Gamasutra: The Art & Business of Making Gamesspacer
View All     RSS
October 22, 2014
arrowPress Releases
October 22, 2014
PR Newswire
View All
View All     Submit Event





If you enjoy reading this site, you might also want to check out these UBM Tech sites:


 
Anatomy in games: An evolution of realism
by Sebastian Alvarado on 05/02/14 04:36:00 pm   Featured Blogs

The following blog post, unless otherwise noted, was written by a member of Gamasutra’s community.
The thoughts and opinions expressed are those of the writer and not Gamasutra or its parent company.

 

Technological advances in hardware and software have allowed video-games producers to create increasingly complex depictions of their characters. With respect to visual depictions of human anatomy in video-games, we have seen characters evolve from two- dimensional collections of pixels into highly-realistic three-dimensional creations capable of the finest facial expressions. To a biological scientist, this ‘evolution’ mimics what we observe in nature; a process of hundreds of thousands of minuscule changes, which result in increasingly more complex and sophisticated organisms.

Assuming this trend continues (alongside the necessary technological advances), we will one day see video-games characters that depict the human form perfectly. We’re not there yet, so it makes sense to question what lengths video-games producers should go to when striving for anatomical realism in their animations. Are certain forms or shapes more relevant to our human senses of recognition and empathy?

Character Recognition

Recognition of a character should not be considered problematic, providing that the recogniser is familiar with the character in the first instance. Studies have shown that a known face can be distinguished from a resolved photograph comprising only 50 pixels. The pixelated images below are of well-known faces that should be easily recognisable. I wont spoil the fun by identifying them here, I’m sure the solutions will appear in the comments.

Of course, recognition relies on a solid introduction and familiarity of a character, which is ideally physically or behaviourally distinctive in some way. It is clear that humans are quite skilled at recognising faces that we are all ready familiar with, but how do we identify something new as human, human-ish or something more dangerous?

Recognising Human Form

Although the processes of shape recognition are complex, phycologists believe that individuals subconsciously observe the relative sizes and locations of geometric shapes (such as spheres and cylinders) when recognising an object. The human form is most clearly recognised by its generally upright position with a (roughly) spherical head protruding superiorly and a central cylindrical trunk which bifurcates into two long cylinders inferiorly, which in turn communicate with the ground. Two long cylindrical appendages also attach to the superior portion of the trunk. More subtle shapes include broad curved shoulders, the flat broad chest and roundness of the gluteal muscles.

Although these descriptions seem comically simple, they (and their relative locations and sizes) are all it takes for our subconscious to identify a shape as human. Camouflage specialists in the elite divisions of our armed forces (e.g. SAS, Marines, Navy Seals) will use this process to their advantage by using camouflage creams and foliage to disrupt their silhouette and blend in with their surroundings.

Of course, the idea of “human” form can be stretched and these principles can be applied to fictional tetrapods, such as mutants, zombies, monsters and orcs.

Underlying anatomy of form recognition

The deepest structures that contribute to our form are our bones. It is important to note that bone is living tissue, and is constantly renewing and remodelling itself to accommodate whatever load it is subjected to. All muscles attach to these bones, and act to create movements across joints between the bones.

An incredible level of attention to detail was apparent in The Hobbit (2012), as New Zealand visual effects company Weta Digital crafted a multi-layered animation of the character Gollum. The process was so rooted in reality that experts in human anatomy will be able to identify each individual bones and muscles in the process, despite the final creation being non-human. The process, incidentally, is very similar to the complex method of forensic facial reconstruction, whereby specialised anatomists work to recreate a face from only a skull.

 In the case of The Hobbit, basing Gollum’s anatomy on an extremely realistic variation of the human form has produced an exceptionally realistic creature despite it being entirely fictional. The crucial point here is that accurate anatomy (human or otherwise) can be ‘tweaked’ to produce wholly original creatures that somehow feel totally realistic and almost familiar to us.

Muscle and bone are living tissues that adapt to load/strain/demand; this ‘form follows function’ relationship is fundamental in explaining the morphology (i.e. shape/structure) of muscle and bone. This relationship between function and morphology should be considered when ‘tweaking’ reality to creating fictional creatures. For example, in The Hobbit, Gollum was given a hunched vertebral column and larger eyes due to him inhabiting a dark cave.

