1ms input lag is BS. The server would need to be within 300km to be even physically possible, and even the practically on the net you're not going to get even close to that. Still, with enough servers scattered about, I could see them bringing the input lag to acceptable levels.
Though one interview said you get a "solid" 30-35 FPS at 720p, which to me is utter garbage. And of course there's the compression artifacts from the streaming.
Though they said that you could just download a small PC client and access the service once it's launched. If that's the case, I'll give it a shot, especially if they have a free trial.
RandomEngy on
Profile -> Signature Settings -> Hide signatures always. Then you don't have to read this worthless text anymore.
1ms input lag is BS. The server would need to be within 300km to be even physically possible, and even the practically on the net you're not going to get even close to that. Still, with enough servers scattered about, I could see them bringing the input lag to acceptable levels.
Though one interview said you get a "solid" 30-35 FPS at 720p, which to me is utter garbage. And of course there's the compression artifacts from the streaming.
Though they said that you could just download a small PC client and access the service. If that's the case, I'll give it a shot, especially if they have a free trial.
It's very ambitious but it is the real deal, this is no Phantom. They've been working on it for seven years and they have the backing of EA, Ubisoft, etc. They're apparently planning to eventually work with cable companies to incorporate it directly like cable TV. It can be done, but not until we improve our infrastructure.
Fixed. Currently network technology has the bandwidth to do what they ask, but not the turbofast response times needed. When people move a mouse or press a key/button, they expect the computer to immediately start processing that. Not "I'll get to it when I can" - now.
PeregrineFalcon on
Looking for a DX:HR OnLive code for my kid brother.
Can trade TF2 items or whatever else you're interested in. PM me.
Ahh. I suppose they need everything they can get to keep the compression/decompression/controller lag from pushing the net lag over into the realm of suckitude.
RandomEngy on
Profile -> Signature Settings -> Hide signatures always. Then you don't have to read this worthless text anymore.
It's very ambitious but it is the real deal, this is no Phantom. They've been working on it for seven years and they have the backing of EA, Ubisoft, etc. They're apparently planning to eventually work with cable companies to incorporate it directly like cable TV. It can be done, but not until we improve our infrastructure.
Fixed. Currently network technology has the bandwidth to do what they ask, but not the turbofast response times needed. When people move a mouse or press a key/button, they expect the computer to immediately start processing that. Not "I'll get to it when I can" - now.
It's just a question of the servers being close enough. A typical broadband connection can easily get <10ms response time within maybe a hundred miles. For the average gamer anything below 50ms is probably tolerable.
Marketing hyperbole aside, I'm thinking about the technical requirements for something like this to work:
1. Server hardware: No virtualization in the world can run games at native speed, so you can't save costs by putting multiple users on a server. One user = one fully equipped computer with GPU. The initial cost will be enormous if they plan on a large userbase, even at wholesale prices, not to mention the power requirements.
2. Dedicated hardware for video encoding: There are hardwareencoders that can do 720P H.264/MPEG2 encoding in <60ms, but 1ms? I'm guessing the bulk of the problem exists here and somehow solved by a proprietary codec that can produce incredible compression (>50%?) in practically no time at all.
3. NOCs in strategic locations: With the speed of light constraint, a cross country connection in the US (4100km) would have a minimum travel time of approximately 15ms (actually higher since electricity travels slightly slower than light). Network operation centers would have to be built out near metro areas for minimum amount of traversal. The bandwidth at each site would have to be substantional, possibly matching regional ISPs, even if they managed to strip packets to the bare essentials and did some serious compression.
The amount of capital needed for something like this makes my head hurt. It seems like a fun challenge, but I don't envy the person who would have to set this entire operation up.
It's just a question of the servers being close enough. A typical broadband connection can easily get <10ms response time within maybe a hundred miles. For the average gamer anything below 50ms is probably tolerable.
Problem is, once you start the "regional servers" shitstorm, you start hitting the problem of "Too many gamers in your area are playing Team Fortress 2. Please try again later." I mean, obviously this system isn't for everyone regardless, but that compounds the issues.
Plus there's the cost of buying/owning/maintaining a server cluster in each area deemed "populous enough to support it" - I'd love to see it happen but we need a much better infrastructure to support it.
PeregrineFalcon on
Looking for a DX:HR OnLive code for my kid brother.
Can trade TF2 items or whatever else you're interested in. PM me.
Marketing hyperbole aside, I'm thinking about the technical requirements for something like this to work:
1. Server hardware: No virtualization in the world can run games at native speed, so you can't save costs by putting multiple users on a server. One user = one fully equipped computer with GPU. The initial cost will be enormous if they plan on a large userbase, even at wholesale prices, not to mention the power requirements.
2. Dedicated hardware for video encoding: There are hardwareencoders that can do 720P H.264/MPEG2 encoding in <60ms, but 1ms? I'm guessing the bulk of the problem exists here and somehow solved by a proprietary codec that can produce incredible compression (>50%?) in practically no time at all.
3. NOCs in strategic locations: With the speed of light constraint, a cross country connection in the US (4100km) would have a minimum travel time of approximately 15ms (actually higher since electricity travels slightly slower than light). Network operation centers would have to be built out near metro areas for minimum amount of traversal. The bandwidth at each site would have to be substantional, possibly matching regional ISPs, even if they managed to strip packets to the bare essentials and did some serious compression.
The amount of capital needed for something like this makes my head hurt. It seems like a fun challenge, but I don't envy the person who would have to set this entire operation up.
Liming for being exactly what I was trying to say last night, but 100Xs more articulate
I was talking about this on another forum, and I felt that this service would really really hurt the modding community if it takes off.
Which makes me dislike it right out of the box.
To be honest, modding has died down a lot. It isn't UT or HL1 anymore. HL2 had barely any modifications. What we have instead is a boom in indie gaming.
Wait what? Seriously? Do you have any idea how many big mods, much less skins and tweaks, there are for HL2?
Heck, even looking at the entrants for the "Make Something Unreal" contest for UE3, there's tonnes of stuff there. Hop on over to Crymod and there's hundreds of user made levels, and loads of mods either in the works or already released.
Look, just have a glance over in this thread, second post, and tell me that there's nothing for Half-Life 2. And that's just for Source games. Over a year ago! That's not even a complete list of what was available back then, it's just stuff that Suds found.
Even leaving all that aside, a vibrant modding community is pretty crucial to the future of the industry, since a large number of people that get their start do so by making mods. Over 50% of Valve's staff started off in the mod scene. Heck, if you haven't played Minerva: Metastasis, I'd say you're missing out. That mod was so well done that Valve actually hired the guy that did it.
I was talking about this on another forum, and I felt that this service would really really hurt the modding community if it takes off.
Which makes me dislike it right out of the box.
To be honest, modding has died down a lot. It isn't UT or HL1 anymore. HL2 had barely any modifications. What we have instead is a boom in indie gaming.
Wait what? Seriously? Do you have any idea how many big mods, much less skins and tweaks, there are for HL2?
Heck, even looking at the entrants for the "Make Something Unreal" contest for UE3, there's tonnes of stuff there. Hop on over to Crymod and there's hundreds of user made levels, and loads of mods either in the works or already released.
Look, just have a glance over in this thread, second post, and tell me that there's nothing for Half-Life 2. And that's just for Source games. Over a year ago! That's not even a complete list of what was available back then, it's just stuff that Suds found.
Even leaving all that aside, a vibrant modding community is pretty crucial to the future of the industry, since a large number of people that get their start do so by making mods. Over 50% of Valve's staff started off in the mod scene. Heck, if you haven't played Minerva: Metastasis, I'd say you're missing out. That mod was so well done that Valve actually hired the guy that did it.
Not to mention games like oblivion and fallout 3 that have entire archives with 1000s of mods.
This is amazing. But there are still many obstacles up ahead. For example this probably wont end buying different consoles for the exclusive titles as I cant see Sony and MS joining the same service.
I was talking about this on another forum, and I felt that this service would really really hurt the modding community if it takes off.
Which makes me dislike it right out of the box.
