AuthorTopic: Game Environment Composition Techniques!  (Read 7630 times)

Online Ai

  • 0100
  • ***
  • Posts: 1054
  • Karma: +2/-0
  • finti
    • View Profile

Re: Game Environment Composition Techniques!

Reply #10 on: December 14, 2015, 01:18:41 pm
Wow, thanks for taking the time to explain that astral. It's really the first conception of 'communicating meaning' I have seen that seems to be actually constructive, rather than making the learner into a product in a factory of 'knowing things that I the authority am handing down to you'. It helps me to see meaning as something slightly more than a kind of instrumentally-useful delusion, which is helpful on both a personal level (doing things that are important) and a idea-communication level (try to make and introduce open systems of ideas, rather than depending on people to simply accept or reject some set of assertions comprising a pat story)
New AA tutorial, about handling irregular lines.

'Better software looks like "people who know what their problem is and why they have it"'

Offline astraldata

  • 0010
  • *
  • Posts: 328
  • Karma: +0/-0
    • View Profile
    • MUGEN ZERO

Re: Game Environment Composition Techniques!

Reply #11 on: December 15, 2015, 04:05:16 pm
Ai, I'm thankful to know my thoughts were helpful to you! :)

To digress a little, I have struggled with conveying this concept of 'meaning' in art for a long time, and despite knowing it intuitively, conveying 'meaning' always seemed so subjective -- something seemingly so invisible -- and yet at the same time, so prevalent, but also essential, in impactful works. I knew that to apply any sense of meaning (and therefore impact) to one's work, it was impossible to outright 'tell' a person the meaning in any direct way, especially where that meaning was complex or seemingly obscure and specific, and expect the observer to appreciate it as fully as it was intended -- discovery is essential to true appreciation. Purposefully convoluting or avoiding simplicity in a work is equally as bad though, as it inhibits discovery in the observer (the example piece is simple but also impactful due to the meaning it held being deceptively simple, but turning out to be heavily complex).

Many regard true 'art' as some lost secret that only long-dead masters and certain other divine beings understand. To these people, you just have to take long walks in the park and practice drawing/painting/etc. everyday and it will soon come to you how to be an 'artist' and convey deep meaning to others in your work (which is a total load of BS, as indicated by my previous post).

As much as I love art with meaning, I personally come from a background of game art more than traditional art, and for all intents and purposes, game art relies on one's ability to develop individual assets as quickly and as effectively as possible with a quality level relative to that of the importance of the asset versus the time you have to produce it. With this as a standard for your work, conveying true 'meaning' in art tends to take a backseat, so even to skilled artists it sometimes feels like you're a machine churning out piece after piece for an overlord client and/or art director. That being said, even in game art, the core issue with creating art assets is still being able to communicate effectively, which means that you are a more effective artist if you can offer 'meaning' in your art that will allow your observer to arrive at his or her own conclusion as to what it is you were trying to represent (while also offering them everything they need to do that in such a way that they arrive at the conclusion you set out for them.) And although everyone experiences stuff just that little bit differently, there will always be a clear commonality in how each person experiences a thing when that 'thing' is communicated well. Not to mention that being able to communicate meaning more clearly makes the art process more fun -- even when applied to things as seemingly mechanical as game art / asset development -- because your art assets can tell a whole (seemingly invisible) story that the main narrative, with all of its fancy exposition and nice character animations & special effects, simply cannot. All in all, communicating meaning is a very powerful and far-reaching ability -- and it will make anyone a better artist to be able to do it well.

PS: @RAV -- If you wouldn't mind, would you care to share what you mean with us? How does Kishotenketsu apply there? I'd love to hear!
I'm offering free pixel-art mentorship for promising pixel artists. For details, click here.

     http://mugenzero.userboard.net/

Offline AlexHW

  • 0100
  • ***
  • Posts: 1029
  • Karma: +0/-0
    • View Profile
    • AlexHW

Re: Game Environment Composition Techniques!

