Discussion in 'Gaming' started by Whalemeister, 24 Mar 2009.
1000 mile range from Datacentres for this to work, 1500 miles with fibre.
I think it will work based on my understanding of it:
None of the game data ever reaches the end user, only video and audio data. Essentially each user will have something like virtual machine running on the onlive servers, which accepts the input from keyboard/mouse/controller over the internet, renders the game on their servers and sends the audio/video feed back to the user through the net.
So input lag will be no greater than with normal online gaming, there could actually be a reduction in overall lag due to the onlive virtual machines having direct connections.
So as long as you're internet connection can stream the A/V feed, you could actually have a better gaming experience than with current online gaming (no framerate drops, etc)
It's ingenious to be honest, wish I'd thought of it.
Edit: http://www.onlive.com/about/team.html the developers are possibly the ugliest bunch of people ever though
I'm going to wait and see it working before I make up my mind, though my default POV is skeptiscism, but if those are key arguements for you then you've obviously never seen Dragons Den
It all rests on the ability of their new video codec, if it really can be compressed and decompressed as quick as they claim, and compress well enough to stream standard-def over 1.5mbps and hi-def over 5mbps, then I think it will change the way games are played.
Everything else is totally plausible, if the codec works I'm in.
True, some people do develop silly ideas, and I remember someone selling their home, losing their family etc over an idea that was so stupid words cannot describe (think it was the US version).
But this is at a different level, if you know what I mean. Why would any big companies and ISPs partner up with an idea what doesn't work.
Some things do seem too good to be true sometimes, it is revolutionary. It they can make the consumer and business side of things work, they have a winner, and the money they make out of it once it becomes mainstream if it does will be quite amazing.
The only problem I can see, which has been brought up before, is consumers like physical media etc. People like to have a PS3 sitting below their TV, people like CD and DVDs sitting on a shelf, they like to feel what they have bought. When you own a PC and make it top end, you like to see the graphics card, it's your own creation, it's yours.
With this, it's slighty different, and some people might not like it. Some people do not like Steam due to this, and Steam downloads the game onto your PC. People spend thousands building a PC so they can have the maximum FPS on games, play Crysis at maximum level and boast about it. If everyone has maximum graphics, does that end the fun of maximum graphics? Will it change PCs as we know them?
Google did say the future of the PC will be server based (or something like that), now I see where they are coming from.
I think there will be rejection from people who like having the bleeding edge of technology and the bragging rights the come with it. But I think the majority of gamers can't/won't afford a few grand a year to keep up with this, so they buy mid-range systems, these are the kinds of gamers that this will appeal to.
I'll accept that a smooth 720p framerate is possible with reasonable image quality over a 5 meg connection, and that - with sufficient processing horsepower - it may be possible to compress and decompress video with sufficiently little latency to make it playable (though 1 ms is, I suspect, total pie in the sky), but network latency alone is likely in my view to kill this. Anyone who has played an online game will be familiar with lag, but it is generally manageable because at least the inputs and graphics are being handled locally, so you have the perception that game controls are immediate. If the game is streamed over a normal broadband connection, that 100ms or whatever will make for nasty sluggish controls. As such, I really can't see this taking off unless a major overhaul of the domestic broadband infrastructure reduces network latency to a much lower, and reliably low, level.
The amount of bandwidth required to send control input data upstream will be trivial (a few kb/s tops), so (aside from the network latency issue) any reasonable domestic broadband connection should be fine.
For RTS games I can play comfortably with lag up to about 1 second. For RPG games I can play comfortably up to around 200-250ms. For FPS games I can play comfortably up to around 150ms. Now I don't mean competitively, but if it was single player mode I don't see how it could cause you trouble. Plus the developer can introduce some compensation. Very little bandwidth is needed up stream, mostly downstream. It would be similar to a game of CS while streaming high quality video.
Like said before, it's all about the codec. They would need some beefy encoder on the server side that will encode it extremely fast, leave great quality, and have it small enough to be streamed. Also, what about frame rate? You might not need 60fps for RTS/RPG but you definently would want more than 23.97fps for shooters and even driving games.
Isn't it like?
Game -> PC -> Unmodded Game Server in some unknown location -> ISP
Watching streamed game controlled by you -> Modded Server in ISP?
The issue is though that when you play online with say 200ms ping, the control inputs and graphics engine are still running locally, so when you move, look around etc. your screen updates immediately. With this system, you move your control stick, the movement gets communicated to a server, the server moves your character or whatever, and 200ms later you get to see the result on screen. Even for an RTS you want a reasonably responsive control interface - imagine trying to use Windows if your pointer moved a quarter of a second after you moved your mouse, or characters appeared 200ms after you typed them. Not fun, and it would be far worse with games where the whole POINT is fun. I'm happy to be proved wrong, but I remain incredibly sceptical.
I don't really follow what you're saying to be honest. The basic difference between this and a "normal" multiplayer game is that in a normal multiplayer game the game is running on your PC (hence immediacy of control input), and is synchronising your actions with those of other players, whereas in this case you will basically be connecting your controller and TV to a PC many miles away and controlling it remotely.
The "lag" in say an FPS is where your PC has assumed someone kept running the same way so displayed them at location X, but in fact they had changed direction and were at location Y but that information hadn't had time to reach the server and then be pushed to your PC yet, so when you shot at X you missed.
The lag on the streamed games will be much more unpleasant, because every action you take will need to make a round trip to the server before it updates. You press the fire button, 200ms later (which is a long time in an FPS) your onscreen character appears to fire his gun.
i for one would absolutely love to see this work. There are doubts it will work but i'd give it my full support because it is ambitious and could be something really special.
Why is the front page always days behind the news?
I'm not going to comment on how well the system works, because neither I or any of you know how well this works.
Fair enough, but I know how well my broadband connection works, and (though I'm happy to be proved wrong) I'm remaining hugely sceptical.
Neither do i, was just some crazy idea.
Exactly. Anyone who has used Citrix (I'm using it now) knows the problem of this.
For those who don't know Citirx, it's exactly the same principal but for Office applications instead of games. You'd think Office apps would be no worries, but you have no idea just how infuritating it is to click on something and it take between 100 and 250ms to respond. It feels as though your computer is running through custard, and that's simply because of the time it takes for the electrons to travel from your PC to the Citrix server and back again. A dedicated billion gigbit line can't speed light up, by the laws of physics there's always going to be a lag between mouse/keyboard input and screen update. Even just typing words is slow. You press a button, and it doesn't appear on the screen for a second, so you're always waiting for the screen to update.
For anything quicker than a turn based strategy game, it's going to get extremely annoying extremely quickly.
The directors at the various ISPs must have sprouted a few more grey hairs, at the thought of even greater pressure on their bandwidth. The main issue I have with idea, is the problem of my normally adequate 8mb Tiscali line, that slows to a crawl between 6 to 10pm. Given that most of us are on shared 50:1 or 25:1 connections, I imagine this would be a common problem. Does anybody know if these streamed games will have any anti-aliasing? I see that the BETA is US only, any UKers been able to successfully sign up?
i think everybody should just wait for it. if it fails, laugh about it in the forums, if it works laugh about it still in the forums. lol
I imagine anti-aliasing should be no problem. It looks like the servers are pretty powerful so they should be more than capable, and AA won't add any network latency.
I think they're saying 1ms on top of your ping, not 1ms connection!
Separate names with a comma.