Uncanny Valley/Cuteness in relation to pixel-art

[Carnivac]

The shape is obviously in the style of the cartoon, but it looks believable. Physics-wise, it's very believable. Your brain doesn't say "Whoa!! WTF is that FREAK!!??" due to the (should-be obvious) deformities, but instead it says "Nice!"

Hm, no.  My brain's definitely not saying anything nice about it.  Respect the work gone into creating such an image but it sorta repulses me somewhat.    The features work fine for a cartoon (though I've never found her 'attractive' as such) but with the realistic rendering it makes me ever so slighly nauseous.

[||||]

I completely agree.. That Jessica Rabbit is CREEPY AS HELL 

This thread is really interesting though. I know I've delved into the valley of uncanny myself and it's a thin valley.


I wish so many people didn't like cute so much.. I mean sure cute is cute and I don't dislike it too much;  but I think the reason people like cute so much is because they like 'simple' or what I'd call 'accessible'. And I think that 'simple' is derived from laziness and lack of cultural aesthetic. Cultural corporatism dominates and is a big part of 'simple' designs taking over I fear. Logos are the new norm, even the fonts we use today are super simplified; Helvetica's taking over for example (a lot of corporate fonts are just minor tweaks of this standard).

Consumer demand is infantile and it demands simplicity.

[rikfuzz]

I completely agree.. That Jessica Rabbit is CREEPY AS HELL 

This thread is really interesting though. I know I've delved into the valley of uncanny myself and it's a thin valley.


I wish so many people didn't like cute so much.. I mean sure cute is cute and I don't dislike it too much;  but I think the reason people like cute so much is because they like 'simple' or what I'd call 'accessible'. And I think that 'simple' is derived from laziness and lack of cultural aesthetic. Cultural corporatism dominates and is a big part of 'simple' designs taking over I fear. Logos are the new norm, even the fonts we use today are super simplified; Helvetica's taking over for example (a lot of corporate fonts are just minor tweaks of this standard).

Consumer demand is infantile and it demands simplicity.

Simplicity is not laziness.  Helvetica is not a modern threat, it has been popular for over half a century.  Simple designs aren't 'taking over'. (There was a trend for over-rendered logos in the 90s / early 2000s which is finally dying down if that's what you mean). 

edit:  Century not decade.  (Haha). 

[Cyangmou]

The shape is obviously in the style of the cartoon, but it looks believable. Physics-wise, it's very believable. Your brain doesn't say "Whoa!! WTF is that FREAK!!??" due to the (should-be obvious) deformities, but instead it says "Nice!"

Hm, no.  My brain's definitely not saying anything nice about it.  Respect the work gone into creating such an image but it sorta repulses me somewhat.    The features work fine for a cartoon (though I've never found her 'attractive' as such) but with the realistic rendering it makes me ever so slighly nauseous.

Well for me it also doesn't work. Mainly because the jawbone is really strong and the cranial area is really short. Maybe just a tad compared to the cartoon, or the value depths are off, but face wise, even to be just slightly off can change a lot.

Simplicity is not laziness.

Well said there.
Simplicity can be very appealing, as everything if it's done on a level of good quality.
And as artists we have to simplify anyways a lot, even if it's just for understanding and all details come in later.

What's simplicity anyways?
Can we say that there is nothing unnecessary in the design, which would distract from the purpose?
Or do we rather define simplicity a sa low amount of overall forms for an object?

Even a full illustrated cute piece of art can be more complex than a simple realistic gesture drawing, so simplicity and cuteness/realism are 2 completely different things - at least for me.

Could we say that cuteness is on an intermediate artistical level of quality easier to pull of than realism on an intermediate artistical level of quality? And because of that cute is more widely spread?

If I look at popular realistic-styled productions they are form an artistical standpoint all really expensive compared to not so realistic productions.

 

[||||]

I didn't say it was flat out laziness.. I meant that it conforms to a standard based on easy access and convenience, which is semi related. The demand behind it is more lazy than those producing it. For the last near decade for example.. a lot of logos and fonts look like they belong in an app.. all glossy and button like. It looks very accessible and 'easy'. 

You're probably right about Helvetica being popular for so long now.. but It seems to me a lot more brands use this basic sans serif font as a base than before.

About the over rendered logos of the 90's/ early 00's.. I can't think of any unless we enter the category of heavy metal band logos. That's over rendered/detailed galore.

I don't want to go too off topic about fonts and logos here though.. cause it's only partially related to the topic.

[Pix3M]

I've been trying to look into this deeper, knowing that I basically set up a false dilemma to simplify the situation. There's cute, then there's realistic, and there's a fair number of things in-between. For the sake of argument, I'm going to set up a sliding scale from cute to realistic.

*Pardon if I come off as bullshitting. These are just my obervations while I'm trying my best to look into this*

Searching through deviantART for what is popular there, I found some things that actually challanges my thinking.

Tier 1, I will stick anything with chibi-proportions into this tier. This is a 'pony', but stylized to an extreme that it doesn't look anything like the original animal: http://fav.me/d7q8v9e However, there is not very many pieces I can stick on to this extreme
Tier 2, which I'll stick in your usual 'anime' style. The face is stylized to exaggerate youthful features, but with *relatively* realistic body proportions: http://fav.me/d7qey6q
Tier 3, moving closer towards realism, I'm a little confused here, but in this tier, I and will stick both this Japanese style http://fav.me/d7obm2l , and this western style in the same tier: http://fav.me/d7rym4n
Tier 4, i'm going to stick here, something I'd call 'real' but cautious knowing that proportions are still highly stylized: http://fav.me/d7qvjg1
Tier 5, is what I'll stick as a reasonable extreme for a realist painter, but an actual painting took a while to find among photomanipulations: http://fav.me/d7qhwkf

First, a word of caution that my experiences on deviantART will put a clear bias towards more realist styles since it's full of artists - people are more likely to appreciate the work put into a realistic style. Again, I found these mostly from what's popular on deviantART in the past month, but what I was surprised was I found a lack of art on the cute extreme, but also on the real extreme as well.

