Streaming gaming going to replace conventional gaming
-
Many consoles are thinking about it.
PS3/4 is experimenting with it.PC services are cropping up.
I mean, I already play MMOs, and you're just pushing image deltas instead of world information....
-
pushing image deltas instead of world information....
If this is already happening, it's news to me.
The big challenge is going to be solving a new class of latency problems, and the cost of rendering, encoding, and transmitting high-resolution frames at a sufficient rate for thousands of simultaneous players.
-
The entire point of ingame spectating is that it's not cost efficient to run everyone's game clients on the server. So yeah, news to me as well.
-
Maybe I just blaming how shit I am at CSGO on lag but if my ping goes above 100ms my kdr just plummets I can't imagine what playing fighting games where every frame counts is like.
I've tried streaming Assassins Creed 2 to my living room. It wasn't a pleasant experience. The input lag drove me mad. A longer hdmi cable would of been a better investment. I might try again with cat6 cables and a router that can support it in my new place when Steamlink is released
-
Nobody remembers this?
Well, they did sell their patents to Sony it seems. Maybe they can make it wo...
Sorry, couldn't end that sentence that with a straight face.
-
When it finally happens, I will be too old for games.
-
OnLive
I used OnLive a decent amount when it came out. My brother even got their streaming box and controller.
The infrastructure wasn't quite there at the time. I was living within walking distance of a phone exchange so had really low latency and it mostly worked. I don't play fast paced FPS games, though, they might have been unplayable for all I know. The concept of playing a monthly fee to have access to whatever games you want is a pretty good one, and streaming is, in principle, probably an easier sell than massively DRM-crippled downloads.
-
Yeah, I kinda figured it was just in that "close, but no cigar" range from some testing on a shitty connection all the way across the Atlantic - it was still almost good enough to play stuff like arcade driving games and such. I'm assuming it was pretty much the same for people in US, but with the bar of "almost good enough" being pushed up to something like FPS.
-
I'm in the UK, where we have a proper broadband infrastructure. It was probably awful for most of the US
-
Depends on where you live. I can only get 18mb down and 2mb up. Fine for most purposes but doesn't hand 1080p streaming well.
Maybe we should move to Korea! I mean what can possibly go wrong.
-
-
I expect anything with user-controlled camera will be nogo, though I'm not sure what acceptable lag is for keyboard actions to be reflected on screen. I think console is 30fps typically, which gives you 33ms. PC is typically 60fps so 16ms. Beefy centralized hardware may leave a window of about 20 ms for:
- Upload keystrokes
- Round-trip latency
- Encode frame
- Download frame
- Decode frame
For contrast, I believe head-mounted VR allows something like 5ms before head movement has to be accounted for on screen. Current hardware can't handle that even in the same room. I've got the lowest-latency internet money can buy, and my network latency to Amsterdam datacenters is 3-4ms. Based on that I'd say VR streaming is still very far away.
-
I believe head-mounted VR allows something like 5ms before head movement has to be accounted for on screen.
I read somewhere that the motion sickness can be massively reduced by rendering your nose. That might allow for a higher latency as well
-
We should have a general thread for technologies that are going to come out "any day now" but never do.
Streaming could go together with VR headsets.
Edit: and those guys that promised "unlimited detail" 3D rendering, but that was to be expected.
-
also the year of the Linux desktop
-
-
Replace? No. There are entire genres that aren't playable over a stream. (For example: fighting games.) Maybe in 20-25 years the latency will be low enough.
But it makes an interesting supplement to.
-
Well, they did sell their patents to Sony it seems. Maybe they can make it wo...
Sony's system is full of shitty single-player JRPGs where latency genuinely doesn't matter. That's pretty much the ideal use-case for game streaming. Or turn-based strategy games.
-
Maybe in 20-25 years the latency will be low enough.
and it will probably still not make sense
-
Let's assume that I'm playing a game on a server in Texas. That's about a thousand miles from me. Assuming the other person is located inside the server and somehow they managed to lay a fiber cable directly from the server to my house, the minimum ping I can physically have is 10-11ms.
10-11ms is about 90 Hz. So it is theoretically possible to play a game remotely. But uncompressed, that's about half a gigabyte per second. Let's say we can get that down to a tenth of the size with compression. Now we need a 400 Mbps connection just to play a game, plus a good enough computer that it can handle the half-gigabyte of raw video every second, and we also need to have a direct link to the server with no processing in between.
That's definitely an improvement over being able to play at ≥60fps with a 3 Mbps connection and achieving the exact same result while still being able to use a voice chat program or browse TDWTF.
-
Ben L are you exercising the power of sarcasm?
-
Yeah, isn't it amazing?
-
Ben L are you exercising the power of sarcasm?
-
I just don't see what's in it for the consumer.