Reply #12 on: December 15, 2015, 05:44:37 pm
I like this Kishotenketsu concept, astraldata.
I was thinking more about it and I while it is nice to think that it offers a way to convey a story without conflict. I don't think that is what it does.
Conflict has to be interpreted. It can't exist without having relativity. I would say that Kishotenketsu actually creates conflict, because the twist(ten) introduces a seemingly different & contrasting idea that must be interpreted. Ketsu is the interpretation of the conflict. The Ketsu then becomes an example of how to handle conflict, and the reader/viewer learns from that example. By learning from examples on how to interpret conflict, they become more capable of doing so on their own. So those who are more comfortable with Kishotenketsu may have exercised their ability to interpret conflict, and thus it may seem to them like there are none.
So the way I see it is that, Kishotenketsu has more of a relationship with the reader/viewer- the sole purpose of the story becomes one that teaches the reader/viewer on how to handle conflict. The conflict is created, and then the sides become blended together.
Am I misunderstanding it?

Offline Conzeit

  • 0100
  • ***
  • Posts: 1449
  • Karma: +3/-0
  • Camus
    • conzeit
    • View Profile
    • CONZEIT

Re: Game Environment Composition Techniques!

Reply #13 on: December 16, 2015, 06:58:19 pm
Oh, Astral thank you for continuing the conversation! I love making halfassed stabs at something I have felt experientially but have no framework to work with, and then have someone give me a framework to use on said thing, sooo pixelation =) LEARNIN!  :D

Alex. I can see what you mean, but honestly a twist doesnt have to be conflict. it can just be mindblowing twist a-la M Night Shamalan. Conflict happens when there is an opposing force, and it hinges on the character's expectation to overcome it, while a twist by itself is shock value, a one off. This is why movies based solely on twists dont work on a second viewing, because the cat is out of the bag, like a magic trick it holds no interest anymore. I think the stairs image is a good example again, you are not surprised the second time you view it, but you still feel for the guy because the image speaks about his mentality and how he cant overcome his opposing force, his own psychology, so it still holds interest. I think the ketsu part saves kishotenketsu from being cheap reveals is the ketsu, if you cant round it all up at the end and make a coherent whole statement, it was not a story after all. I think I'd have to see more examples of kishotenketsu to know wether it avoids conflict or not.

The Kishotenketsu concept reminds me of setup-punchline as well.
EDIT: more than setup punchline it reminds me of essay structure. Thesis-antithesis-thesis. only with required development for the thesis after it's introduction. it's got a bit in common with shakespeare's 5 act structure as well.  I think all these structures to deliver ideas have a lot in common, but vary slightly because of the specific thing they deliver.
oh boy. I'll have to rumiate on it some more to bring more into the conversation
« Last Edit: December 16, 2015, 07:42:26 pm by Conceit »

Offline API-Beast

  • 0010
  • *
  • Posts: 260
  • Karma: +0/-0
    • API_Beast
    • http://pixeljoint.com/p/35725.htm
    • View Profile

Re: Game Environment Composition Techniques!

Reply #14 on: December 20, 2015, 03:14:41 pm
@Astral:

You mentioned that single art assets themselves can have meaning, as someone who's background is in Game Art too I would be very interested in seeing a example of that.

The way I see it each art asset is a small puzzle piece, no more, no less. It itself has very little meaning, the meaning starts to show when all these puzzle pieces are combined into a picture by the player. Another analogy is facts, a single fact has no meaning, but as you collect facts they become interconnected and you understand more and more of the story.

Offline astraldata

  • 0010
  • *
  • Posts: 328
  • Karma: +0/-0
    • View Profile
    • MUGEN ZERO

Re: Game Environment Composition Techniques!

Reply #15 on: December 26, 2015, 08:32:23 am
I like this Kishotenketsu concept, astraldata.
I was thinking more about it and I while it is nice to think that it offers a way to convey a story without conflict. I don't think that is what it does.
Conflict has to be interpreted. It can't exist without having relativity. I would say that Kishotenketsu actually creates conflict, because the twist(ten) introduces a seemingly different & contrasting idea that must be interpreted.
...
So the way I see it is that, Kishotenketsu has more of a relationship with the reader/viewer- the sole purpose of the story becomes one that teaches the reader/viewer on how to handle conflict. The conflict is created, and then the sides become blended together.
Am I misunderstanding it?

To be honest, I think you're close with your understanding, but the thing that separates the 'twist' in eastern vs. western story-structure is that, in the east, conflict is not necessary to have a story. In Japanese, and in other similar cultures in the east, it is very common to observe that a thing is simply what it is -- a thing -- and that 'thingness' is celebrated over there. While that seems 'boring' to western culture, the 'thingness' of a thing, being as authentic as it can possibly be, is the essence of substance and truth. While 'ten' is generally used to throw an idea into the mix that isn't expected -- the purpose of the 'ten' aspect as it is used in eastern storytelling isn't to make a story more interesting or more palatable to the audience as much as it is to help promote their understanding of the subject from less obvious and more telling angles -- in other words, teasing out that 'click' or 'aha' moment that everyone enjoys feeling when they finally put together a seemingly 'hidden' meaning and/or relationship that was actually already there all along. It is the idea of creating an understanding that was just beyond reach in which you were able to make the jump (with a little help from the story) on your own that makes Kishotenketsu so effective. Conflict, or any sort of perceived (or actual) contrasting setups, is actually a byproduct of that goal and should never be the focus, or otherwise you run the risk of breaking the suspension of disbelief you worked so hard to create in the development phases of 'ki' and 'sho' when you arrive at the 'ten' portion of the story.

I know it's subtle, but it's very important to keep in mind that, as Conceit hinted at, there are underlying truths that make the work more long-lived than a simple "does the guy save the girl, or does the bad guy kill her?" sort of setups. For example, the fact that she dies because she's attracted to the guy out to kill her while the good guy risks his neck for her, even after warning her and then being ignored, changes the overall nature of the story towards more universal truths, perhaps about blindness and ignorance and a potentially rationalized 'love' on the guy and girl's part -- things that aren't very easy to put into words -- leading to something that's deeper than a simple reveal of the plot twists to the audience, and also to something that 'you have to see to understand' when you share it with your friends -- versus, of course, 'it's just another guy with a superhero complex who just needs to get laid' sort of story that you can easily convey to anyone.

To further elaborate on the lack of necessity of conflict, in the example I gave above, the 'ten' could be that the guy is killed by the bad guy when he tries to save her (despite being ignored), leading him to be at fault almost as much (if not moreso) than the girl -- or even the bad guy! The 'ten' (which, in this example, actually eliminates the conflict in this story altogether, rather than simply resolving it as most would expect) offered alongside the rest of the story development gives the overall story a more long-lasting shelf-life (or short-lived, assuming the 'ten' was absent or poorly executed). In the end, the characters could suck, the writing could suck too, but as long as there is still some human-level understanding reached from the combination of all the parts (the guy's part, the girl's part, and the bad guy's part) in the story, Kishotenketsu is probably present in some form. In the end, conflict wasn't necessary as the guy didn't save the girl, and further, it was a moot point anyhow since he was killed -- so the story is no longer about him saving the girl, but it instead it became about you, making you think about how and why things ended the way they did, meaning that it is your story now -- and that is exactly what makes it transcend simple plot devices, like, in this case, conflict itself, while still keeping the story 'interesting' in the end. If it were a movie, you could rewatch it a second and a third time, and you may come to a different conclusion about why the guy should or shouldn't have gone to save her every single time, but just like in art, the perception you want to convey vs. the one you actually convey depends upon how strong you are at communication via your breadcrumbs of understanding to allow the audience to come to the conclusion you want them to. Bottom line is, conflict aside, Kishotenketsu is about effectively communicating an authentic and universal understanding through some medium. Generally, the understanding you wish to convey is only possible when you communicate it on a level of truth and authenticity that anyone can see and feel (sometimes that truth even has to be a little dirty) -- but to convey authenticity like that, you have to actually understand it on a level you can also put into physical form! Helping the audience come to an understanding on their own by providing them the tools and resources to do so is a skill that requires many skills, but the most important of which is the ability to cut through to the most base and experiential level possible to actually see the subject from every angle possible from both above and below (and never just from your own pov [or those similar to it], because those views will only make it more difficult to see more truthfully) -- From that point onward, you can determine what it is you want to actually say about the subject once you have gained the ability to actually see.

(As a side note now that that's out of the way, I forgot to mention earlier in this post when speaking of another 'form' of Kishotenketsu in the example story, the 'ketsu' in the above example would likely be 'mixed' with the 'ten' just after the 'ki' and 'sho' exposition are complete since, the 'understanding' in this example story can almost be reached in very nearly 3 parts (but not completely!) instead of the usual 4. I say 'almost' because the 'ketsu' and 'ten' concepts are not actually merged, they're just tightly coupled -- the core concepts of 'ten' and 'ketsu' are still completely there functionally, regardless of the form the exposition takes that masks their presence by allowing them to happen almost simultaneously.)



@Mr. Beast
It wasn't anything complicated. As it's fairly obvious that most of your game assets are, by default, supposed to further a game's story/lore/world/etc., individual assets can tell their own individual stories as well. Kishotenketsu, since it is about communicating effectively in a way that teases an understanding out of its audience, allows one to use virtually anything that can communicate to tease that understanding out. A simple example is a sprite character's posture and appearance -- he's confident because is back is straight (ki) he's determined because his eyes are straight ahead, even with weird girly-looking eyelashes (sho), but he has the body of an ugly yellow sponge-with-a-face that wears a dumb-looking hat and having tiny stick arms (ten), which tells us he's unaware of his appearance since he's obviously somehow ready-to-go-at-a-moment's-notice like any other go-getter superhero with an overly-positive attitude (ketsu). This art asset offers meaning to all of those who gaze upon him because there's an underlying thing there in your head that says "this guy is totally stoked -- but why? He has nothing to be stoked about, right?" and yet your audience, as players, secretly feel encouraged by the effort you, as the artist, put into making him so true to his own personal uniqueness, leaving them with a sense of meaning from nothing more than a simple drawing that puts forth clues to who he might be, making them feel stoked themselves to play as him. Even the story itself couldn't tell you why he stands the way he stands or feels the way he feels about the world -- only the asset itself can truly (and most effectively) convey its own meaning (indirectly) through visual clues that tie together their invisible story. This generally helps you to arrive at an understanding about their character (even a tile in a tileset is a character, after all!) deeper than you would have if they had been told to you explicitly through dialogue and cutscenes in the story itself. These tiny stories always have their place, and they are personal and private to each asset you create, making them very effective at communicating an understanding of themselves. And, by extension of these tiny personal stories arrived at by way of Kishotenketsu, the world you're building becomes much more rich and substantial, making it a million times easier to relate to, and also to understand!

(I apologize in advance if this doesn't make sense -- I might have to approach this again when I'm not falling asleep at my desk. Maybe I can condense it into something more readable tomorrow... D: )
« Last Edit: December 27, 2015, 03:03:03 am by astraldata »
I'm offering free pixel-art mentorship for promising pixel artists. For details, click here.

     http://mugenzero.userboard.net/

Offline API-Beast

  • 0010
  • *
  • Posts: 260
  • Karma: +0/-0
    • API_Beast
    • http://pixeljoint.com/p/35725.htm
    • View Profile

Re: Game Environment Composition Techniques!

Reply #16 on: December 26, 2015, 07:10:52 pm
This reminds of what Feng Zhu is teaching about concept art in this one hour long video.

He doesn't talk about meaning, he doesn't talk composition either, just about designing "cool stuff". Copying something existing 1-to-1 is boring, e.g. something we are 100% familiar with does not make good concept art. Something that is completely new and foreign is just "weird", and puts a lot of people off. In between those two extremes is what he calls the "Goldilocks Zone", which is the "cool" stuff. People can relate to it, but it's also new and original and thus exciting.

