|
We've all heard it before; since the PS3 and Xbox 360 generation,
our games' color palettes have moved towards desaturated tones. I'll
try to explain why this has happened, and focus on one of the less
obvious reasons.
Since textures are now of higher resolution, dirty surfaces such as
rusty metal, rocks, muddy grounds, damaged concrete, etc., can look
pretty good. On top of that, using specular highlights implies metallic
or wet materials.
Dynamic lighting coupled with normal maps leads us to
make environments where surfaces are not flat; we're more likely to
make damaged or rocky surfaces to get that extra detail in our
environments now that our shaders allow us to, and metallic surfaces to
make specular highlights and normal maps more apparent as the lighting
moves over the surfaces. So by default, the new tech leads us in a
certain direction. We could make some nice looking clean world, but it
would imply new challenges to overcome.
Imagine you were to look at a painting of a person. You know it's
not real, hence there are various errors that might subconsciously
bother you, even though you wouldn't have noticed them if those same
apparent errors were edited into a photo. You might not be able to
pinpoint what's wrong with it, but instinctively your mind noticed
something wasn't right.
Video game worlds are by their very nature
artificial. There's all sorts of factors that we would normally not be
bothered by that will simply feel wrong when seen in a video game. A
very clean hallway will look unfinished and a perfectly straight edge
will look like it lacks detail.
So we can tell that already, the artistic direction has been
influenced by the development of the tech we can now use, and that the
video game medium makes it easier for the viewer to doubt what he sees.
But why desaturated colors? There is one thing that our current consoles
are terrible at; lighting. Our current lighting solutions are
improving, but for the moment we have much difficulty simulating
indirect lighting, especially in real-time. In the previous generation,
graphical quality was not high enough for us to be bothered by the lack
of indirect illumination in our saturated environments, but once again,
as graphical quality rises, so does our expectations of how the world
should be presented. Just as wonky animations will shatter immersion,
so will poor lighting.
To hide this problem, we tend to instinctively desaturate
everything. The mere presence of saturated colors unbalances the rest
of the image. Since we often have some form of ambient occlusion in our
environments, this visual effect makes the game look more visually
convincing. The lack of indirect illumination, or more specifically the
lack of radiosity, brings this level of believability off balance.
Here is an image that illustrates the problem:
The
top image doesn't look bad, ambient occlusion (the dark edges around
the areas where the different surfaces are close to one another) works
well to add quality to the image. But the lack of radiosity doesn't
feel right. Imagine if this scene was actually a colorful sunlit living
room in a penthouse. The lack of bouncing colors would really cheapen
the quality of the image.
Here's the same image as above, but in black and white:
Now
that the radiosity can't really be perceived, the visual quality doesn't go
off balance.
The game Mirror's Edge used some nifty tech to simulate
indirect lighting, which was really vital to the game's visual quality.
It simply could not have been set in a clean white city with brightly
saturated surfaces if it wasn't for this tech; it would have made the game look cheap,
fake, and not immersive at all.
Gears of War
on the other hand had an artistic direction developed around the idea
of using the new graphical developments to their full extent, so Epic
went for dark environments with bumpy rocks and dirty metallic surfaces
which would allow them to light the scenes up with multiple dynamic
lights, allowing them to get the most out of their normal maps and
specular highlights.
Uncharted 2: Among Thieves has an even more saturated palette than Uncharted: Drake's Fortune
because their lighting solution has vastly improved since and can be
showcased rather well in colorful environments.
They also use saturated
colors to make certain objects stand out so as to guide the player
throughout the level, it's subtle but it works well. Uncharted 2
will probably be even more of a trend setter than Mirror's Edge since
it manages to pull off the gritty look while still using a unique color
palette. It really allows the game to set itself apart from the
competition.
As
our lighting solutions unify and become more dynamic-oriented, we can
expect the next-generation games to have a much wider variety of
color palettes as real-time translucency and indirect illumination
become easily achievable. Expect saturated colors to be the new brown.
[Originally posted at http://www.allegory-of-the-game.com/]
|
That is a good point. I think when it comes down to it, either in the case of figurative art or realistically rendered video games, it's a question of intentions. The one who makes figurative art might make it for a reason; maybe it is what the client is asking, maybe it is for his own personal satisfaction, etc. The same can apply to video games; if realistically rendered games are an easier sell, something that appears to be true, then we will continue to make them. Unlike in the case of paintings and photography, video games are first and foremost a product, so this is what will lead us in a certain direction over another.
