[FILLED] Looking for artist(s) to develop new style for A Valley Without Wind.

[x4000]

Update: If you just want to see the latest samples and not all the discussion, then please see the following locked thread: http://www.arcengames.com/forums/index.php/topic,11063.0.html

If you want to actually discuss those samples (and we'd love you to), please come back to this thread and do so.  It was suggested in this thread that we make a locked thread with just the latest versions of the art, so that there's not so much scrolling around to see things side by side, so that's what we've done.  Cheers!


Originally here: http://forums.indiegamer.com/showthread.php?30388-PAID-Looking-for-artist%28s%29-to-develop-new-style-for-existing-2D-Steam-game.

And here: http://www.wayofthepixel.net/pixelation/index.php?topic=14189.0

Figured I'd also make a post here in case there are any pixel artists (or similar) that olks here know.


The game in question
You may or may not have heard of A Valley Without Wind, or my company's other projects AI War and Tidalis (also both on Steam). AVWW is our latest title, just having arrived on Steam in late April.  So far it has done reasonably well for us, but there seems to be a very big divide in opinion of the game based around the art in particular.

The Problem
The art that is currently used by the game is all spritesheet animations, mostly which was rendered in Vue or DAZ 3D and then which was post-processed rather heavily in photoshop to give it a painterly look.  Some of the tiling backgrounds were from Genetica or other sources.  In other words, mostly the art used here is royalty-free stock art which I then lit, rendered, and post-processed into a semi-consistent somewhat-painterly style.

Some people seem to genuinely like the art style, and I think that it is beautiful in parts.  Other aspects do not gel however, and I know that some things could be better.  There are also inherent limitations of doing prerendered 3D for enormous numbers of objects: all those textures make loading slower, take up RAM, etc.  That puts extreme limits on the number of frames we can have in a given enemy animation, for instance, and makes something like a transition from one state to another impossible.

That's less of a problem with something inherently more abstract, like pixel art; but I think we're falling into something akin to the "uncanny valley" with the current art style.  It's causing lots of moaning on various forums, and it's my belief that we could see wider success with an art style that was more universally accepted.

The catch-22 with this is that AVWW has already been a very expensive game to make, and it has thousands of images already comprising it.  And we are always working to add more enemies, more spells, more objects, etc.  So this means that if we want to try out a new art style, that's going to be very expensive because of the amount of rework to be done, as well as then the ongoing cost of adding ever more content.  On the surface, this would make it sound like we are either stuck with the current art style or going to have to rely on incredible amounts of charity from artists -- the former may wind up being what happens, but the latter is not ethically acceptable to me (or remotely sustainable in the long run).

The Project
Ultimately, right now what I want to do is work with several artists to develop out several different prospective art styles.  Some of them will be pixelart-based (hence my posting here), but I'm also going to look at other forms of 2D art via other forums.  Right now I can honestly say that I'm not sure what I'm looking for, except that I'd like for the artists to take the foundation that I've laid down (the game is already released, after all), and transform it into something that makes people go "wow" instead of "ugh."  I could see a number of ways that could happen, but the specific style depends on the artist in question.

As I said, right now I'm looking to develop proof of concepts.  This is a paid offer, not a speculative one.  The process would be:

1. I'd like to see portfolio work from anyone who is interested in this, and what your usual hourly rates are.

2. If you have any specific ideas that you would like to propose for the art direction, feel free to mention them, but this is not required.

3. For the artists whose style I believe would be fitting with the project, I'll be back in touch with more information about the specific prototype art I'd like you to work on.  Even if your style isn't quite what I'm looking for, I'll at least have the courtesy to let you know.

4. When we look at the specific scene in question, based on your rates and estimate of time it would take to recreate the scene in your style, I can then be certain that this particular prototype fits into my budget for this.

5. Ideally within the span of the next two weeks or so, multiple artists will complete the work in their style, and then if any further tweaks are requested by myself to the initial work, those would also be paid at the agreed-upon hourly rate.

That's the whole of this project, at the moment.  It's essentially for R&D only, from my perspective.

Next Steps After This Project
Once an art style is selected, the next step will be to work with that artist to determine what the likely total project cost would be to convert the entire existing game over to that style.  Not to mention the timeframe.  My expectation is that multiple artists will be needed, ultimately, so then the next challenge will be seeing if we can find multiple people who can work in that style.

At that point I'll have a more accurate idea of exactly what the revision costs as well as the ongoing-development costs of the new art style would be (although I already have a pretty decent idea, having worked with 2D, 3D, and pixel artists on other projects in the past).

With those cost figures in hand, plus an actual visual prototype that people can see and look at, the next step will most likely be a kickstarter.  It's possible that a discount sale from our game could give us enough of a windfall to fund the entire thing ourselves directly (this sort of thing has certainly happened to us before), but we already have a fulltime staff of 5 and so there's a limit to how much extra financial burden we can take on even if we hit another major windfall this summer.  Most of our general budget goes toward expanding the game itself, which is what keeps bringing in new players the most reliably at the moment.

My hope is that with a solid prototype in hand, plus reasonably accurate conservative cost figures (aka, with overage built in just to be safe), a kickstarter for improving the art of this game should be a popular idea and something that people can get behind.  This has been a pretty high-profile game as far as indie games go, but you can gauge that for yourself.  Short-term this is basically just a paid stint doing a prototype of what the game could look like, and then if we're able to procure the funding to make that a reality, then this would turn into something larger.

More Project Details

Currently the game is prerendered 3D, with most characters being 128x128 images with running animations that are typically 21 frames or so.  Some of the monsters have a lot more frames than that, and vary in size from down to 32px square all the way up to 1024px square.  Most monsters don't get larger than 512px pixels square because of RAM limitations.

Generally speaking, aside from the running/flying animations, everything else is single frame -- characters spin as they jump based on rotation of a single sprite frame, then have a different frame for falling or ducking, etc.

The particle effects are something that I will probably continue to handle -- I'm quite pleased with how they've been turning out, and I don't think anyone has been particularly complaining about them.  Those are generally sprites rendered in particle illusion individually, and then animated procedurally in-game using our own custom particle system.

At the moment the game supports anything from 800x600 screen resolution all the way on up to anything above that.  The size of the viewport simply changes, rather than scaling individual elements.  That gives a fairly different feel to playing at 1080p and makes repeating background textures a lot more glaring, for instance.  My thought is potentially to cap the maximum resolution at 720p and just employ scaling above that rather than allowing the viewport to get any larger.  I mention this because I know that has some substantial effect on the amount of repetition and thus the size that background repeated textures must be.

There a few different art directions that seem viable to me, and which I'm interested in exploring.  But if you have ideas beyond the below, don't hesitate to suggest them:

Option 1: The simplest and probably least expensive option would simply be to tidy up what is already there.  There's a painterly aesthetic already, so if you have a tablet and good drawing skills you could probably add some flourishes and corrections that would help quite a bit.  Plus potentially there are some things that I'm overlooking about differentiating the foreground and the background that might be worthwile, etc. 

I don't think that this option would ultimately be transformative enough to win over the people who absolutely abhor the current style, but I could be wrong.  And if it turns out that I'm not able to procure a large art budget for the actual implementation phase of this project after the prototypes are all done, then this could be a middle-of-the-road route that would at least improve things somewhat.

Option 2: Going hardcore pixelart.  This would be replacing basically everything except the particles and probably the font, and it would be a pretty big job for several people, to put it mildly.  One immediate challenge of this is the scale of the art that we are currently using: 54 pixels per meter at the level of the characters (though some things are closer or further away, obviously). 

Doing a pixelart character that is extremely detailed at 128px square strikes me as expensive and perhaps infeasible, though I'm sure it could be very beautiful.  Perhaps a better idea would be to instead do the art at half-scale (which I know is still twice the resolution of the SNES), and then I can program the game itself to upscale your sprites to the appropriate in-game scale.  That would make things look extra blocky, but the scaling should be sufficiently high-quality that we wouldn't see blurriness.  That's something we'd have to experiment with.

However, that option appeals to me the most because it is relatively simple technically-speaking, and lets us save a lot of RAM.  Presumably with pixelart, instead of 21 frames for running we could get away with 2-4 instead, too.  For some of the larger monsters that have animations, rather than them being completely frame-by-frame of course they could have their component parts individually animated like the larger bosses were back in the SNES days.  That would take more coding, but not a whole lot.

Option 3: Whether it's pixelart or some other 2D art style, we could convert the entire game to a radically new look (again keeping the font and the particle effects, and potentially even the GUI if it's not a pixelart look).  But rather than doing frame-based animation, we could do a system that worked based on smaller component parts that then are animated via something like Spriter (or, if need be, an in-game surrogate tool that we'll code; we did that for our game Tidalis, to excellent success visually speaking).

This route would mean that we could have very high-res art, but it obviously would have to be of such a style that hero arms and legs could move against the body as separate entities without looking strange.  So that limits things and makes anything remotely 3D prerendered right out of the question.  But it would let us keep a very high resolution on things while also having very low RAM usage.  Depending on the nature of the artist, I imagine that this could be relatively faster to do since pixelart takes longer based on the size of the image whereas traditional 2D drawing does not.

Conclusion
It might seem odd that I'm mentioning the non-pixelart options I'm considering on a pixelart forum, but I figure that it's best if I'm as upfront and clear about what I'm trying to do as possible.  Right now I'm just looking to work on some prototypes with a handful of talented artists and to see what we can cook up that would be inspiring for people who might fund us in a kickstarter.

