of pulling out your weapon, which is just a few milliseconds. Older games, the second you clicked was the second you shot.

I have a feeling this will be how some future games resolve the online latency streaming delay. Games will just be slowed down even further to make it appear as if there is no lag.
 
The problem is if it is "good enough" for many players, then it will be adopted and all of the customer-unfriendly policies and monetization possibilities will spread through the industry like a virus - as we've seen with literally every other idea from crappy DLC, to DRM, to Battle Passes, with gatcha/random and microtransactions as a whole being the most visible problem.

While I am glad to hear that there are still technical issues like latency that are an impediment to something like this, that's something that will eventually be conquered to "good enough" for many people. As high bandwidth connections spread and Google (along with a few others like Amazon) has the money and power to put local servers in more places around the nation, latency will be less of an issue. Sure, it may be a problem for those playing the most twitchy, online, pro-level titles...but that's a SMALL group. I'm sure Google will have the money to make sure there's a "Local Stadia Hub" physically backstage at any tournaments or big e-sports events on the platform. Consider something like consoles - there are whole shooter communities on consoles that play with controllers and find it "good enough". Now, will Grandmaster League controller users be at a disadvantage playing against Grandmaster KB/M titles? Depending on the game, probably. However, for most others it won't matter so much.

I'm not saying this to be a downer, but rather simply realism - we can't expect a technical shortcoming to "save" us in this case, because those with vested interest will be looking at minimizing said shortcoming. "Cloud game streaming" of this sort needs to be highlighted as anti-consumer, controlling, and overall bad for gaming... but sadly even that is a hard sell against convenience, for many. We can't let our arguments hinge on outdated examples (ie OnLive) and technical limitations exclusively, as if they are overcome to a point that many users say "good enough", then we lose credibility. We can still take things like latency into consideration, but we must also focus on the things that no matter the technical competency of the technology, will always be centralized, anti-consumer, and bad for gaming.
 
One thing to consider is that input latency for your controller and even the mouse and keyboard are higher than in the past.

http://danluu.com/input-lag/

This is one major reason that while I can pretty much one shot Super Mario Bros. on a CRT with a real NES, I struggle with the NES Mini and the Virtual Console on my Wii.

Then you have the fact that lag compensation has changed dramatically over the years. I played a lot of Unreal Tournament Instagib. You always had to prefire, even with a T1 connection and 20 ms ping, because if you shot where they were, you missed because the opponent was no longer there. FPSs have changed now so that they calculate where you were and if your shot fired, rather than having to guess where they are.

One major thing that's changed with FPSs though, is that they've slowed down the game, even with twitch shooters. Modern FPSs, there's almost always some sort of delay firing. It's not much, but it's something along the line of pulling out your weapon, which is just a few milliseconds. Older games, the second you clicked was the second you shot.

I have a feeling this will be how some future games resolve the online latency streaming delay. Games will just be slowed down even further to make it appear as if there is no lag.
The Wii VC version of SMB runs at the wrong framerate, in addition to having 2 frames of additional input lag before the image is even processed by your display. The NES mini version has the same issues. It's why you don't see anyone speedrunning it on any version but the original game on original hardware.

The article you link is also only measuring keyboard inputs in a terminal, and the author even expresses later on that more gaming-oriented peripherals are just as fast or faster than the oldest and fastest machines. Polling rate of the device has a lot to do with this, and the author even mentions it at the start (the Apple IIe polled at 556 Hz and he compared it to an obscure custom mechanical keyboard that I've never heard of at 167 Hz). Not all devices are made equally: A Corsair K70 has a polling rate of 1000 Hz, while your typical $10 Walmart Special has a polling rate of 125 Hz.

Will agree with on lag compensation. I started playing Counter-Strike with a flag that made the game on the client update when it received information from the server, and that was a game changer for me.

Movement? Yes, but it's not because of "increasing input delay." Designers are striving to make games more realistic, and along with that comes with not allowing players to run at 10 m/s or jump 20 feet into the air. Built-in input delay? I don't think so. In real life there is a small delay that is almost imperceptible as the firing pin needs to trigger the explosive charge in the bullet or shell casing. It is not on the order of magnitude of milliseconds, though. Are you talking about projectile travel time? Because not all games or even individual weapons in those games are using hitscan anymore.

Games being designed around the technical limits of streaming? Unfortunately I can see that happening if it becomes the primary revenue generator for publishers.
 
I'm not saying this to be a downer, but rather simply realism - we can't expect a technical shortcoming to "save" us in this case, because those with vested interest will be looking at minimizing said shortcoming. "Cloud game streaming" of this sort needs to be highlighted as anti-consumer, controlling, and overall bad for gaming... but sadly even that is a hard sell against convenience, for many. We can't let our arguments hinge on outdated examples (ie OnLive) and technical limitations exclusively, as if they are overcome to a point that many users say "good enough", then we lose credibility. We can still take things like latency into consideration, but we must also focus on the things that no matter the technical competency of the technology, will always be centralized, anti-consumer, and bad for gaming.
I'm getting pretty fucking tired of the industry's push for loss of ownership myself.
 
I think the secret tech sauce will be to get AAA developers to stop making regular games so in the future it's your only choice.

The day this happens is the day I give up on gaming going forward and focus on my backlog of shame which is large enough to keep me entertained for several decades.
 
Stepping back, I dont think this is marketed to [H] gamers but to kids who don't have access to the latest gaming consoles and PCs. They'll probably suck a lot of these kids in with free games but get their parent's money using microtransactions.
 
Back
Top