Slashdot Mirror
OnLive One Step Closer
hysma writes "It looks like OnLive, the remote gaming system that streams HD video over the Internet, is one step closer to becoming reality, according to an article on DSL Reports in response to a lengthy video presentation by founder & CEO Steve Perlman at Columbia University. Perlman demonstrated the UI, spectating, using the service on an iPhone, and other features."
Until round trips between the server and client are guaranteed to be under 50ms, the lag will feel unbearable. If someone is playing a racing game and has to deal with a second between the time they begin turning and the time they actually see it turn this service will be dead before it begins.
There are still no plans to support alternative platforms outside of Windows and Mac which is actually a bit disappointing. Onlive could have knocked out one of the major reasons why many people stay with Windows.
Sigs are too short to say anything truly profound so read the above post instead.
Why "Games:"?
Wake me up when it can be deep-fried over the internet.
I also strongly doubt that any kind of game on this platform can be enjoyed by people who are sensitive to input latency. For example my old high quality PVA TFT panel used an overdrive circuit to reduce ghosting. The overdrive logic in TFT panels usually buffers about 1 or 2 full frames to analyze and optimize the pixel voltages which leads to about 20-50ms input latency. I for one already notice it when I just work to the point where it annoys me when the desktop or terminal sessions somehow always feel sluggish, let alone fast 3D games.
I can't imagine that the complete round-trip time for sending my input over the internet, waiting for a frame to be rendered and encoded remotely, sent back over the internet, decoded and displayed locally would be less than 20ms and then you'd still have the latency of your display. It might be bearable with a very fast internet connection and a CRT display which has 0ms input latency.
Maybe others aren't that sensitive to latency and can enjoy at least slower games like turn-based strategy with this service. Good for them.
So basically, this is Cloud Gaming.
Ok, so in order to improve and maintain a consistent FPS rating (rendered, not just streamed), do you have to purchase "upgrade points". Basically, a virtual hardware upgrade for your virtual gaming rig session?
Life is not for the lazy.
Yeah sure... replace a low-bandwidth, local application with a remote one that heavily relies on a fast network [...] Am I missing something here
You're missing ssh + screen + emacs. That used to work fine on 9600 baud terminals; it should work fine over even a measly 1 megabit intertube. In fact, I know it does; I use ssh + screen almost every day (sometimes including emacs).
i know that for investors they need to show some virtual activity, but in the world where 800kB youtube video loads few seconds it is not gonna work for a long time... Even video-conferences almost perfected at this moment have to be crippled in peak times in most of the world, so at least we dont need more /. posts on it
God's gift to chicks
If ofuscated text is easily readable (anti-captcha spam bots), text with not distortion at all will be perfectly readable, so if you send video, the other side will save that video (either with hardware or software) and use OCR to get back the text.
-Woof woof woof!
A game like rock band could 'tune it in'. [it=500ms lag]
I really don't think it could.
Here's the reason: suppose I'm to go red-green-blue-yellow-orange-yellow-blue-green-red really fast (say, at the end of the TTFAF intro), and it's one big hammer-on-pull-off sequence which can't realistically be strummed (or the rules of the game have changed so I have to HOPO).
I miss the first green.
I only get to know that I missed the first green 500ms later. I have already HOPO'ed the rest of the sequence. There's no way I can go back in time 450ms and strum the blue I HOPO'ed, undoing the not-playing-correctly.
It's not just that you have to compensate for lag between inputs and outputs. You also have to make the lag inside a feedback loop very small. A minimal lag of 500ms is too much for rock band. ... Even if the audio and video is perfectly synchronous, and the game compensates for the output lag by virtually moving your inputs back in time. The game can never move the reaction to the output, which happens in your brain, back in time to before the output.
User with 50 ms ping, the optimistic traceroute
0 ms: User see on the screen his car moving left,
200 ms: User press "D" to move his car right.
201 ms: OnLive process the "D" and send the message to the server. The message is on the home router.
251 ms: OnLive server receive the "D" command.
301 ms: OnLive server generate the next frame.
321 ms: OnLive server compress the frame, and send it to the client. The data is on the server router (80KB)
(322 ms: User press "A" to move his car to the left.)
371 ms: Home router receive the data.
471 ms: Onlive download the whole 80KB of data.