Occasionally the ‘form follows function’ relationship is ignored in video-games. Take Hell Knight from the Doom series, for example. He is essentially human, apart from the horns, hooves and excessive musculature. Closer consideration of Hell Knight’s form suggests that he was developed after mythological depictions of The Beast of biblical end- times, or the devil himself, without consideration of how or why his anatomy is the way it is. This purely mythological depiction produces a character that, while quite disconcerting, is far from realistic.

The disconnection between anatomical form and function can be explained away within the narrative of the game, of course. For example, the super mutants of Fallout 3 were morphologically realistic, but had no functional basis. This was explained away by the introduction of a modified strain of a fictitious virus, which resulted in increased muscle mass. Somehow. Just don’t ask about the green skin.

With these points in mind, I would offer some general guidelines for character artists when designing creatures, fictional or not:

I.      Perfect the shape of your character before worrying about textures. Firstly, consider an accurate depiction of body shape; users will subconsciously rely on this when interacting with other characters, especially in a fast-paced action game. Secondly, bear in mind the relative dimensions and shapes of constituent structures, especially with regards to faces. Players will rely on this when recognising specific characters, which may assist in allowing the player to form a relationship with your character.

II.     More pixels don’t always result in a more convincing depiction of a character; concentrate on specific physical attributes which instil a sense of relevance in the character.

III.    When adding embellishments to your character, think like a functional anatomist and consider how your character behaves and interacts with their environment. This will not only result in a believable character, but will also give the impression that your character truly belongs in virtual environment your colleagues in world/level design have concocted.

Creating believable characters or creatures is surely the best way to evoke genuine emotion and reaction from an audience. Truly believable animations must go beyond simply mimicking the underlying structures of “real life” anatomy to consider the relationship between form and function. With believable characters or creations, an audience becomes open to genuine empathy with a virtual personality or story, and will undoubtably produce more immersive and enjoyable games.

Andrew O'Malley Is a lecturer in anatomy and human identification at the University of Dundee and consultant for Thwacke LLC. 

For more scientific insight in games visit Thwacke and follow us on Twitter, Google+ and Facebook


Related Jobs

University of Texas at Dallas
University of Texas at Dallas — Richardson, Texas, United States
[10.22.14]

Assistant/Associate Prof of Game Studies
Avalanche Studios
Avalanche Studios — New York, New York, United States
[10.22.14]

UI Artist/Designer
Bohemia Interactive Simulations
Bohemia Interactive Simulations — ORLANDO, Florida, United States
[10.22.14]

Game Designer
Petroglyph Games
Petroglyph Games — Las Vegas, Nevada, United States
[10.21.14]

Illustrator / Concept Artist






Comments


Alan Barton
profile image
Thanks a thought provoking read. The problem is essentially the Uncanny Valley. The concepts of the Uncanny Valley shows how people will react to a design and the problems designers will encounter the nearer we try to get to photorealism. Its also not just its physical form where we face these design problems, because its also even in the quality of the animation. Its also even in the quality of how the skin moves and reacts to movement. It even extends to the quality and form of lighting in the scene (not least of which the Global Illumination Radiosity problems) and how light reacts to the movement. (Subsurface lighting isn't even enough, as there's a lot of gory lighting details going on under the surface, e.g. veins and bones occluding some of the light and blood and fat altering the light etc.. etc..). If any of this looks wrong, it helps add to the feeling something is wrong and it doesn't take many things wrong to add up to the big feeling its wrong. Humans are very good at spoting "wrong" :)

Its a very interesting subject and a subject in one form or another I've been very interested in for the past 30 years (I'm a games programmer but I'm also into Robotics).

Michael G
profile image
1. Edward R. Murrow
2. Christopher Lloyd's great-grandfather
3. Jessica Alba
4. John Boehner
5. Tom from Myspace

Alan Barton
profile image
6. Michael G on Gamasutra?

... Its the most logical answer to your random list of people.

Ben Newbon
profile image
He's listing the people in the obscured 50 pixel pics.