In the commercial IT market, the cloud brought in the ability to scale relatively cheaply. As in, instead of going through the hassle of having your own IT department and infrastructure, just lease it as needed. If you have highly variable traffic (eg. graveyard most of the time, but then have a huge spike around Christmas), it can pay off .
You can see they are trying to push the same deal for the gamers too. "You never need to worry about hardware updates and Windows ever again! Just press the ON switch and play!".
Frankly, I don't see it. Gamers don't have the same pain points as the companies. They won't suddenly need 100 game consoles ready to play RIGHT NOW. Hardware updates are slowing down anyway, and consoles give you the same "Just Works" system as the cloud, except with less online-related hassle (at least in theory).
I mean, I can see why companies like Nvidia would like to transition from the fickle hit-based market to a nice stable fiefdom they control. Unlike AWS, I just don't think they'll have enough of a carrot to entice people into the snare.
-
Let's assume that I'm playing a game on a server in Texas.
As opposed to one in Illinois or Wisconsin? Seriously, for this sort of thing to work, the servers have to be pretty well distributed.
The video and audio compression would likely have to be done directly in hardware, which is doable with something custom I think; nobody's really optimised for that yet, but this would be an application that could drive it. The control stream in the other direction would be much lower bit-rate so that would be much easier, and I'd guess they'd be looking at paying for improved QoS shaping at the network level, probably going for lower latency and higher priority at a cost of dropping late or missing packets entirely. Doable, though difficult.
This is unlike multiplayer game servers though. Those are about sharing game world state between multiple people, not giving someone a dedicated remote rendering engine.
-
There's nothing wrong with the "Netflix for games" concept. I might subscribe to it.
-
I just don't see what's in it for the consumer.
It's either cheap, almost disposable dedicated hardware, or the ability to hop PCs without having to download and keep sizable games on them, depending on where they take it. I personally find the other option more enticing - think of sites like Newgrounds, which you can load anytime and anywhere and play something instantly, except instead of shitty Flash games you can play your AAA titles.
Plus it will probably drive the prices lower either via a subscription model (the good idea), or by forcing it upon people and effectively putting an end to piracy (the evil idea).
-
effectively putting an end to piracy
The #1 way of doing that has always been obvious: offer a service than the pirates that the customers think is better. As a rule, people don't mind paying a bit for something if they think they're getting something for it (though how much is variable between countries and areas, of course).
No idea if what the market will bear is economically viable once the costs of offering at that price are taken into account…
-
I personally find the other option more enticing - think of sites like Newgrounds, which you can load anytime and anywhere and play something instantly, except instead of shitty Flash games you can play your AAA titles.
I could never get into Newgrounds, mentally. Even when I liked a game, I always looked for a way to get it downloaded and playing on my computer.
Probably just a generational thing.
-
Isn't that Gabe Newell's entire approach to piracy? He views them as a competing service.
-
This is unlike multiplayer game servers though. Those are about sharing game world state between multiple people, not giving someone a dedicated remote rendering engine.
Notwithstanding that lossy compression of quaternions is a thing in game netcode, and that's just to synchronize orientations for individual objects. Suddenly, we're going to throw all these optimizations out the window and start sending whole frames? Maybe there was, uh, a reason for those optimizations?
-
Isn't that Gabe Newell's entire approach to piracy? He views them as a competing service
Is that why Steam is almost as usable as your average BitTorrent client?
-
ddenly, we're going to throw all these optimizations out the window and start sending whole frames? Maybe there was, uh, a reason for those optimizations?
Different use case. Current game netcode is optimized for dealing with the case where there's rather a lot of crappy networking kit between the communicating parties. It's what you have to do to make things work when synching state across a bunch of geographically distributed computers.
The on-live stuff is different, in that they're probably looking at hosting dedicated specialized servers relatively close to the players (in network terms) through deals with ISPs. They're also offloading a lot of the smarts onto the server; the clients in users' homes are much dumber and thinner: the network is going to be carrying a lot more traffic. It's more like VNC than game netcode.
-
The #1 way of doing that has always been obvious: offer a service than the pirates that the customers think is better. As a rule, people don't mind paying a bit for something if they think they're getting something for it (though how much is variable between countries and areas, of course).
Some decades ago, government here thought it would be a good idea to make the IT market to be a "reserve". So it was ilegal to import anything. Of course, what happened was a big black market and all sorts of shit. Up to 1991, the only way we could get software and games was trough piracy, so we got used to it. Legit software is slowly geting out from being a SJW thing here.
-
I could never get into Newgrounds, mentally. Even when I liked a game, I always looked for a way to get it downloaded and playing on my computer.
That's because it was wrapped in a window.
If you could have a desktop shortcut directly to the game, without the site's frame, you wouldn't know the difference.
-
That's because it was wrapped in a window.
If you could have a desktop shortcut directly to the game, without the site's frame, you wouldn't know the difference.
If everything worked exactly the same as an offline game, maybe.