Grahpics

Well, I am sitting down the other day reading my recent E3, and I started to wonder.."What will games do now with graphics?" I mean, I dont think its physically possible to make better grahpics than Doom 3, Half-Life2, or Far Cry. What is the next keen feature of developers to focus on? Perfect physics (Ala: HL2), Perfect Gameplay, Or just small details? Or will people start investing into virtual reality, I mean, actually GOOD virtual reality (although I heard its been done, yet its quite expensive).

Well for a start, even the next-generation games are not photo-realistic. It is certainly possible to improve graphics - just watch the movie Final Fantasy and then tell me that game graphics are perfect :razz:

Anyway, I hope the focus will return to gameplay. Eventually it won't be possible to impress people with graphics alone (I think that Painkiller demo is a good example).

Vash, when was the last time you went outside? doom3 and HL2 are nowhere near realistic in terms of lighting. if you ask me the ultimate destination in graphics is in physics simulations. you don't seem to realize that physics has more to it than bullets and biomechanics.

it's probably a few steps away, but I expect you will be able to set a given materials absorbtivity, flouresnece, dielectric constant, roughness, microstructure ect.. all things which effect how an object interacts with light.

Also, much work still needs to be done on creating realistic outdoor environments. to be honest, far cry did not impress me much.

At the moment, everything looks very wierd and plastic like. especialy in doom 3.

I agree with TB. Real physics is the next step. Anyone played de_ffraction? :biggrin:

Real-time light rendering would be good, but we dont have good enough puters for that.

or how about real time lighting . . . ? obviously that would take an immense amount of processing power (or a faily fast pipelined system) but it would greatly add to the interactivity. But over all, Tracer is right.

I don't think you could perfectly simulate the real word without some sort of encapsulation of your self.
I mean, yeah Far Cry looked good, I mean be honest it did, even if it looked like plastic, it looked like good plastic lol.

I personally think the problem is physical movements. They've got environments pretty much down (not visually, just physics-wise, such as cables and ropes). The way a person moves seems just too far out of reach. I had an idea awhile ago for a plug-in for a modeling program (any modeling program really) such that when the model is compiled into a model and not in it's editable state if the joint coordinates were actually sent to the game, the game could catch them or not, it's up to it, but if it did, you could practically have real time human reactions, as in, you'd move a certain way if you got hit with a bullet and so on, it would also be rather hard to replicate the same series of movements twice (so animations aren't set in stone).

Anyway, it was just a thought I had.
I'm sure my entire post looks like rambling in heinsight.

I have often asked myself what the gap is between game/3D Animation graphics is and the visual aspects of real life. Then again, we've only been in this field for a maximum of what, 20-25 years, and it took painters a couple thousand to get their work looking 'real'.

Of course the scale and speed of computing power is exponentially larger than painting. They had the same tools and people died in the processes of learning. They never really got "faster and more powerful" painting tools. A few advancements but nothing near what we're capable with computers. And seeing as how our societies like to waste money, I'm sure it won't be that long until all of our resources end up creating games with true-to-life physics, lighting, control etc.

http://collective.valve-erc.com/index.php?go=photorealism

Apparently photorealism isn't going to happen too soon.

I could have told you that, Loco.

I mean seriously, think about everything you'd need. It wouldn't just be light acting in a right way. It would be the actual composition of an object. Because if you just deal with light you're ignoring some important factors, like tempurature, in warm conditions light is refracted through heat. There's also water, water's refraction is pretty random, in regards to you trying to determine how it works.

The only way a game would be photorealistic (which I personally wouldn't want on the models of people, it would be disturbing if it were too realistic) is if we actually programmed at an atomic level. And by doing that you'd in advertantly be creating light and world physics.

But, we're not going to be able to do that very soon (obviously) so.

Also, there's one small thing that articles author forgot to think about...64-bit pipelined machines. Piplelining is a step in the right direction for real time computations (because several things are calculated at once instead of just one thing).

But, we're not going to see most of this stuff in our lifetime, but we'll see damn good simulations :biggrin:

Vash said:
Well, I am sitting down the other day reading my recent E3, and I started to wonder.."What will games do now with graphics?" I mean, I dont think its physically possible to make better grahpics than Doom 3, Half-Life2, or Far Cry. What is the next keen feature of developers to focus on? Perfect physics (Ala: HL2), Perfect Gameplay, Or just small details? Or will people start investing into virtual reality, I mean, actually GOOD virtual reality (although I heard its been done, yet its quite expensive).

Thats what people said about doom2.Graphics WILL advance untill they look 100% real like.

I just want game developers to focus more on gameplay ... people will stop buying those top-notch systems to play a game that looks awesome, but plays like a bunch of crap.

It's all about polycount ... but where's the love ?

Loco said:
http://collective.valve-erc.com/index.php?go=photorealism

Apparently photorealism isn't going to happen too soon.

lol I love that 'number line' that tries to tell us something or other- what's it got to do with large numbers? Where's the origin, what's the scale? If it's a 'normal' scale then 0 and a billion should be in practically the same place as a million, seeing as it's 1/1000th the length of that line. If it's a logarithmic scale, it should be halfway.

I really shouldn't have posted that!

The point indicated in the 'answer' appears to be just under 30 billion. I think the point of it is to show us all how far a trillion is from a billion in relation to how far a billion is to a million, just in case we didn't know. I've read this artcile before and thought it was a complete waste of time - it doesn't say anything, setting out no aims, and reaching no conclusions.

Next step will be refining of performance, so that all these new-fangled engines are less machine intense - allowing for more to be going on in the background i.e physics etc.