From experience, art styles leaning on the extreme end for cute runs into several issues.

• It is more difficult to create something memorable with more extreme cute art. My sister told me of a moe-style artist who got away with tracing for so many years, and I was not surprised. The face style they traced were done in a popular (but IMO highly generic) moe style, allowing the tracing artist to get away with art theft for years.
• Cuter art styles are arguably pretty over-saturated as a result too, if said artist got away with tracing for THAT long. Might also be why the only 'extreme-cute' art happens to me MLP fan art, and not those little chibi dolls you see all over pixel artist circles on dA 
• Limited on striking compositions, animations, and costuming with the most extreme chibi proportions. This might explain why there's more tier-2~3 art that's popular on deviantART because there's a larger body size that gets you far better potential for good composition

That might explain why I found less 'extreme-cute' than I thought there would be. Most of the art that makes it up high in my search results actually falls somewhere in the middle from cute to real.

However, cuteness at around tier-2 is still hugely popular if you have entire imageboards catered to the otaku subculture, such as this: https://danbooru.donmai.us/explore/posts/popular [warning: explicit images!] I also don't take much interest in anime, but I'll tell you that I think that the art found there is pretty damn generic.

I wanna say that I find most potential around tier-3 where I stuck in western-style illustrations, particularly styles that enjoy a great deal of creative liberties with face shapes and cuteness is not always a goal. However, cuteness is still highly possible with the given art styles, just that not every character has to be cute.

I'm still looking into this, but this is where things have brought me. Cute is good, but going for cute comes with its own challenges

[Tourist]

I'll toss an uninformed opinion in.

There's a bit of philosophy that says that says the self is not a single object, but an aggregate of things.  Of relevance to this thread, it includes

1) The external world and stimulus (the pixel image)
2) Feelings toward the thing (emotional reaction)
3) Recognition of the thing (memory and identification)
4) The collection of associated memories, ideas, and whatnot that are connected to the idea of the thing.  The surrounding mental map in the viewer's mind as it were.
5) Items 2, 3, and 4 in aggregate make up consciousness, which we perceive as a whole (the mind)

With this model,
1) the cute image tends to stimulate on an emotional response
2) the realistic image tends to stimulate the recognition and memory
3) the cartoon tends to stimulate the association response.  Jessica Rabbit doesn't look like a real person, but she has all the associated pieces.

Of course these aren't singular responses.  The cute image also triggers the recognition and the viewer's mind is adding in all the associated concepts to what they see.   The others work the same way.

It's also possible to have some responses but not another.  An image can provoke an emotional reaction even if you can't recognize what it is.  Or you can recognize it but feel no emotional reaction.  I expect the uncanny valley happens when associations of an image conflict with the recognition of the image as a real thing.

We normally operate at the aggregate (conscious) level so we aren't really paying attention to the different divisions.   I could blather on, but I imagine you get the idea.  This all comes from Buddhism, and I'm no expert on Buddhism nor philosophy, but maybe you can find it useful.

Tourist

[astraldata]

I expect the uncanny valley happens when associations of an image conflict with the recognition of the image as a real thing.

This is probably why some people find the Jessica Rabbit image I posted above to be creepy, whereas I myself have no real feelings toward it one way or the other aside from the simple respect for the fact that it was well-crafted. I think the reason I don't find it creepy is because I've seen much creepier images (that realistic Mario face for example -- waaay more highly-distorted than Mrs. Rabbit there, who just seems a lot more believable [and less gross-looking] imo, who only triggers the creepy-meter when I look way too hard at the image hoping to make more sense of the features than at first glance).

That being said, Like Tourist, I feel the uncanny valley is more of a sliding scale on a per-person basis. I suspect people like Carnivac and |||| might have more vivid associations between fictional images and reality than I might, making images that contradict (their idea of) reality so blatantly a lot more intense to them. If that's not the case, my second guess would be that maybe they look much closer and harder, and/or maybe longer, at image details than I do. Some people like more detail-rich images than others (I personally favor clarity -- too much detail is distracting imo -- but I do prefer clear and quickly-readable (concise) images over staring at a detailed, but overly complex, paintings for 30 mins [unless, of course, I'm studying them], so that tendency not to look too closely at an image [preferring to see the "gestalt" of images instead] might be a significant factor to the reason why my feelings seem to differ about such an [admittedly creepy] image [but I only feel it's creepy after I decided to look back at it really really hard, while second-guessing my initial impression of the image].)

[lachrymose]

Necroing an old thread!

But I have an interesting video to show. Shows how a guy animates robots to interact with humans. While he doesn't really get into the Uncanny valley I think there is still a lot to be learned about the uncanny valley and how to avoid it using this guys techniques/methods.

http://www.ted.com/talks/guy_hoffman_robots_with_soul?language=en#t-226690

Enjoy!

[tocky]

well if the thread has been necrod i will post in it i guess.

this is a good thread. good advice. but let us have another good advice.

sometimes its good if the player is made to feel uncomfortable. there are reasons to wish this.

certainly, we should still love those children who are ugly because they might teach us new ways to be pretty.

anyway, i dont know if you guys really know what im talking about but I understand ive been trying to do too much lately. tried to prove too much. i have made too many people mad. i am sorry.