Unlimited Detail

Technology and Gaming
#21

correct me if I’m wrong, but aren’t google’s “server banks” used for storage rather than computational power? And isn’t the search time determined by your network bandwidth rather than CPU power?

Like I said before, the easiest way to understand what “unlimited detail” is talking about is to imagine google maps - which takes an infinitely large image file, determines what you’re looking at and only renders that tiny subsection of the large image - only working with 3d geometry rather than 2d images.

#22

yes but you kinda glossed over the “determines what you’re looking at”. That part isn’t the simplest task when your in a 3d space.

#23

Of course its not simple, thats why people are talking about this tech - “ray casting into a sparse voxel octree.” “Ray casting” is how the engine determines what you’re looking at. “Sparse Voxel Octree” is a data format which is structured in such a way that the information it stores can be looked up extremely efficiently.

edit - looks like the next Cry Engine might use this, too.

#24

I was thinking about this kind of 3D rendering system, you could create a model with the points defined, but on rendering create tiny differences to make more unique objects. I’m glad to see somebody has done the effort to make this.

#25

I think with unlimited detail, as long as you have enough memory, graphics processing will be bottle necked by the cpu (or whatever hardware is dedicated to the individual pixel calculations) and your memory access times. No longer will your FPS drop because some GPU cant push the geometry fast enough. Instead, your FPS will drop because the individual pixel calculations may not be occurring fast enough or you are unable to pull the pixel information from the data array in time.

Unlimited detail does not work with geometric shapes like current GPUs. GPU’s draw and redraw polygons created by the developers. Instead, with unlimited detail, everything is calculated at the individual pixel level. So, I think, the number of objects unlimited detail processes is equal to the number of pixels of your resolution (i.e. 1920x1080 = 2073600 pixels/objects to be calculated). Currently, the number of objects a GPU processes is proportional to the number of polygons displayed on screen at that time. Keep in mind that geometric calculations are more complex and time-consuming than individual pixel calculations.

Unlimited detail requires a large library of data (cloud) and fast access times, not powerful GPUs intended to create geometric shapes.

#26

Isn’t the OnLive idea going to make things like this irrelevent anyway?

The basic premise being that you have a game installed on server somewhere, as you would a website, and you connect to the server to play a game. You use all the servers hardware resources to play the game and stream the video to your computer over the magical interweb. That way it doesn’t matter what system you’re running at home, only that the hardware in the server can cope (which is what your paying for) and you have a decent internet connection.

At the rate the internet is speeding up, that sounds much more likely to happen soon and make PC gaming hardware updates a thing of the past. At which point, finding solutions such as this are no longer viable because no one will care what kind of performance it takes to run a game because it won’t matter to them.

#27

But it will matter to the company running that system, if they can do the same thing cheaper, they probably will.

#28

Rule #1 of new technologies: They are never as good as the advertisers say.

#29

True dat.

#30

The problem with that is that you always have some delay over the internet. Even if it’s only 10ms, it’ll be 20ms to go back and forth. Add to that the delay to other players in multiplayer games, and it can get pretty high.
Everyone involved has to have a lightning fast internet connection.

#31

I’m aware of the limitations but the system looks to go live from later this year. By the time the Unlimited Detail gubbins is ironed out enough to launch an OnLive type system will be well established I’m sure.

Neither system is going to run perfectly today, I’m just saying that UD will most probably be largely irrelevent by the time it is launched from a consumer point of view. Someone mentioned a 10 year timeframe, which I don’t think is unrealistic. That would put the technology presumably 8 years too late.

#32

One huge advantage I’ve just thought of with the onlive thing is compatibility. You could play games through that from any system, be it Linux, Mac or Windows.

#33

OnLive wont make new operating methods such as Unlimited Detail irrelevant, something still has to process the game data whether by software or hardware. The OnLive computers have to use something, right now it is hardware. In the future however, if projects such as Unlimited Detail succeed, they will use the most efficient and cost effective method.

#34

I’m not going to subscribe for it, and I don’t know anyone who would. Steam hasn’t made physical retail irrelevant, and onlive won’t either.

#35

There’s no way onlive will be able to make anything irrelevant when something as comparatively simple as Assassin’s Creed 2 DRM can’t even work right.

#36

I fucking hate the idea of onlive.
I seriously think it should be made illegal, as a matter of preservation of freedom.

#37

It’s not a bad OPTION, and the CONCEPT is good. If it can work the way it’s supposed to it’ll be a huge asset for people who don’t have the knowledge or time to devote to handling purchasing of a new PC suitable for gaming, upgrades, and keeping track of specs for games. Theoretically, with onlive even someone with an old netbook would be able to run the latest PC games at full specs.

BUT it still requires an incredibly solid net connection, and the more people who are using the service would mean more traffic that your data has to compete with. It also means that you never really own any of your games, since you only have access to them for as long as you pay the monthly fee.

If it were the only way to game it would be a failure, but it has the potential to be a good optional method.

#38

Onlive is a great concept for causal gamers. And by casual I mean part-time. Gamers who aren’t students but people who work for a living and don’t have the time to get £30 game play from a game, be that a one off payment or a WoW-type pay as you go game. It is also a good concept for families who run an average PC, mac, or laptop. They can all benefit from one decent internet conn, which is one monthly payment plus subscription. I know that whatever the cost of OnLive, it would save me money.

The cost of upgrading one server which supports 10 gamers at a time is considerably less than the cost of 30 gamers (see the part time bit) upgrading their personal computers, which is a large saving that can be passed down to the consumer. Not to mention the fact that you don’t need to pay full price per game to only play each for a couple of hours.

Unlimited Detail is a good idea, but not practical in the next 10 years (for example). OnLive will make the idea practical in the next 5 years (relatively), but make the impact much reduced because the consumer will not care about the difference it makes - it just won’t impact most consumers.

Also, OnLive is not a massive step from renting movies, or games for your PS or Xbox. You can go to your local blockbuster and rent games for consoles, but not for PCs. I appreciate that that is because they are easy to copy, but this makes it a non-issue - so why the problem for legal reasons?

Honestly, I think that the OnLive concept will massively reduce the impact of graphics advancements. Infact, let me rephrase that. I think that the OnLive concept will greatly increase the rate of graphics advancements because game devs won’t have to lower their standards to that of the average gamers setup. So better graphics means a better experience, right?

I can’t see any downsides for those with an internet conn that can cope, and at the rate the first world is developning that should be 50% of people within 2-3 years IMO.

I’m a little tipsy, so I’m rambling, but honestly can 95% of hardcore gamers say that their internet conn would not cope with OnLive? And would those people then say that they would rather pay £40 per 8 hours of new gameplay than £40 per month for all the games they fancy and at a higher quality that they could render on their home machine?

Seriously, UD is a cool idea, but is a lot more impractical than OL, yet OL gets the hostile reception? I don’t get it.

#39

I just physically couldn’t do OnLive. My internet is so awful it would be better for me to just play on my pc normally (and that is saying something).

#40

The savings ISN’T passed to the consumer. If a person pays a subscription for a service, invariably, eventually, the subscription will end up costing more than the actual product. Say you pay 5 dollars for every hour you play a game in Onlive…then after 10 or 12 hours, you’re paying more for the same game, and getting a horrible ping if playing online. That’s not even counting the subscription to use the service itself. If you play for a year, at $10 a month, you’ve got $120 you could spend on a mid-range GFX card that can run almost all of the games you’re subscribing to.

The only way OnLive is a benefit is for people who have amazing Internet, shitty graphics, and only have $30-$50 extra dollars a month to spend on gaming. The rest of the time, it’s redundant.

Founded in 2004, Leakfree.org became one of the first online communities dedicated to Valve’s Source engine development. It is more famously known for the formation of Black Mesa: Source under the 'Leakfree Modification Team' handle in September 2004.