To be honest, modding has died down a lot. It isn't UT or HL1 anymore. HL2 had barely any modifications. What we have instead is a boom in indie gaming.
Wait what? Seriously? Do you have any idea how many big mods, much less skins and tweaks, there are for HL2?
Heck, even looking at the entrants for the "Make Something Unreal" contest for UE3, there's tonnes of stuff there. Hop on over to Crymod and there's hundreds of user made levels, and loads of mods either in the works or already released.
Look, just have a glance over in this thread, second post, and tell me that there's nothing for Half-Life 2. And that's just for Source games. Over a year ago! That's not even a complete list of what was available back then, it's just stuff that Suds found.
Even leaving all that aside, a vibrant modding community is pretty crucial to the future of the industry, since a large number of people that get their start do so by making mods. Over 50% of Valve's staff started off in the mod scene. Heck, if you haven't played Minerva: Metastasis, I'd say you're missing out. That mod was so well done that Valve actually hired the guy that did it.
Not to mention games like oblivion and fallout 3 that have entire archives with 1000s of mods.
The modding community is very very far from dead.
I forgot about that. Bethesda games wouldn't really be complete without the mod community.
Lack of personal mods might be a downside of the service, but it's not hard to imagine that they could at least install custom maps and total conversion mods on their side.
I am NOT letting myself believe that this is coming out anytime soon. I don't want to get my hopes up and then get crushed.
However, can you imagine the shitstorm Gamestop will throw down if this gets real? I'm willing to bet that they will threaten to withhold shelf space to any title that goes up on this service.
It's very ambitious but it is the real deal, this is no Phantom. They've been working on it for seven years and they have the backing of EA, Ubisoft, etc. They're apparently planning to eventually work with cable companies to incorporate it directly like cable TV. It can be done, but not until we improve our infrastructure.
Fixed. Currently network technology has the bandwidth to do what they ask, but not the turbofast response times needed. When people move a mouse or press a key/button, they expect the computer to immediately start processing that. Not "I'll get to it when I can" - now.
It's just a question of the servers being close enough. A typical broadband connection can easily get <10ms response time within maybe a hundred miles. For the average gamer anything below 50ms is probably tolerable.
As has already been said, not for input lag. 50ms of input lag would render an FPS pretty much unplayable, hell anything above 30 would, and 30 would still be noticeable and annoying. Another thing to keep in mind is that generally with the equipment currently in use routing large amounts of data to residental users increases their latency. If you ping 20 ms to a game server you'll get significantly more latency while transferring 5Mb/s
Honestly, I don't see them being able to get input lag down to the point where it wouldn't get in the way with twitch gaming. Heck, even with platformers like Super Mario World I can't really play it on anything but its original platform because the infinitesimal input lag screws me up. With FPS-es where success depends on lightning-fast reflexes, I can't see this working well at all.
But I think a service like this could lay the groundwork for the perfect MMO. If the gfx and server logic were in the same room, then tons of the crappy little technical restrictions that make MMOs so silly could be done away with.
Imagine real-time terrain modification. Being able to cut down trees and plant new ones. Being able to interact with each plant in the forest. Being able to leave items on the ground. Being able to customize your equipment, to design your own crafted items. Imagine a MMO with actual physics applied to each item and character model. All of this would be possible if the entire game was just running on an gigantic cluster of supercomputers, sending out graphics streams to each of the users. The game could be designed around a 100-ms input lag.
While the "LIVE" buttons are annoying, are there any ABXY configurations that aren't patented? Perhaps they are trying to get this thing out without having to license any patents from anyone?
While the "LIVE" buttons are annoying, are there any ABXY configurations that aren't patented? Perhaps they are trying to get this thing out without having to license any patents from anyone?
I wasn't aware that an arrangement of buttons and their naming could be patented. They could always give the buttons numbers, unless Logitech has patented that.
7 years in the making, huh? I bet these guys had a heart attack when a lot of the big broadband providers started adding monthly bandwidth caps. 3 hours of gaming per day is like 160 GB in a month.
RandomEngy on
Profile -> Signature Settings -> Hide signatures always. Then you don't have to read this worthless text anymore.