Albeit incorrectly, I think:

1. Abraham Lincoln
2. William Shakespeare
3. Mona Lisa

Not sure on the last 2 though

Alex Popescu
profile image
I'd say
4. Mao
5. Mark Zuckerberg

(I feel wrong for knowing the last one)

Jennis Kartens
profile image
"5. Mark Zuckerberg

(I feel wrong for knowing the last one) "

Haha! Exactly my thoughts when first seeing that picture :D

Alan Barton
profile image
"He's listing the people in the obscured 50 pixel pics"

I see now :) ... I was having a dense no coffee moment. :)

Jennis Kartens
profile image
I actually see this as a problem, because it leads to the same pattern over and over again.

The human nature limitis itself, one could say. We need to compare everything, that is how we function. We are used to how our world functions with the constant physical environment that we live in and every other living form has to adapt to that environment too.
Weirdest things we know are deep sea creatures, living under different conditions without sunlight and under high pressure.

That however leads to unimaginative creations over and over again, especially when it comes to non-realistic themes and abstractions.

Aliens, fantasy figures and robots are too often designed very close to the human, for the sake of recognition or connection. Topped with some stereotypes in their behaviour and voices: perfect videogame character/enemy.

A breakout of these pattern is very difficult and just from the top of my mind, I cant think of a game where that has been achieved (I am sure there are games and movies too which went another direction). There are different designs found here and there, but overall the above is the rule.

Alan Barton
profile image
@"however leads to unimaginative creations over and over again"

Yeah films are the same, but then the films that are experimental, rarely do well financially.

Same with art and same with music and unfortunately same with games. I remember for example, levels in Quake 1 where people used water to make tunnels up into the air. So it looked like big columns of water in the air, totally impossible in real life and I loved the idea from a game play point of view (as you could jump out the water at any height you wanted and shoot through it etc... But it was bitterly put down as bad because it was unrealistic by many of the people playing and these levels were hated for it and later quake games never had anything like it as a result. I've seen this all too often with good fun features in a game. If its not reality, then some people can't cope with it.

Some (and unfortunately too many) people are all too often like the Grammar Nazis of reality. They have very ridge expectations of the way they want things and don't want to accept anything else. Trying to bend their expectations can be awful and very detrimental to the potential of your product as a result. So its no wonder many games take fewer risks and look like tired formulaic generic designs as a result.

Kevin Fishburne
profile image
You could make highly-recognizable hybrids. Take a human form, add two more arms, and rig it with a cockroach armature and animate it as such. People would probably freak out.

Chad Wagner
profile image
One thing I think is interesting is that when creature designs go completely alien - they all seem the same! All tentacles and spines or whatever. That's probably why you can't think any game that has done it - they all blend in to weirdness. Other.

We are FAR better at telling fine differences in things we know, and disturbing variations on what we know (intentionally uncanny).

I'm not sure how to solve this issue. Perhaps to create a new "species" and get players used to expectations in this new structure.

Jennis Kartens
profile image
Often with design I think maybe a scientific (evolution, or ideas of evolution under other physical circumstances as we have on earth) approach rather as an artistical (which ultimately is a psychological one too, since artists are often tought in regard of pleasing what we already know) would help.

A visit to the local terrarium or museum might help as well, but in respect of function and behaviour, rather then visual impressions of forms that then are again modeled onto the basic human head...

Also the stereotypes need to go... they're an additional burden that holds evolvement back.

David Cummins
profile image
While I agree that typical creature design has only progressed so far from the "person in a rubber suit" school of thought, the same techniques could be used to create much less humanoid creatures. If you focus on functional structures, the resulting creature will have a certain grounded realism that you don't necessarily get from a naive conversion from concept art to 3D.

Luis Guimaraes
profile image
"however leads to unimaginative creations over and over again"

This makes me immediately think of the StarShip Troopers movie. They still tweaked the designs of bugs and soldiers to fit the technology of the time, but still managed to do amazing effects for that time that still look awesome today.

Ben Newbon
profile image
Was just talking with some game dev friends about this very topic last night. With games like LA Noire and the likes over the past few years we've seen great advances in facial representation for video games, however we still see many games with frankly sub-par body representations, ranging from basic anatomic structures to complex movement. And then there's issues with super realistic faces being attached to poor depictions of heads and bodies and poor merging between them and even on top of that clothing (especially hats) that don't meld well with the character model as a whole.


none
 
Comment: