You'll notice that I haven't spent much time updating, recently, and there's actually a really good reason for that: it's too damned hot.
The room that has my workstation isn't air conditioned. I'm not worried about overheating my computer, mind you, I'm worried about the river of sweat that starts at my neck and flows like the mighty Mississip down to my butt crack. So, to avoid looking like the foil in a deodorant commercial, I spend most of my time downstairs in the air conditioning on my laptop. My laptop is a super-portable with naught but a 13" screen; not exactly an epic amount of screen real estate on which to get graphical work done.
Hopefully, with summer nearing an end, I'll be able to spend more time upstairs grinding out lovely images of cartoon breasts. Because, seriously, what else is there?
Sunday, August 22, 2010
Tuesday, August 3, 2010
Of Figures
Since I began my work on cartoon figures in earnest about a year ago, I've become well acquainted with the lines that make up the human form. Cartoons, especially some of the more simplistic ones, are essentially distillations of the human form down to a set of gross lines. This sort of perspective gives you an interesting take on people, and it also helps me to understand the world of fashion and the uncanny valley that they seem to have hit.
Put briefly, the uncanny valley is the "region" on a curve representing the realism of a human form. On one extreme is a simple sphere, and on the other extreme is a perfect photograph of a person. How realistic the image appears climbs as you move away from the simple sphere, by adding eyes, a mouth, and ever-more details.
Oddly, though, once you hit a certain amount of details, the realism breaks down and the human eyes ability to "see" it as a person plunges. This is why full CGI human faces are a sort of Holy Grail for movies, and why Benjamin Button was such a breakthrough.
I think that the crossover happens when the brain switches from interpolation to sensory-acceptance. What I mean by that is when the brain stops filling in gaps with its pattern recognition "software" and instead accepts what's coming in via the eyes without alteration. We're incredibly good at finding patterns, even faces, so things that are roughly human have just enough detail to trigger our brain's pattern searching mechanism, which happily fills in the blanks. With this, we can see things in clouds, make abstract art, and even see Mary in a cheese sandwich. But after enough details are found, the brain stops filling the gaps and hands it over to the holistic processing part of the brain, that gels all of the perceived details into a whole figure for storage in memory. People who are unable to see the gestalt of the human face are suffering from what's called Prosopagnosia, or face-blindness. I assume that these people would be unable to experience the uncanny valley.
In the same way that creating ersatz human forms is a bottom-up approach, it makes sense that we can also take a top-down perspective, where we start with entirely realistic photographs and then push them down into the uncanny valley on a search to exaggerate positive characteristics.
Advertisers have always used airbrushing and illustrations to create idealized forms. But back then, our ability to manipulate photographs was limited, and illustrations were created by hand, which meant that if one characteristic was pushed too far, the artist could easily compensate. But with an illustration, 100% of the output was from the human perspective. With a photograph, 100% is already there and the artist is instead trying to draw out exaggerations.
Since I spend all of my time working with very broad lines, I know exactly what they're doing when they totally screw up images of women. They're attempting to use the same lines that work so well in cartoons but fail miserably when applied to pictures. In a cartoon, you can extend or squash lines all you want and the form can remain consistent, as long as you manipulate the rest of the form accordingly. They've been moving towards this for some time. Making legs longer, waists thinner, lips plumper, etc. Only recently, I think, have they effectively hit the limits of what they're able to do. The watershed moment came with the infamous Ralph Lauren photo.
I'm sure that you've seen it at least once. It hit every major news site and has been a source of constant discussion for the last eight months. The model's figure is so grotesquely malformed that she has entered the uncanny valley. She no longer looks human. But the principles of the image manipulation are identical to earlier efforts, and her form's lines are similar to many cartoons. In fact, if we remove the details and break her down to nothing more than a silhouette, it begins looking alright again.
We have another example of photoshopping gone awry with Ann Taylor, who received their fair share of flak for some editing awhile back.
