Are the graphical elements of gaming starting to peak?

So look at a good CGI now, it will be possble in games at 30+ fps in 2014.
Photorealism in 2014 then :D


Next things games really need to step up in quality is real time raytracing, some half decent simulation of radiosity (indirect illumination) and Sub Surface Scattering for more real organic textures.
 
Good luck trying to model photorealism. Several million polygons per square centrimetre. Ultra-high resolution textures. Ray-tracing everywhere. Physical and material behaviour of single objects and alloys. The list goes on.

We're never gonna see 'photorealism'. We'll be stuck in uncanny alley and forced to go for stylized graphics in the end.
 
KNives, whilst your correct in theory it doesn't really apply in theory. Top end games have already got to the point when additional polygons would never really show on anything other than a still rendered image. Lighting and Texturing are the areas that need to be improved along with a higher level of polish and attention to detail.
 
Hopefully in the future, all these so called "photorealistic" games will have more colours other than the millions of shades of browns and greys. Until then, I'll still be playing colourful "kiddie" games like Mario :D.
 
Surely, there has to be a point where by the games themselves have little or no more room for graphical improvement and I think we are starting to see this.
_

No, the next step is 3d displays. Is it zalman who has already released 2 3d displays, no idea what they are like, but I would love to see one in motion.
 
Nowhere near... look at something like robotics and see those natty little things at the cutting edge that can walk up stairs and say hello to you. Gaming graphics are as close to their peak as those things are to Star Trek's Data.
 
Good luck trying to model photorealism. Several million polygons per square centrimetre. Ultra-high resolution textures. Ray-tracing everywhere. Physical and material behaviour of single objects and alloys. The list goes on.

We're never gonna see 'photorealism'. We'll be stuck in uncanny alley and forced to go for stylized graphics in the end.

Agreed I don't think photorealism is ever going to be achievable, however CGI level graphics should be attainable within the next 10-15 years (Imo CGI is still quite easy to spot). Personally I don't think I would ever be comfortable playing a shooter which had photorealistic graphics. Anyone else feel if photorealistic graphics were possible it should be kept to the non-violent domain of gaming...say racing?
 
Mixed view on this as of course we all want the best contemporary gaming experience we can. If we still had graphics to the standard of the Amiga 15 years ago, it's doubtful gaming would have survived.

However I do share the concerns over the development issues. The higher and more technical the spec, the more specialised (and expensive) it becomes to find the artists and programmers to do the coding. People are still only going to pay the £30 or £40 (or whatever the going rate is at the time) for a new PC game.

In addition, raising the graphics bar can hinder homebrew development for games that support it. A minority scene I know, but I'm quite into train sims and the graphic detail required has jumped from the original MS Train Sim in 2001 to the new EA Rail Sim released in 2007 and likely to jump again later this year when MS release Train Sim 2.

Now for MSTS(1) I had access to reasonably cheap (Train Sim Modeller, 3D Canvas) or free (Gmax) 3D modelling packages which allowed the amateur builder to produce reasonable representation of required objects. However make it so that content can only be produced in expensive programmes and you kill off the homemade content that some games (not just train stuff) thrive on.
 
However make it so that content can only be produced in expensive programmes and you kill off the homemade content that some games (not just train stuff) thrive on.

Its interesting isnt it. a lot of innovative stuff you see nowadays comes not from the big corporations but from user-made content or some nobody nerd who's expressing their ideas over Youtube.
 
People say this every generation.

Graphics have an immense way to go before they become photo realistic and we will need to switch over to ray traced graphics before that can happen.

Ray tracing won't become feasable end efficient until 2020 imo, the way i see it is graphics will make drastic improvement but as we get better at producing higher quality games the time frame to create the games and make the more photo realistic games will mean longer waits.
 
Ray tracing won't become feasable end efficient until 2020 imo, the way i see it is graphics will make drastic improvement but as we get better at producing higher quality games the time frame to create the games and make the more photo realistic games will mean longer waits.

It will take forever to be feasible at good levels of detail, but my point is that until it happens we haven't reached the limit of current displays and we definitely aren't anywhere near reaching the graphical peak of games.
 
I think what people forget is that the vast amount of game engine development time goes into having it render as little detail as possible whilst still achieving the graphical level they desire.

Crysis for example, only assets immediately surrounding your POV are being rendered at highest details or rendered at all, just going a few ingame metres away from your char and all of a sudden all the litter such as cans / leaves start to vanish, shadows become much more simple, textures are dropped back.. go back even further and you end up with pretty much just a basic representation of the trees / major buildings being rendered and nothing else.

The only reason that game (or any other) is in anyway playable is because of this. So just to get Crysis running without that happening, also then bare in mind that compared to real life that sandbox environment of Crysis starts to feel like a claustrophobic corridor shooter, then I doubt you will see a computer this side of 2020 that would handle it at playable framerates... so to then want it to look real, all of a sudden a single blade of grass contains more polygons and data than that entire beach level combined? yeah, I don't think you should be wondering whether it will be this decade or the next, but rather this century or the next.
 
Guys, its all about creating the illusion of reality or photo-realism. Some of you are taking it literally, meaning things in real-life to be properly created would take millions, maybe billions of polys, milions of ray trace/radiosity bounces etc...That is terribly ineffiecient :) Even when doing 3D renders with a whole render farm of PCs to help you its still inefficient.

Clever lighting techniques, clever bump or normal mapping, parallex mapping, etc...these things just as an example get around that headache with great efficiency. Other techniques will pop up that do the same in the coming years. Software and hardware architecture is changing all the time. What seems impossible now will be laughed at for being so simplistic and yet incredibly inefficient in several years time...

Games are not much different to a render when it comes to optimising. When I create a 3D scene I do the bare minimum to get the image accross, just like a good game developer would employ clever streaming of data/distand LOD in an open world game.

Its all about the illusion....;)
 
Interesting philosophy Streeteh. You may have a point - even if we see technology and processing power continuing to improve, at the end of the day, the visuals we see have to be created by someone in the first place (in terms of textures, character models etc).

Personally, I still think there is a lot of room for improvement in terms of graphics. It's easy to find flaws even in the latest and greatest like FC2. Remember too, that even if you have incredible design and artwork, all that can be ruined by something as simple as an engine which doesn't handle clipping properly, so you have bodies stuck inside rocks or something stupid.

My general view on graphics in gaming over the past decade is this:

-Pre-2000, most games look very dated and it has a significant negative impact on my enjoyment of the game
-From 2000-2003, most games look good enough to play and enjoy - things like Deus Ex, NOLF2, CoD, NFS5/6/7, Chrome etc while not impressive by today's standards, at least portray the game world in a good enough light to be able to play without worrying too much about it.
-In 2004 there was a big improvement, with FC, D3 and HL2 all being released. This set the benchmark for a few years and IMO those engines still look good.
-The release of Crysis a little over a year ago shifted the goalposts again, and that's the kind of level most developers are/should be aiming at for titles being released today.
 
Yep, waiting for Raytracing. It shouldn't be long with the way Intel is pushing it.

Take a look at the Beowulf film for an example of what I expect games to peak at.
 
Last edited:
Back
Top Bottom