On the Subject of Graphics

Kelly Brook

I was unfortunate enough to stumble across this painful example of "journalism" on the otherwise quite respectable bit-tech.net. It's (thankfully) short, but in order to save you from wasting even the smallest amount of time reading what is in reality a bunch of ignorant, idiotic wank, let me distill some of the "best" bits.

"Crysis with all the sauce turned on looks just as good as Left 4 Dead"

What a dick. Define "good". L4D has graphics which are appropriate given what it's trying to achieve. Heaving hordes of zombies are more important than rich environments. Indeed, the sparsity and starkness of L4D lends to it's post-apocalyptic feel. Crysis, on the other hand, attempts to render a rich, lush tropical environment with a much smaller density of NPCs. Similarly, WoW is quite openly designed to run on as many machines as possible, and looks very basic as a result. Do the X million WoW players care? Do they fuck. L4D would arguably not benefit much from running on CryEngine 2. Crysis, on the other hand, would be impossible to build on Source. The fact that someone compares the two demonstrates a tragic ignorance of graphical effects.

"how good a game looks isn’t about how technical the graphics are"

This is really the same point, and again relies on the tiresomely ambiguous use of the word "good". How "realistic" a game looks IS dependent on the lighting pipeline supported by the engine. How "good" it looks is a meaningless, subjective metric. This is just another example of people not understanding the difference between graphical complexity and art design, or at the very least not appreciating that a more sophisticated engine can only increase the options available to designers, while no amount of clever design will achieve realism beyond the capabilities of the engine technology, even if it can disguise its shortcomings.

"I really hope that graphics advancements slow down some and gives the hardware a chance to catch up"

What a fucking retard. How about hardware advancements speed up to enable better graphical effects? Let's put the brakes on progress just so you fucking morons who can't afford a decent machine can run "high end" games on your piece of shit PCs. These are the same sort of people who still claim Crysis is "poorly coded" to this day, despite a) having no experience of graphics programming, and b) the fact that nothing has been released since Crysis that is anywhere near as visually complex.
Anyone with a clue knows that video card technology has been sadly stagnant for the last couple of years, while nvidia sat on their thumbs waiting for amd to catch up. The 9x00 series in particular didn't deserve to see the light of day, and nvidia are still reguritating three year-old tech in the form of the GTX250. Let's give them the benefit of the doubt and blame the global economic fiasco. The fact is still that a combination of weak tech and the increasing trend towards porting games from aging console hardware to the PC are holding back much-needed advances in graphics.

If Crysis is indeed the pinnacle of PC graphics then in a few years or maybe less than that, we’ll be able play our games at the native resolution of a 30in display with all settings on max using mid-range hardware

This is such a meaningless statement. There are plenty of games right now which will run perfectly well at 2560x1600. Some of them even look good. This quote borders dangerously on a "we will have photorealism in X years" prediction. Crysis is a triumph of graphics programming on current hardware, but it is a long way from photoreal. And it's not just about being able to push enough pixels to be able to run it at a high resolution. I'm sure even Crytek's own developers would be happy to point out the many approximations and cheats they had to employ in order to squeeze every last frame-per-second out of the engine, at the expense of realism. Did you really mistake the Koreans in Crysis for real people? Or even one of the trees and bushes for real trees and bushes? The graphics might be "sufficient", they might be "impressive", but they're by no means photoreal, in the same way that you would never in a million years confuse that Final Fantasy film with live action. This quote is just more whining from someone who can't afford a decent PC and who thinks the whole games industry should just spin it's wheels churning out second-rate, 5 year-old visuals just so that they run well on his shit PC.

Of course the author is simply trying to align his own worthless opinion with that of industry luminary Doug Lombardi in order to look cool and connected. Unfortunately he's also misunderstanding the issue. There are other factors involved when it comes to game graphics, most importantly cost. Let's say Crysis is 80% realistic. I just made that up, but the point is that the last 20% of realism would cost a disproportionate amount to achieve, even ignoring the fact that it would be nowhere near possible on existing hardware. The sheer amount of data that would be required would demand so many modellers, texture painters, animators and level designers that it's simply not realistic to expect genuinely photoreal games. Consider the enormous budgets consumed by current AAA titles, and then look at what they achieve technically. Yes, there are the constraints of the hardware to consider, but bear in mind that modern games also seem to be shrinking in terms of hours of game play.

Another issue that sometimes pops up during discussions of game graphics is that of ray tracing. It is currently being waved around as some sort of magical technology that will single-handedly make all games photoreal, if we can only get hardware that can process it in real time.
Unfortunately this is simply not true. The fact is, there are some effects which can be achieved more effectively, or more accurately, by ray tracing, but it also comes with a lot of it's own limitations, and not just when it comes to speed. Even if the hardware was available, there would be no reason to have an engine ray trace everything all the time, it's just not necessary. It's not really the simple choice of "ray tracing vs. rasterisation" which is often described. Visual effects artists in film or television don't ray trace everything even though they don't suffer from the limitation of having to render a frame in 1/60th of a second. Because it doesn't add anything to the visual quality in a lot of cases. Of course if you're dealing with pretty, glossy reflections in complex surfaces there are few alternatives. Not to mention other applications like shadowing, occlusion and colour-bleeding, although in those cases there are alternatives, and alternatives which often produce superior results. The point is it's not an "either, or" situation, and while it would be nice to have access to real-time ray tracing functionality, don't expect it to suddenly make all your games photoreal. Ray tracing won't improve your models, your textures or your animation.