- The servers are located within a 1000 mile radius of population centers, so it appears that 4 locations can effectively cover the US.
- As expected, the video compression takes alittle bit away from the visuals
- Crysis runs at 30~35fps in 720p, input lag seems barely noticeable (from a controlled enviroment with a server in Santa Clara, CA)
- Dedicated box (MicroConsole) no bigger than a PSP, with HDMI, USB, and Bluetooth. Onlive guys hinting that the box may be cheap enough to be given away with subscription contract.
- The servers are located within a 1000 mile radius of population centers, so it appears that 4 locations can effectively cover the US.
- As expected, the video compression takes alittle bit away from the visuals
- Crysis runs at 30~35fps in 720p, input lag seems barely noticeable (from a controlled enviroment with a server in Santa Clara, CA)
- Dedicated box no bigger than a PSP, with HDMI, USB, and Bluetooth. Onlive guys hinting that the box may be cheap enough to be given away with subscription contract.
A 1000 miles between populated areas can cover the US effectively? I'm not trusting on input lag until I try it myself, or I hear of tests not plugged into the computer that's hosting.
Now, the box thing I don't doubt, since you can get motherboards a little bigger than that for about $40, if you look
edit- If this thing does work, I want a Linux client, so I wouldn't have to dual boot to stream games
I was onboard, with some Phantom-related doubts, until the input-processed-server-side info came out.
Hell, my Internet connectivity here in town is iffy half the time at best. I'm not confident this will be feasible outside areas in immediate urban centres -- Internet infrastructure sorta sucks ass in my corner of the world.
FireWeasel on
AC:CL Wii -- 3824-2125-9336 City: Felinito Me: Nick
I was onboard, with some Phantom-related doubts, until the input-processed-server-side info came out.
Hell, my Internet connectivity here in town is iffy half the time at best. I'm not confident this will be feasible outside areas in immediate urban centres -- Internet infrastructure sorta sucks ass in my corner of the world.
And to add to this; just about every ISP in the world has oversold their capacity (well maybe except for asian-region(s), I hear they got crazy infrastructure over their).
The only way I could see this really working is if they struck deals with major cable companies in the US to QoS their traffic. Which leads to a tiered internet.
Imagine the rush of high school kids logging into Xbox Live on a Friday afternoon. Now imagine instead of a dozen kbps to play a Halo game per kid, you have a 5mbps video stream. All that competing with other kids (and even the same kids) downloading the latest TV shows, movies and music over bittorrent.
Even if they have adaptive compression, what kid is going to be impressed when his awesome Crysis 3 game goes from720p to 640x480?
The idea of Onlive sounds cool but I'm kind of skeptical about it. Some college campuses, like mine, have bandwidth limitations and penalize any student that goes over the limit. Plus there's the factor of internet connection disruptions. Overall, it sounds cool in that it would allow anyone who doesn't own an expensive high-end gaming PC to play the PC games that usually require high-end PCs but I highly doubt it will be replacing console gaming.
720p? IF you have a fast enough connection...Unless you're playing a Wii, chances are any network hiccup is going to reduce your resolution to unplayable levels.
I don't get the comparisons to Crysis. If you're playing it at 1024.x768 (about 780p, right?) you're doing it wrong. Unless HDTV2 is coming out soon, these resolutions are already behind most PC games. Sure consoles are at this resolution, but PC hardware keeps pushing screen resolutions bigger and better. They make it sound like everyone will be happy with 720p for a decade or so. Or maybe they plan to market this to the Wii-buying market.
El Guaco on
0
Olivawgood name, isn't it?the foot of mt fujiRegistered Userregular
edited March 2009
I think they said 1280x1024 max resolution, equivilant to about 780p or whatever the fuck the HDTVs are these days
Frankly that's my go to resolution for most graphically taxing games these days, particularly on my laptop which does not have a serious GPU, and if the technology does work and Onlive does succeed, then I have a feeling they'll be working on getting better resolutions as time goes on
This is the exact concept I have been discussing with my friends for years now. I always said eventually people wouldn't need to bring their super computers around to lan parties, that at some point a streaming service will exist to allow a regular old laptop to interact with a tricked out computer running the game somewhere in real time. I just didn't think this tech would be revealed so soon.