We can make games that are not necessarily realistic. It's simply a question of threshold; how much margin you have before you produce a game that goes against the market. Uncharted is a good example of a game that struck a good balance I think, but this can change from one type of game to another, or as time goes by based on the current sentiments of the market (too many games looking similar, etc.).
As per the article, the key problem is lighting, though I think it's not so much about technology. Put simply: real life is generally a bit dull and dirty looking! The closer we get to simulating reality, the worse the problem.
Back to the 256 shades of brown, a game released around the same time as Quake (around 6 months later) took a different approach. Shiny Entertainment deliberately decided to focus on good texture mapping, a good draw distance and a high poly-count: things like dynamic lighting and gouraud shading were left out.
Combined with some sharp level design, this made MDK stand head and shoulders above the competition - it still looks fairly good today (and runs fine on both Vista and XP, thanks to Shiny releasing a Direct X patch). It may not be realistic (and it doesn't lend itself to the sandbox-world environments being created today), but it's a useful reminder that technology needn't be the limiting factor.
Admittedly, I may be slightly biased - I've dumped a couple of videos of MDK gameplay onto Youtube:
http://www.youtube.com/watch?v=IXxSet2PKM0&feature=channel_page
It's worth taking a look though:)
In my view, the English market (particularly the development side) tends to have a focus on photorealism rather than approaching visual art from a broad perspective and offering the consumers a wide variety of styles and presentations. Photorealism is not needed for a fun game, so there's no reason to limit development with such a focus. I hope this trend will reverse itself; after all, there's a good reason why various games that are not photorealistic are very popular with consumers while some games that are very photorealistic wind up being generally panned.
It seems there is currently a problem with my host, so the images are not displaying right now. I'm guessing it will be resolved soon.
Also, I sort of prefer the top image; perhaps I've been conditioned by current game graphics, or perhaps vibrant saturated well-lit objects aren't something that I'm used to seeing in nature, but all the colour-bounce and lack of nice crisp dark shadows in the bottom image make it look quite surreal.
Another obstacle/technical impediment currently is the distortion caused by linear perspective projection, where rotating the camera makes things look "wonky". This can be reduced with the right FOV selection but it's still present. Probably I'm not explaining this very well; I don't know if curvilinear projection would fix this problem, but _something_ is not right with the way things currently get all swimmy around the edges of the screen when rotating the camera around the vertical axis..
Actually, I think both of you are reinforcing Eric's point even if you meant to disagree (and Alex's too, for that matter) because Philippe is offering a possible explanation as to why current (mostly English market) games focus on certain visual presentation choices, but underneath Philippe's explanation is the assumption that developers need to focus on attempting to use photorealistic art. Everything in Philippe's discussion revolves around why things look "off" but his points are only applicable to photorealistic visual styles.
What is "interesting" for one person is not "interesting" for someone else. However, the current development trend is to focus on photorealism as "necessary" or "the only approach" to visual art, and the assumption is that this view is the only one that consumers find "interesting." This is a false assumption, of course, but the trend continues (with all its associated problems, some of which Philippe discusses here). This assumption has been largely in the English development community but Japanese and other East Asian companies are beginning to focus on "making games that appeal to Western markets" (according to related articles and interviews here on Gamasutra and other reliable news sites). This is not a positive direction for the industry; it undermines diverse tastes in visual styles. The same is true for other aspects such as writing, of course.
Personally, I have argued that a game that is appealing due to story and character development, as well as visual style, causes the player to put up with gameplay that may be inferior. Obviously, we,d prefer to have excellent gameplay, too, but many players will tolerate gameplay flaws if they feel emotional attachment to the characters and events in the game. They want to see what happens, so they keep playing.
Alex's point about costs is important, though. The focus on photorealism happened when the graphic card makers caused too much influence on the industry. In addition, it's much easier to market "cool, flashy graphics" than "innovative gameplay techniques." The former can be seen, but the latter must be experienced through actual play sessions. The industry has been choosing the easier selection by focusing on graphics for their cost expenditures, but this isn't necessarily the best choice for longterm benefits for a game product. It's easier to justify and market, though, so this direction probably won't change anytime soon, at least for larger publishers.
The best hope for more creative and innovative approaches is probably independent, small developers. Of course, I tend to feel that Japan has a very strong, vibrant doujinsoft community, so I think they have a big advantage in this area. At the same time, there are more exchanges of information and styles worldwide than there used to be, and I've seen influences from East Asia making their way into the independent art community in America and other Western markets. I'm hopeful that this will help counterbalance the trends of the past 10-15 years in the English market.