472 ms: Onlive uncompress the 80KB of data has 1024x768x16 bits of video data (??) ( compression: 153.6 % )
482 ms: Data is rendered on the screen on the next retrace.
What the user see:
User see his car to collide, press D to move right, wait 322 ms and nothing occur, press A to move left. 482 ms after the first keystroke, the car move right.
-Woof woof woof!
I remember reading in the Time/Life book on the brain that adding an appreciable delay to the auditory feedback you get makes it very difficult even to talk properly. But I'm sure that's old research.
There doesn't seem to be that much on it on the net (or maybe I'm not searching properly). The WP article on Delayed Auditory Feedback has a link to a paper with similar work but it is also from 1979.
Every damn time there is a story about OnLive or other similar services, you see dozens of messages by people speculating that it's impossible and latency will kill it. This conversation is get BORING. We get your point, and we got it the 500th time we heard it. You all sound like people having a heated argument over whether some comic book character can jump over a building or not. Unless you have tried the system, then STFU. At this point it's just vaporware! Maybe you are right; maybe you are wrong - these guys are working with the big boys - major ISP and hardware providers, and may be working to change the actual structure of the network itself to support the game, as opposed to changing the system to work on the current network. In any case, just wait and see and quit boring us with your armchair network admin speculations.
There is a fine line between being a cultivated citizen and being someone else's crop. - A. J. Patrick Liszkie
the richer control setup of keyboard and mouse.
It's really only the mouse---you can make a perfectly fun FPS that's playable with the buttons on, say, a wiimote plus nunchuk (one stick for moving, 8 buttons in B/C/Z/d-pad/A).
What's really missing is the fine-grained relative motion of the mouse.
It needs to be fine-grained; anyone who has tried to aim with a joystick will understand why.
It needs to be relative, as anyone having played Metroid Prime 3: Corruption on the Wii will know.
Roughly speaking, you point at an absolute point on the monitor plane; your character yaws and pitches gradually to aim at that point, the speed being monotonously increasing in the distance between your current aim point and the target aim point.
What are the implications? Either you point at where you want to shoot and it takes a while to aim there, or you point way past where you want to shoot and you get to where you want to go really fast but move away again really fast.
What you really want to do is overshoot by infinity (or $BIGNUM), then aim at the target point when your character points exactly at it: then you get to your target fast and stay there. This is virtually impossible, and trying to do it is unpleasant.
That's why you need a mouse for FPSs; you can make games that only take 8 buttons, so I don't buy the "you need a keyboard". Maybe specific FPSs require keyboards, and maybe there's really no way to design around that without making the game a different game---I'll buy that. But really it's the mouse.
I still maintain that this simply can't work, and that it's an absolutely braindead money pit of an idea if it's not a total scam.
Idea: let's take the most latency sensitive, computationally demanding, and visually intensive thing you can do with a modern computer and try to apply the thin client model to it.
A single instance of the application in question will demand the full resources of the most powerful PC you can throw at it, but we'll just wave our hands and mutter something about virtualization to convince stupid investors that we have magic at our disposal. Because they are morons and because we put on a good show, they'll believe that you can somehow run many instances of a game on the equivalent of a single PC. We'll also be encoding 720p video in realtime at a quality / bandwidth ratio that no codec today can deliver; this will presumably happen on the same computing hardware that's already running multiple instances of cutting edge 3D games.
Finally, we'll throw in some shit about the iphone, because people can't stop fellating apple lately.
Anyone who believes this is technically feasible, much less economically viable, is fucking *retarded*.
Great concept, I can't wait to play with it in person :) A few thoughts:
I'm a little skeptical about how robustly it will perform, but I am sure they will have a chance to prove their technology soon. I'm sure everyone who's played online has dealt with lag spikes, just due to random congestion, noise, route changes, that sort of thing. It seems that this system will be much more sensitive to those kinds of network delays.
One thing that they didn't talk about was really how high latency-sensitive games fit into this framework. I'm talking about timing games, such as Guitar Hero, or games that need twitch reactions like Street Fighter. Players of these games frequently complain about latency that happens between the console and the display, nevermind the the latency of a network. (HDMI decoding delay has made playing these games kind of a pain on some HDTVs) Anecdotally, I notice that in these kinds of games, delays around 5-15ms delay can make a huge difference.
Even still, I think a lot of action games can be very successful on this kind of system. Platformers (like Super Mario, Little Big Planet) and driving games do involve real-time reaction, but the way we play these things, our actions are more predictive, than reactive. ("I'm running towards the edge of the roof, and I'll get these in about half a second, so press jump.... now!") In these kinds of games, our brains sort of "build in" the latency into our actions, so we're not as sensitive to them.
One point that I think they should have emphasized is cheat prevention. On many PC games, there's trainer's, aimbots, all sorts of cheats that make playing on public servers very suspect. With this system, they're reducing everything down to just control input and video output, so the opportunity to cheat in a game is significantly harder.
I have a funny feeling that the monthly subscription fee will be something more than I'll want to pay. They have to deal with the cost of maintaining high-end gaming servers, and what I'm sure will turn out to be an enormous network bill. I'm sure it will be reasonable, because in their model, instead of maintaining my own high-end gaming PC, the burden will be on them to keep the hardware up to date. It will just be more than I want to pay.
--
#include <malloc.h>
free(your.mind);
I think the most interesting part was the (lack of) answer about how the compression works.
They claim 80ms round-trip latency from button push to image display. Running a game on a server and screen-scraping in ~20ms is fairly easy. With proper datacenter placement and peering agreements ~50ms round-trip ping times are reasonable (if somewhat optimistic). The issue is how do you compress the 720p image and send it back in 10ms with reasonable bandwidth.
They're claiming 1ms compression, 8ms decompression (125hz), and 5mbit 720p streams. The compression is using a custom ASIC, so that's completely believable. Decompressing at 120hz on any generic hardware (they specifically said no GPU help) means it has to be an extremely simple protocol. The biggest question is how do you reach "HD-quality" at only 5mbit when you are not doing group-of-pictures compression (keyframes and diffs from the keyframe). Mind you that a standard DVD is 10mbit, so they're claiming higher resolution with half the bitrate and no keyframing. Obviously H264 gets better quality/bit than DVDs, but it does so by using even more complex keyframing and diffs and is far too CPU intensive for their target platforms (it's hard to watch 30fps H264 trailers on many machines, let alone a 60fps stream). The only hint he gave was some mumbling about visual perception, and the statement that their compression only looks good in motion (if you paused the stream it would look terrible).
Any ideas as to how the compression works?
And after that you will have to reconstruct the document from the OCR output, which will take quite some time because people don't scroll by nice segments (i.e. by consecutive pages) while programming. And there's another issue, libraries. You can't OCR binary files which you don't see the contents of.
And I for one cant wait until they release Duke Nukem Forever on it!
Because you always need a smart fox!
Is to rent console time over the Internet, to people with enough money to have a PC that will run this stuff, and a fast Internet connection... ...or an iPhone, a platform known for its cost-effective pricing model... ...but don't want to buy their own console, because it would clearly be too expensive?
Of course, people don't want to all play computer games at the same time, so I can see they'll be balancing load throughout the day... erm... or not (and certainly, they're not going to be running connections internationally with latency that's anything less than abominable for this).
In summary: WTF?
The latency problem is of course the most apparent and thus the most discussed but there are others.
One I wonder about is what kind of servers they are supposedly using. The problem is that modern games demand a modern GPU to look good. The kind of processing needed cannot be done on any sort of reasonable processor in realtime. Also, GPUs aren't really set up to work in parallel these days. What I mean is if you try to have a system with multiple GPUs and running multiple 3D games on them, you are going to find that doesn't really work. That sort of thing is coming, DX11 generation hardware is much better at multi-tasking and such, but it requires apps to be rewritten for it and still isn't there.
So what it comes down to is that to run a modern 3D game, well you have to have a desktop system more or less. You need to have a system running Windows with a powerful GPU at its disposal, and it needs to be tasked to running that one game.
Well that isn't a situation I'm seeing as working real well for a hosted business model. You have a whole bunch of individual desktop machines set up that then load up the software and whatever handles the encoding.
If they are claiming they are doing it with "virtualization" then I'm saying they are "lying." As it happens, doing virtualization related things is a big part of my job, so I'm fairly up on the tech. When it comes to 3D with VMs there are two things that are true of every technology that supports it:
1) It doesn't work real well. It is on the slow side, and there are bugs of various sorts. It is for sure usable, but nobody is going to confuse it for being 100% good to go, and newer games are the thing it has the most trouble with.