Tasteticle on
Uh-oh I accidentally deleted my signature. Uh-oh!!
Posts
Though one interview said you get a "solid" 30-35 FPS at 720p, which to me is utter garbage. And of course there's the compression artifacts from the streaming.
Though they said that you could just download a small PC client and access the service once it's launched. If that's the case, I'll give it a shot, especially if they have a free trial.
They said 1ms video compression lag.
Fixed. Currently network technology has the bandwidth to do what they ask, but not the turbofast response times needed. When people move a mouse or press a key/button, they expect the computer to immediately start processing that. Not "I'll get to it when I can" - now.
Can trade TF2 items or whatever else you're interested in. PM me.
Ahh. I suppose they need everything they can get to keep the compression/decompression/controller lag from pushing the net lag over into the realm of suckitude.
It's just a question of the servers being close enough. A typical broadband connection can easily get <10ms response time within maybe a hundred miles. For the average gamer anything below 50ms is probably tolerable.
1. Server hardware: No virtualization in the world can run games at native speed, so you can't save costs by putting multiple users on a server. One user = one fully equipped computer with GPU. The initial cost will be enormous if they plan on a large userbase, even at wholesale prices, not to mention the power requirements.
2. Dedicated hardware for video encoding: There are hardware encoders that can do 720P H.264/MPEG2 encoding in <60ms, but 1ms? I'm guessing the bulk of the problem exists here and somehow solved by a proprietary codec that can produce incredible compression (>50%?) in practically no time at all.
3. NOCs in strategic locations: With the speed of light constraint, a cross country connection in the US (4100km) would have a minimum travel time of approximately 15ms (actually higher since electricity travels slightly slower than light). Network operation centers would have to be built out near metro areas for minimum amount of traversal. The bandwidth at each site would have to be substantional, possibly matching regional ISPs, even if they managed to strip packets to the bare essentials and did some serious compression.
The amount of capital needed for something like this makes my head hurt. It seems like a fun challenge, but I don't envy the person who would have to set this entire operation up.
Problem is, once you start the "regional servers" shitstorm, you start hitting the problem of "Too many gamers in your area are playing Team Fortress 2. Please try again later." I mean, obviously this system isn't for everyone regardless, but that compounds the issues.
Plus there's the cost of buying/owning/maintaining a server cluster in each area deemed "populous enough to support it" - I'd love to see it happen but we need a much better infrastructure to support it.
Can trade TF2 items or whatever else you're interested in. PM me.
http://www.fallout3nexus.com/downloads/file.php?id=16534
I can think of plenty of places with vastly superior broadband infrastructure to test this with. How about South Korea, where they have 100 Mb/s right now and will be moving to 1000 Mb/s by 2012?
PSN:RevDrGalactus/NN:RevDrGalactus/Steam
Liming for being exactly what I was trying to say last night, but 100Xs more articulate
Wait what? Seriously? Do you have any idea how many big mods, much less skins and tweaks, there are for HL2?
Heck, even looking at the entrants for the "Make Something Unreal" contest for UE3, there's tonnes of stuff there. Hop on over to Crymod and there's hundreds of user made levels, and loads of mods either in the works or already released.
Look, just have a glance over in this thread, second post, and tell me that there's nothing for Half-Life 2. And that's just for Source games. Over a year ago! That's not even a complete list of what was available back then, it's just stuff that Suds found.
Even leaving all that aside, a vibrant modding community is pretty crucial to the future of the industry, since a large number of people that get their start do so by making mods. Over 50% of Valve's staff started off in the mod scene. Heck, if you haven't played Minerva: Metastasis, I'd say you're missing out. That mod was so well done that Valve actually hired the guy that did it.
Not to mention games like oblivion and fallout 3 that have entire archives with 1000s of mods.
The modding community is very very far from dead.
Hey, I have a blog! (Actually being updated again!)
3DS: 0860-3240-2604
I forgot about that. Bethesda games wouldn't really be complete without the mod community.