Again, look at what they've done. They're tried to reduce a complex human figure down to gross lines. They removed detail and exaggerated other details. They've pushed the figure towards the uncanny valley. It's not all the way there, but the website's little screw-up that reveals the original photos illustrates how even mild, unskilled rejiggering makes the figure appear off. All of this because they're trying to reduce real women down to cartoon-like lines.
What the industry may find is that, going forward, they will either have to accept that this is as far as they can go with manipulation of the female form, or they might want to start hiring cartoon character designers and animators to take over Photoshop duty, because, man, this shit is just getting weird.
Put briefly, the uncanny valley is the "region" on a curve representing the realism of a human form. On one extreme is a simple sphere, and on the other extreme is a perfect photograph of a person. How realistic the image appears climbs as you move away from the simple sphere, by adding eyes, a mouth, and ever-more details.
Oddly, though, once you hit a certain amount of details, the realism breaks down and the human eyes ability to "see" it as a person plunges. This is why full CGI human faces are a sort of Holy Grail for movies, and why Benjamin Button was such a breakthrough.
I think that the crossover happens when the brain switches from interpolation to sensory-acceptance. What I mean by that is when the brain stops filling in gaps with its pattern recognition "software" and instead accepts what's coming in via the eyes without alteration. We're incredibly good at finding patterns, even faces, so things that are roughly human have just enough detail to trigger our brain's pattern searching mechanism, which happily fills in the blanks. With this, we can see things in clouds, make abstract art, and even see Mary in a cheese sandwich. But after enough details are found, the brain stops filling the gaps and hands it over to the holistic processing part of the brain, that gels all of the perceived details into a whole figure for storage in memory. People who are unable to see the gestalt of the human face are suffering from what's called Prosopagnosia, or face-blindness. I assume that these people would be unable to experience the uncanny valley.
In the same way that creating ersatz human forms is a bottom-up approach, it makes sense that we can also take a top-down perspective, where we start with entirely realistic photographs and then push them down into the uncanny valley on a search to exaggerate positive characteristics.
Advertisers have always used airbrushing and illustrations to create idealized forms. But back then, our ability to manipulate photographs was limited, and illustrations were created by hand, which meant that if one characteristic was pushed too far, the artist could easily compensate. But with an illustration, 100% of the output was from the human perspective. With a photograph, 100% is already there and the artist is instead trying to draw out exaggerations.
Since I spend all of my time working with very broad lines, I know exactly what they're doing when they totally screw up images of women. They're attempting to use the same lines that work so well in cartoons but fail miserably when applied to pictures. In a cartoon, you can extend or squash lines all you want and the form can remain consistent, as long as you manipulate the rest of the form accordingly. They've been moving towards this for some time. Making legs longer, waists thinner, lips plumper, etc. Only recently, I think, have they effectively hit the limits of what they're able to do. The watershed moment came with the infamous Ralph Lauren photo.
I'm sure that you've seen it at least once. It hit every major news site and has been a source of constant discussion for the last eight months. The model's figure is so grotesquely malformed that she has entered the uncanny valley. She no longer looks human. But the principles of the image manipulation are identical to earlier efforts, and her form's lines are similar to many cartoons. In fact, if we remove the details and break her down to nothing more than a silhouette, it begins looking alright again.
We have another example of photoshopping gone awry with Ann Taylor, who received their fair share of flak for some editing awhile back.
Again, look at what they've done. They're tried to reduce a complex human figure down to gross lines. They removed detail and exaggerated other details. They've pushed the figure towards the uncanny valley. It's not all the way there, but the website's little screw-up that reveals the original photos illustrates how even mild, unskilled rejiggering makes the figure appear off. All of this because they're trying to reduce real women down to cartoon-like lines.
What the industry may find is that, going forward, they will either have to accept that this is as far as they can go with manipulation of the female form, or they might want to start hiring cartoon character designers and animators to take over Photoshop duty, because, man, this shit is just getting weird.