2) It requires a 3D card on the host. All of the virtualizaiton solutions do 3D by processing the guest 3D calls and translating them in to 3D calls to the host. 3D hardware is then needed to do the actual rendering.
I'm afraid I don't buy that these random guys have a more advanced technology than VMWare, Sun, Microsoft, and so on. If you could easily virtualize a system and emulate full modern 3D in software, well they'd be doing it. Hell, MS would be interested in doing it non-virtualized. Be a cool selling point of a new Windows if you didn't need a GPU anymore.
So the only way I see this working is lots and lots of systems with big graphics cards in them. This I do not see as a profitable proposition, even assuming all the rest of it works flawlessly.
They have been pretty open about latency issues. The server needs to be reasonably close (max 1000 miles) to keep round trip time below 80ms.
Well, they say 80ms and 1,000 miles. But first of all, even that seems a little dubious, for two reasons:
In general, OnLive (/GaiKai) are interesting, but to work, they need radical breakthroughs in several separate areas, primarily compression and networking. It is possible they solved both. However, if you had a breakthrough in even one of them, it would be lucrative enough by itself. Whereas they are relying on multiple breakthroughs for their specific product. So, I guess it's possible, but it just seems odd. Add to that the concerns about lag even with 80ms, and something is fishy.
Furthermore, this will be expensive. Even with the 'cloud gaming' model of sharing hardware, they will need at least 5 machines per customer (TechCrunch guesses more actually), which means just for the hardware costs people will need to pay hundreds of dollars a year. Add to that the cost of bandwidth - this is very, very bandwidth intensive, ISPs will raise prices if this takes off - and the other OnLive hardware costs, plus additional OnLive costs for salaries and so forth, and this will not be cheap. Even if this worked as promised, would enough people be willing to pay a large amount of money for it? I'm not sure.
The "User press A" is because the problem ocurrs on the clientside wen the player overcompensate. A accelerator can have a delay of 1000 ms, and you will not notice, because the engine have a builtin delay, but If your direction have a 1000 ms delay, your car will move like you are drunk. Withouth "User press A", the experiment is somewhat moot point. The end result is that after the player ask the car to move left, the car will move right.
----
} 471 ms: Onlive download the whole 80KB of data.
} wtf? what's that 100ms for?
downloading the 80KB of data. is that too much / too lite / wrong?
} 472 ms: Onlive uncompress the 80KB of data has 1024x768x16 bits of video data (??) ( compression: 153.6 % )
} wtf? 1 ms to uncompress?
i assume really fast CPU processing. my log is the "optimistic" log.
-Woof woof woof!
parent is a troll
Even if this purple unicorn was possible, giving them the benefit of the doubt that they can pull off everything they claim on server-side, as packets trickle from conglomerate to backbone and up and down the tiers of ISP's, JohnDoeNet Inc. will not appreciate such a surge in traffic if this became popular. Facebook is one thing, but streaming 'HD' video alongside a bunch of gaming data, good luck. I have had multiple ISPs (local and national) enforce a cap on my bandwidth because of comparitively lightweight PC games. I call VaporWare.
I thought the days of using buzzword dot-commyness for the sake of luring investors died ten years ago. Some folks never learn.
'We are trying to prove ourselves wrong as quickly as possible, because only in that way can we find progress.' RPF
So far this thread is all "Well, I'm sitting here in my armchair and I believe..."
History will judge them, not self-important speculation.
PS: They're aiming at iPhones, not your $10,000 nitrogen-cooled neon-lit gaming rig.
No sig today...
The biggest selling graphics chip is Intel - selling more than NVIDIA/ATI combined.
This could open up 3D gaming to all those people with 'underpowered' machines (generally their CPU/RAM/Disk is as good as yours).
No sig today...
All these whiners claim latency will kill this before it ever starts. Guess what? FPS is not the only game category. I game all the time, but latency means nothing to games like Civ4, Neverwinter, and thousands of others. Sure, lower latency is a great goal to aim for, but this platform is a good step towards moving MMOGs onto lower powered clients. The games are just an excuse to extend their reach to more customers in a "new" way, so they can sell them things.