He then talks about how designs can be created that fall into the "Goldilocks Zone", he does that by combining two familiar concepts that aren't seen together, usually. For example, he takes the pyramids in Egypt and puts them into a snowy tundra. Or he takes spiders and "reskins" them to be mechanical and bigger and he gets spider tanks. Or he takes a train and turns it into a budding city.

Offline API-Beast

  • 0010
  • *
  • Posts: 260
  • Karma: +0/-0
    • API_Beast
    • http://pixeljoint.com/p/35725.htm
    • View Profile

Re: Game Environment Composition Techniques!

Reply #17 on: December 27, 2015, 01:42:03 am
Let's distill this down:

  • Humans are familiar with certain concepts and objects.
  • In order to learn the new needs to be put into a context with something that is already familiar.
  • Learning provides enjoyment.
  • Learning can mean very trivial things; Like understanding the character in a story a little better or just understanding what is displayed in a picture.
  • The concepts and objects we are familiar with are interconnected.
  • It's much easier to teach and come up with new connections, than new objects themselves.
  • So to create some kind of "meaning" we need to find and isolate a unfamiliar connection between two or more familiar things and portray it.
  • Every piece of art we create need something familiar as a foundation, because we need a context for the connection or concept we want to introduce.
  • The network of familiar concept of a individual is constantly growing and changing, so a art piece might be interesting for one individual, but not for another.

For example, everyone should be familiar with the concept "robot" and "sadness", but we are rarely reminded that those can be interconnected, and thus drawing a sad robot provides some sort of meaning.

Pictures are limited to displaying only a limited amount of connections, but other art forms such as films, writing or video games can develop these connections much further, capable of creating deeply nested networks. In this case we can speak of "depth".
« Last Edit: December 27, 2015, 01:47:25 am by Mr. Beast »

Offline astraldata

  • 0010
  • *
  • Posts: 328
  • Karma: +0/-0
    • View Profile
    • MUGEN ZERO

Re: Game Environment Composition Techniques!

Reply #18 on: December 27, 2015, 05:19:59 am
Yep, applying 'ten' to a concept is a great way to achieve an interesting take on it while also expanding the concept with, perhaps, backstory to help establish (ki), and develop (sho), an initial visual design into a harmonious visual concept that makes sense and is relatable (ketsu), but also remains understandable when you apply the twist (ten) to the design as well.

In the idea of the spider tanks, for example, the 'ten' would only make sense if there was a functional (real-world) advantage to the tanks being on spider legs. In this case, form follows function can applied particularly well if you're creative enough, and even inspire real-world designs if you're authentic enough too. I can think of a few real-world functions that would allow a spider-tank to be particularly advantageous, for example. However, for them to be appealing -- I mean /really/ appealing -- they must also work well enough to continue suspension of disbelief.. even in unrealistic settings and situations.

The idea of taking ordinary stuff and turning it into something fun and interesting is essentially the core of what a concept artist does, imo. No offense to Feng Zhu intended here (amazing concept artist!), but to use vague terms like "Goldilocks Zone" to describe something between "cool" and "relatable" is kind of a cop-out. At the same time though, many really great artists (or maybe even /most/ great artists!) don't really /know/ how they do what they do quite specifically enough to explain it. Instead, a lot is intuited, and a lot is learned, but to pass that knowledge on to others, you need to be able to verbalize it.

In my opinion, to make something "relatable" /and/ "cool" simultaneously, you don't need some vague and mysterious idea like the "Goldilocks Zone" concept you mentioned, being that it's only real 'use' is to a master artist who already knows how to employ it -- you simply need to develop that human/real-world element in your work, no matter if your work is a book, a movie, a drawing/painting, completely abstract or cartoony, or even if it is inanimate like a tree or a building, or whatever -- you /need/ to be able to characterize it somehow, and the way to give something character is by offering that 'ten' or that 'twist' to make it unique enough to be its own thing while also retaining the fact that it also makes sense why it is (visually or inherently) who (or what) it says it is. In the end, it needs to be understandable and relatable by way of its characterization, at least in order for it to be appreciated and taken seriously for what it is. That's it's shelf-life, as mentioned in my previous post -- and it's not something one would want to overlook when offering a concept to the world that the world will also appreciate. In the end, people /always/ love a good character-driven story. It doesn't even truly matter what the context is -- if the characterization is smart and clever, the work (be it a concept, a story, a song, a picture, a sprite) will always be well-received.


Digressing a bit though, but I do highly agree that copying something 1:1 is quite boring to the viewer. On the other hand, it is also boring to the artist as well, so most people, I feel, only do it when they have little knowledge of the subject. This is, frankly, why I think so few people practice life-drawing or reference-drawing when first starting down a professional path to doing art. It sets one waaay back in their art ability to not have a solid visual library in memory to work from for whenever you need a twist in your subject to form a concept, but you're not as familiar with the given (and potential) subjects you're trying to combine as you might have liked to think you were before actually trying to draw them.

So, some friendly advice to other artists who might be struggling with boredom in reference images -- You will need a visual library to be able to put the 'ten' into your concepts as effectively as someone like Feng Zhu, for example -- so try inventing parts of references/subjects you're drawing using your existing visual library in your head, and try only to understand the parts you don't have any ideas for twists on by drawing them as you see them. This will give you the ability to learn to see a thing as it is, while also flexing your creative-memory muscles in accessing the library you already have more quickly, teaching you to churn out ideas as fast as you can draw them. This will prove invaluable to you as a concept artist, or even as a game artist, since it allows you to learn to see and be creative simultaneously.

This is how I do it -- but your mileage may vary. Give it a shot and tell me how it goes! :)
« Last Edit: December 27, 2015, 05:33:58 am by astraldata »
I'm offering free pixel-art mentorship for promising pixel artists. For details, click here.

     http://mugenzero.userboard.net/

Offline astraldata

  • 0010
  • *
  • Posts: 328
  • Karma: +0/-0
    • View Profile
    • MUGEN ZERO

Re: Game Environment Composition Techniques!

Reply #19 on: February 14, 2016, 09:16:06 am

Been thinking a lot about game level composition (as indicated by this topic apparently), up to and including procedural level generation (I was recently introduced to "No Man's Sky" thanks to this!), and my journey led me to a very insightful blurb about the UbiArt engine that brought me back to this (apparently almost dead) topic to share another quick (but amazingly overlooked!) technique with you!


The Origin of the UbiArt engine is in Procedural Content Generation.

Yep. You read that right. It was originally built as a fast way to create 2d game levels.

"No way!!" you say? "Yeah! Rayman Origins didn't include any procedural art!"

Well... I beg to differ. According to the article linked after this paragraph, it was originally built as "a visual pattern generating-based engine" as a continuation of what they were doing with Beyond Good and Evil 2:

http://www.gamasutra.com/view/feature/6562/the_evolutionary_process_of_.php

According to Michel Ancel (the guy who started the UbiArt engine), "Artists make some patterns, like samples, and we use them to quickly model levels, as to have a forest, for example."

This tiny sentence says a ton about the idea (and the creative mindset!) behind this amazing engine -- but most importantly -- it says what the engine actually does!

For those who haven't figured it out yet -- the UbiArt engine creates procedural art by using an algorithm guided by work produced by the hand of an artist.

Although, in and of itself, this is not entirely new, it is the way procedural generation is applied to level generation here that is quite interesting. I'll get into that in a moment.


In English Please!

To put it simply, everything in the Rayman games' environments since Rayman Origins has been produced to be as modular as possible. This means the same setpieces can be used multiple times in multiple places in multiple ways -- and when you add in a lighting system that allows 2d art to appear 3d lit, colorized, and then add customizable fog to give depth, and then vertex-colored sprite assets that blend with the light and have nice normal maps, allowing them to be stretched and scaled as well as rotated -- you have a system that lets the exact same assets be used hundreds of times without ever feeling stale. BUT, when you design a system that essentially creates your lush backgrounds FOR YOU -- you've got little work left to do aside from producing some nice assets and dropping them into the game!