Thanks for reading!

Chris Park
[email protected]
Founder, Arcen Games, LLC
http://www.ArcenGames.com/\

[Goncyn]

I'm really happy to see you exploring this possibility. I think you would be able to reach a much larger audience with more appealing character art, in particular. I like a lot of the background art, but the player characters look out of place and awkward. Best of luck with the search.

[eRe4s3r]

It may be worth mentioning that if the awkward animations are the main reason for this saying that pixel art animations get away with 2 to 4 frames is a bit misguided If anything, pixel art needs as much or more animation frames, because you can't do transitions with IK you need to account for hair movement, cloth movement, and assorted and it gets worse if you do transitions.

And the reason the characters look awkward is really because they have no transition animations whatsoever (and are missing some variation in movement styles ,p). So making less animations is not exactly the direction that seems to be wise. That's not saying a redo of the character art wouldn't be good, but they need proper and more animations too.

If anything the decision should made to search for a distinct more or less unique art-style, and not just pixel art vs render art. The problem is there is nearly nobody around who could do this for a 2d game anymore. I mean creating a unique (coherent throughout) art style. And as characters tend to move in a scene, scenes and characters need to fit, so to speak.

[Hearteater]

I am also very excited to hear you exploring this.  Uncanny Valley is my main issue with AVWW and the few friends I've shown AVWW to have had "meh" reactions entirely because of the look.  So anything that gets them to see the light works for me!

That said, 21 frames for running seems insanely high.  On my current project I'm working with 6-8 frames for movement and it looks pretty smooth.  And from what I've seen from various sprite exports from common games, that's pretty standard.  Have you tried cutting your running frames in half and seeing how that looks (e.g., remove all even frames)?

[BobTheJanitor]

(e.g., remove all even frames)?

Don't do this, then the running will just look odd. (*rimshot*)

Also glad to hear that you're looking for a fix for the graphical complaints. I've never had an issue with it, but I can completely understand how aesthetics can completely make or break a game for some people. It's pretty much how popcap games makes all their money. I swear they could release a tic-tac-toe game and pretty it up with shiny happy rabbits and sparkles and music and people would still buy it like mad.

[eRe4s3r]

Imo, the most important thing for animations is that they blend together. But this is about revising the graphic style for everything right? Seems like a huge investment... that said.. aesthetics do matter. And it's true characters, their animations, and the bg.. are not exactly in harmony. But it does not bother me all that much.

[TechSY730]

Hmm, so when you said art may go under revision, but later, I guess now is that "later"?
Good to see you are at least exploring options.


How about another option, use a graphic (or graphics) for each "part" of a character/monster, and the animations are defined by how they move, scale, and animate relative to one another, similar to what you do with animating 3D models.
I'm not well versed enough in computer graphics to know what this technique is called, but I know several games, like flash games, to use this technique so you can get new character animations without tons and tons of new full images, saving quite a bit on how many textures you have to keep around (though at the cost of extra CPU cost to actually move the components around as defined in the animation rather than just fetching the next part of the spritesheet)
This also lets you do cool stuff like rag-doll physics and slightly varying the animation based on things like the shape and slope of the ground you are currently on.

Also mentioned in that forum post, this would be very difficult to develop, get right, still look natural, and fit the art style of the game (though you mentioned that completely changing the art style of the game is not out of the question), so this might not be a practical option, but you could leave it open so that someone could try to take a "stab at it". Who knows, maybe someone can create such a set of parts and animation definitions that would work with game.

[Bluddy]

I think the background art generally has a good style already. My suggestions would be:

- Render all buildings together in the same environment, with a uniform sun positions or a night sky. You can render your buildings as many times as you want -- since the sun is always ever in one position, the assets will never co-exist in memory. The only difficult part is when you have sun transitions -- there you'll have to switch the assets in realtime.
- Avoid the excessive glow that exists around every object. It's too much.
- I think if characters fully animate, that'll be much more impressive than having better looking characters. What you could do to save RAM is to create polygons that aren't square.  Currently, because you're using squares, every texture has to have a lot of wasted space. This isn't what 3d games do, and it's not what you have to do either. If you make flat polygons that form a minimal convex shape around the character/monster, you can then cram a lot more frames into a much smaller space in the texture. Each frame will then have data telling it where to read the frame from in the animation. The trick is automating this process.
- Limit the amount of detail on background entities, much like the ValleyWithoutDepth texture pack does.
- Limit the background objects that can spawn together in the same area. If the materials or designs are very different, they shouldn't be allowed to spawn around each other.
- I never liked the look of the mushrooms. IMO they detract from the look of the caves. Part of this has to do with the fact that they just have too much variety, and part of it is just that they don't look that good.
- I'm glad to hear that you're thinking of limiting the maximum effective resolution. It really kills the graphics IMO to see them shrunk down and repeated.

[TechSY730]

- Avoid the excessive glow that exists around every object. It's too much.

+1 to that. What has glow and what doesn't isn't very consistent right now. Such as many unimportant things have glow, and some things that probably should have glow don't.

- Limit the amount of detail on background entities, much like the ValleyWithoutDepth texture pack does.

I don't think the detail of background entities is the problem as much as the brightness and saturation of them. Background entities should generally be "not as bright" or "not as color saturated" or related as game objects that players normally care about.

- Limit the background objects that can spawn together in the same area. If the materials or designs are very different, they shouldn't be allowed to spawn around each other.

I would agree with reducing the spawn rate of background "doodads" and plants and stuff in most areas, especially in caves.
While classifying stuff into groups of "aesthetic similarity" and aiming to only allow things from the same group or things from "aesthetically compatible" groups near each other (and keeping things from two groups that are deemed "aesthetically ugly when near each other" away form each other), that seems like a very difficult problem to figure out a good classification system, classify all the objects and art assets, and then some hard, quite subjective work to figure out which classes of aesthetics would work with each other, and which sets should be kept apart at all costs. But it might be worth it to make those occasional "piles" of background objects (which there still should be some of, but maybe not the the frequency they currently occur at) less likely to be visually, well, odd.

[x4000]

When I say fewer frames, and pixelart, I'm thinking of this: http://www.arcengames.com/forums/index.php/topic,10653.0.html

Not quite that extreme, but something more along those lines is in keeping with what we can reasonably accomplish.  Perhaps Super Mario World would be a better example of what I mean, or Super Mario Allstars.  That has more colors and so on, but still a simplicity that would look cohesive and which doesn't need more animations.

The only way that more animations could possibly happen is if we were to do something like Spriter and have multi-part animations where body parts move relative to one another in code, not in frames.  We need fewer frames than we have now, as people are already starting to have RAM problems on occasion.  Adding more frames is out of the question, so it has to be one of two radically different approaches.

[eRe4s3r]

But wouldn't it be better to just.. make a new AI War expansion? 

[TechSY730]

But wouldn't it be better to just.. make a new AI War expansion? 

Ahh, the joys of trying to factor in opportunity costs.

[Misery]

I think the background art generally has a good style already.

This.  I've always been pretty positive about the art in the game, and that's the major reason.   The backgrounds and terrain/trees/objects/whatever that make up the actual areas are, for the most part, very nice, and have a unique style to them.  I have a hard time picturing them looking any other way.   Some may say they're too bright or whatever, but that's the sort of thing that might do for an option in the graphics settings, something like that.   There are exceptions in that occaisionally I'll see a background (almost always in a building) that looks really funky for whatever reason, but most of the time, definitely liking the way all of the areas and the things in them look.   Games like, say, the original Metroid, have a certain special "ambiance" to them (hard to explain), and really, this is pretty rare in games these days;  many games dont manage this.    I really think THIS one does, though.

That being said, the main issue with the art, at least as I see it, is stuff like animations and whatnot.   The player characters obviously animate rather oddly, but they sure arent the only ones.  Like those wasps.... what the heck ARE they doing?  Or the T-Rexes, which sort of moonwalk from side to side while biting at nothing..... stuff like that.   I dont think the actual models for any of these look bad at all, but they do often animate very strangely indeed.


I do find the spell effects pretty nice, also.

[eRe4s3r]

But wouldn't it be better to just.. make a new AI War expansion? 

Ahh, the joys of trying to factor in opportunity costs.

Well as an outside onlooker I have the pleasure of just being able to watch I wouldn't want to have to decide something like this or calculate what would be the best choice.

That said I think with some downscaling of the amount of units (but not the gameplay) AI War could be made very very nice looking and that would maybe bring more people to buy that. Or hell, AI War 2.0 as a new release so all addons inside + beefed graphics with downscaled unit numbers and some game play improvements and additions like civilian traffic more point of planets and stuff, pre-owners of AI War 1 get 50% off and so on.

Well, yeah Sure am glad I can just watch what happens That said I have no idea whether this is a profitable endeavor or not. I can't really believe different graphics would help much, they would have to be clearly better. And that is a huge effort for an already released game of this size, no?

[doctorfrog]

FWIW, I still like the original art style, it kinda grew on me. But this move is definitely in the best interest of the game.

I'd nominate Auntie Pixelante. Though her style tends to look more 8-bit than 16-bit, she has an amazing way of coaxing detail and personality out of blocky sprites.

Actually, what I'd really like is to see her take a whack at reskinning AI War, even if (and perhaps, because) she'd turn the whole game into an elaborate bondage metaphor. The world needs at least one sci-fi space-based bondage RTS.