On the one hand you take life too seriously, and on the other, you do not take playful existence seriously enough. Seth
Then all the people bitching about 5ms here and 15ms there and divide-by-fps-so-see-it-can't-possibly-work will either be vindicated, or look like idiots.
Though, you never know - it might be playable by 99.999% of the populace, but not for them.
Either way, I'd love the chance to see. Doubly so if I can pull off my Longest Yard railgun hits.
"Sometimes a woman is a kind of religion, she can save your soul & set you free from all your sins" - Bad Examples
Your other responder mentioned OCR and libraries, but the other issue from the original post was that they would take the entire revision history, which is generally worth considerably more than the head revision (which might not even build at the moment, let alone work correctly!).
I feel fantastic, and I'm still alive.
At least, not in any way, shape or form for people who don't live in close proximity to the servers. For me, I can usually get an average ping of about ~80ms to any given person or site on the internet; Even if the video feed had no additional overhead (and it invariably does), that still takes at least 80ms from when I send input and get to see a response on-screen, at the best of times. People already have issues with some LCD displays and input lag, and that's only on the measure of up to 10-40ms. Add in minimal extra latency for things like wireless controllers, some added latency in-game by console developers trying to squeeze in some extra FPS (sometimes upwards to another 150ms for console games), and you've got some pretty decent lag going on - even more if you've decided to play multiplayer with someone not also on the cloud. In addition, compression artifacts will be present, and requiring a constant 1-2 mbps at minimum for the duration of the session (for "SD"; 5mbps for "HD") is going to make it difficult for most people to play anything with a resolution higher than what consoles were using half a decade ago, or what PC's were using nearly two decades ago.
If the service actually does ship, I have a hard time believing it's going to "revolutionize" anything. I'm all for letting people in on the gaming experience without having to go out and buy consoles or killer PC's, but this certainly isn't the way of the future.
Screw the rules, I have green hair!
Unless they have figured out how to go faster than the speed of light this simply can't work for action games. Their only hope can be to get bought by a cable company that would offer casual games or non-action games.
http://www.popularculturegaming.com -- my blog about the culture of videogame players
Why would i want someone spying on my game play?? or even what I'm chatting about with another player.
Jack of all trades,master of none
The only way this could NOT be a scam is if they have accomplished two or more of the following:
1. Made incredible leaps in computing power that nobody is aware of yet
2. Overcome many laws of physics
3. Decided to charge people so much that they might as well have bought a computer/console anyway
Also, my television has about 100ms of input lag. Add another 50-150 for Onlive to render, encode video, and stream back, and it won't just be FPS games that are unplayable. Nobody is even going to want to play Plants vs. Zombies when it feels like you're moving your mouse through hot tar.
Let me get this straight: People with low-end hardware can pay a fee to run the games on OnLive's hardware, and stream the video over the internet ? How is that better than spending money ONCE on a mid-range GPU and running the game yourself ? What about bandwidth surcharges ? High-def video doesn't come light, most people will blow their monthly caps in no time.
A modest gaming-ready card is $100 or so. Even if OnLive charged only $9.99 / mo, in less than a year the GPU is a cheaper option. The other thing is when you cancel your OnLive subscription, you can't play anymore. That GPU is yours forever until it dies in a fire. If and when you feel the urge to upgrade, you can sell the old GPU for 20-30% of the cost of a new one.
Am I missing something, or is this whole business idea an exercise in futility ? I get that most people are dumb as dirt, but come on...
-Billco, Fnarg.com
.
.
Now think those videos are prerendered, onlive say they are streaming 1080i that is rendered in real time.
I just don't see it working for any real time game, FPS, RTS, etc, racing, which is what they seem to be selling it for. Adventure games would be fine, but we all know how much money there is in those.
Who is the market here? How many people have high bandwidth AND low leniency connections close enough to their servers not to have lag? 99% of potential customers will have to deal with lag between several 3d party network/server hops. It could work when everyone has much much better internet connections then they do now, but in 2010? ha.
As I check around the web for this thing, I see a lot of dates of April 1st. When was this video actually shot/released?
It's pretty horrible in 720p HD, sluggish feel, blurry visuals until you're still etc... and we were not able to get to extensive MP gameplay. Our office is less than 10 miles away, ping 20ms.
Not to worry, basic laws still hold, reading recent WIRED - failure is just as solid as success!