However, can you imagine the shitstorm Gamestop will throw down if this gets real? I'm willing to bet that they will threaten to withhold shelf space to any title that goes up on this service.
Yeah that is pretty awful. At least the L is the left button. And I won't need that, playing on a PC.
And according to the interview, you get 720p at 60 fps. So if it delivers on that and the super-low response time, this could have something to it.
Hopefully it goes live in 5 and a half hours and we can try it out.
Hell, I'm still sort of worried that once the website goes live it'll just show a looping video of Rick Astley
I mean I don't think the technology is able to make something like this work yet, but I would love to be proven wrong, you know?
PSN ID : DetectiveOlivaw | TWITTER | STEAM ID | NEVER FORGET
But I think a service like this could lay the groundwork for the perfect MMO. If the gfx and server logic were in the same room, then tons of the crappy little technical restrictions that make MMOs so silly could be done away with.
Imagine real-time terrain modification. Being able to cut down trees and plant new ones. Being able to interact with each plant in the forest. Being able to leave items on the ground. Being able to customize your equipment, to design your own crafted items. Imagine a MMO with actual physics applied to each item and character model. All of this would be possible if the entire game was just running on an gigantic cluster of supercomputers, sending out graphics streams to each of the users. The game could be designed around a 100-ms input lag.
If something like this ( http://www.circleid.com/posts/cerns_the_grid_broadband/ ) becomes available to consumers, you can bet on games becoming absolutely awesome.
I wasn't aware that an arrangement of buttons and their naming could be patented. They could always give the buttons numbers, unless Logitech has patented that.
- The servers are located within a 1000 mile radius of population centers, so it appears that 4 locations can effectively cover the US.
- As expected, the video compression takes alittle bit away from the visuals
- Crysis runs at 30~35fps in 720p, input lag seems barely noticeable (from a controlled enviroment with a server in Santa Clara, CA)
- Dedicated box (MicroConsole) no bigger than a PSP, with HDMI, USB, and Bluetooth. Onlive guys hinting that the box may be cheap enough to be given away with subscription contract.
From IGN:
- Burnout Paradise running close to 60FPS, with a vague statement about it being pretty responsive
- Invite only beta in the Summer, release expected in Winter 2009
A 1000 miles between populated areas can cover the US effectively? I'm not trusting on input lag until I try it myself, or I hear of tests not plugged into the computer that's hosting.
Now, the box thing I don't doubt, since you can get motherboards a little bigger than that for about $40, if you look
edit- If this thing does work, I want a Linux client, so I wouldn't have to dual boot to stream games
Hell, my Internet connectivity here in town is iffy half the time at best. I'm not confident this will be feasible outside areas in immediate urban centres -- Internet infrastructure sorta sucks ass in my corner of the world.
Imagine the rush of high school kids logging into Xbox Live on a Friday afternoon. Now imagine instead of a dozen kbps to play a Halo game per kid, you have a 5mbps video stream. All that competing with other kids (and even the same kids) downloading the latest TV shows, movies and music over bittorrent.
Even if they have adaptive compression, what kid is going to be impressed when his awesome Crysis 3 game goes from720p to 640x480?
I don't get the comparisons to Crysis. If you're playing it at 1024.x768 (about 780p, right?) you're doing it wrong. Unless HDTV2 is coming out soon, these resolutions are already behind most PC games. Sure consoles are at this resolution, but PC hardware keeps pushing screen resolutions bigger and better. They make it sound like everyone will be happy with 720p for a decade or so. Or maybe they plan to market this to the Wii-buying market.
Frankly that's my go to resolution for most graphically taxing games these days, particularly on my laptop which does not have a serious GPU, and if the technology does work and Onlive does succeed, then I have a feeling they'll be working on getting better resolutions as time goes on
PSN ID : DetectiveOlivaw | TWITTER | STEAM ID | NEVER FORGET
Uh-oh I accidentally deleted my signature. Uh-oh!!
EDIT: Oops, read the counter there wrong, ignore this
Uh-oh I accidentally deleted my signature. Uh-oh!!