Okay, so that probably doesn't explain what I'm trying to get at clearly enough, so I'll try to point out some (admittedly well-hidden) artifacts of this procedural process. Let's take a look at this image with the house-looking things near the coffins:



As you might notice, the general level theme is these houses and stacked coffins. However, less noticeable is that the houses you see, as well as the piles of coffins and wood are heavily repeated (even in the same screenshot there!) and, even knowing this, you still have to look really close to determine this because there's lots of mirroring, scaling, and general color changes going on all over in that image that the one 'unique' house/mausoleum/whatever on the far right of the screen is just enough variation to fool your trained eye that you're looking at a nice painting thanks to the amount of use each of the elements have when combined with the various options available to them. This engine even has areas of light/shadow/colorization it can overlay on top of things to break up the hues and create separation, adding variety using very few elements. An artist's dream indeed.



In the forest portion of the image above, you will see that, although there are other bits and pieces scattered about between them, there is a single large tree used repeatedly across the image (hint: it is the most detailed and in-your-face tree there) and yet, somehow, it just doesn't register in the brain that it is repeated at all. Although this image is labeled as concept art, judging by the previous in-game image, it is highly likely the game engine would produce these trees at random intervals where there is terrain (using said algorithm, which is strangely missing from the UbiArt demonstration video on youtube -- that portion seems to have been cut out of the demonstration entirely).



The rocks in the icy world above show the most visible seams in the system when the BG rocks behind the action transition to the darker machine-looking blob above them. This indicates to me that there was an algorithm at work here moreso than an artist's hand in particular. The plain-looking rocks would have been the swathe the computer generated overlays upon -- overlays that were originally placed by an artist into some sort of template most likely. These overlays of machines and flags over the rocks would have been spaced and scaled randomly by the computer, while the artist would have only needed to select an area that needs a particular template, and the algorithm would generate the randomized assets following the template specified by the artist using something akin to Perlin Noise to shift stuff and scale it according to the original artist's settings. The artist placing it would just pick a region where it is necessary to use that area (such as the generic rock repeating texture with some stuff overlaid on top of it, shifted around semi-randomly, following the noise settings in the template.)

This is, of course, my own theory based off what I've seen of the UbiArt engine, and gleaned from images like the ones above. Unity has a plugin that attempts to mimic the functionality of the platform portion of the UbiArt engine, and does it pretty well from what I've seen. However, it is the amazingly detailed-looking backgrounds that make Rayman stand out so much. That being said, an engine like the UbiArt engine takes a lot of the tedium away from you when it procedurally generates portions of the game world visuals out of your assets and designs while still allowing you to go back in and tweak them to be exactly what you're after. I'm pretty certain there's some sort of rough "fill out the level" algorithm leveraging templates akin to the ones I've mentioned used in these very large, very detailed (fast-moving!) levels. I highly doubt a level designer slaved over art assets to fit just perfectly in that section where that first platform falls when that big thing is chasing you.

Just my two cents. Feel free to add on to or take away from my ideas here -- I'd love to explore them more if anyone's interested.


Rules of Procedural Generation -- No Man's Sky

I mentioned "No Man's Sky" at the beginning of this article because it is a game that takes Procedural Level Generation to the next level (pun intended -- don't hate.) It is another interesting way to use an artist's touch to generate not only levels, but worlds, using rules in the form of a pyramid to generate content vs. rules in the form of a tree or heirarchy. This interesting approach could also be used to design levels, perhaps in a similar way to Rayman (which, I feel, might have already been done as my theory stated above), but if it hasn't, it'd be a hell of a way to make and decorate levels very quickly.

Just some food for thought. :)
I'm offering free pixel-art mentorship for promising pixel artists. For details, click here.

     http://mugenzero.userboard.net/