Abstract
One of the most difficult things to do when it comes to art is adapting to a completely different style. Especially if you and/or your team have already settled into your own style, switching styles is the same as learning a new language – in this case, a visual language. However, when working on a team, and/or in a professional setting, everybody needs to be on the same page in order to create a cohesive end result. Luckily, if you know what to look for, how and when to dissect, test, or implement, adapting to these styles can actually become easy and fun.
In addition to just adapting to a new style, I will discuss how to combine multiple very different styles to create something entirely new. While helping to create our game Etherian Sol, I led a team, creating one such style, borne from four others – Art Nouveau and Minoan architecture, Ukiyo-E texturing, and Cyberpunk lighting.
Remember that what worked for us may not work for you. The order that steps are taken is not critical whether it be a video game, movie, tv show, or art piece – the concepts behind it will translate no matter what.
Rules
- There are no rules. Only suggestions. If it doesn’t work for you, don’t use it. (That is a rule).
- No matter how smart you or the combined intelligence of your team is, as soon as you start making things, you all become stupid. Get feedback. Constantly.
- Communicate constantly to make certain everything works together.
- Review and revise your schedule on a regular basis.
- Get feedback.
- Everything can change at any moment. Be aware of the team’s needs and adapt accordingly.
- If it ain’t work, delete it. Don’t be afraid to throw things out. When trying to create styles any one single asset can break the illusion and sometimes a piece just cannot be saved.
- Schedules are your friend. Make them and run them by the rest of your team to make sure it is feasible. Stick to them. If you find at any point that you can’t, update them.
- Get feedback.
The Actual Thing You’ve Come to Read
To start off, I will be using two terms in regards to combining styles: Layering and Amalgamating (no those terms are not official, but considering I can’t find any record of anyone else talking about such a subject in this field, take it or leave it).
Layering is the process of breaking down each aspect of a project (i.e. characters, environments, lighting, texturing, and so on) into it’s own style. In this situation, the styles do not overlap in any meaningful way. Amalgamating on the other hand means that it is impossible to tell where one style ends and another begins. The shape language, textures, lighting techniques, and everything else all have the most essential elements of every style mixed together indistinguishably.
While the technique you choose should be decided based as much on what will work best for your specific project, there are technical considerations as well. Layering is simpler to figure out and far easier to explain when in a group setting than Amalgamating. This in consequence also means that it is more reasonable for projects with shorter turnarounds. There are also middle grounds between the two that may work best for your project.
Etherian Sol was one of those cases. While there was some amalgamating when it came to the texturing and lighting, in general, we tended more towards layering, because we very much wanted our players to be able to tell the difference between the hard, ancient-Greek temple, the clean, smooth spaceship, and the almost magical effect of the Oracle Supercomputer.
Step 1: Mood Boards and Color Palettes
If the Concept Artists and Art Leads don’t spend enough time on getting references, the team will find that later in the project, one or more of the styles is not right. Additionally, some members of the team may run the risk of getting an idea of the style in their heads that is not accurate to what that style actually is, and you will either finish up and find that any player who knows the style you are trying to recreate will call you and your team out on it or there will have to be significant time invested into redoing a lot of the work. Sometimes both.
There are many different tools out there to create mood boards, but I like to use Pinterest, because there’s a very large inventory of images and I’m able to sort them into folders painlessly.
Once your team has a basic understanding of the styles, move on to creating color palettes with Adobe Color.
Switching back and forth between working on the mood boards and the color palettes allows the concept team to make a lot of mistakes early on and fix them within minutes rather than halfway through the project. After looking at references, we would realize that one of the oranges should be more muted, or that the blue I had didn’t quite compliment the reds enough, which would then go back and affect what types of images we would add to our mood boards.
Step 2: Dissecting the style(s)
In Etherian Sol, the goal was to take four art styles and combine them all into one. In order to do this though, each style needed to be dissected on its own. The concept team should spend time familiarizing themselves not only with the style they are attempting to emulate, but also similar styles and variations of the style. There are many situations where people will post incorrectly labelled images, and being able to identify this is invaluable.
About a month and a half of the way through Etherian Sol, I had to go back and adjust what I was doing, because I realized that I had some Sumi-E art instead of Ukiyo-E, and while it is easy to look back and think how boneheaded I was, it wasn’t so obvious in the moment.
Once your team as a whole can understand what the style is and is not – shape, color, texture, sizes of one object compared to another – the concept artists can begin their traditional duties.
Spoiler alert: even then, none of you actually understand the style.
Step 3: Concept Art and Texture Pipeline Prototype
This is possibly the most important step in the entire process. Concept Artists should create art for all of the styles separately, and get a hold on what makes them what they are. Very quickly, it will become clear that there is something that the concepting team did not understand with the style. Jump back to step 2 as many times as you need, because if you make a mistake at this stage, but continue on with the project, you will waste far more time in the long run trying to clean up your mess. Don’t be too concerned though. After the second or third time looking back, everyone should understand it enough to move onto the next step. Once it feels right, begin combining the most essential parts of each style.
If your team’s goal is to amalgamate the styles, there are two ways to go about it, entirely based upon personal preference.
The first option is to start with one style, incorporating parts of the other style(s) onto it one by one, making sure it works until the styles are inseparable from each other.
The second option is to build them both up from scratch at the same time. This method means more thrown away ideas, but allows you to find combinations that you may not have thought of otherwise. For example if we have “Style 1.0” and “Style 2.0” and you were to combine them in one manner, they would become “Style 3.1”. Then, if you were to go on and make a completely different combination, you would have “Style 3.2”. You might then find that you could take elements from both 3.1 and 3.2, and you then have a combination of a combination (3.21).
The more iterations of this you do, the more you will find what pieces work together best.
If you can, it might also behoove you to begin creating basic texture pipelines. The reason to do this now is because like writing down notes, figuring out the texture now will help you better remember what makes each style its own. It isn’t absolutely necessary to do it yet, but it might help. Just be aware of your team’s resources, because these texture pipelines may need to be scrapped once you start making the final art, and if your texturing/concepting team has dedicated time that they don’t have to it, that can get dangerous very quickly. For Etherian Sol, we decided to make basic versions of brick, wood, and metal, before moving on to other materials in order to test that the pipeline actually worked. The Texturing Artists should choose which materials are most important to your project and make those.
Step 4: Set Up a Diorama
The 3D Artists should now create a vertical slice of the assets that will be in your final version from each category (architecture, characters, fauna, etc.), but in the simplest form possible, while still being legible. If there is an older project already made from a previous project (it doesn’t matter what), the lighting and texture artists should work together to play around with their styles.
In Etherian Sol, we did not do this, because we had been planning to make the lighting inspired entirely by Ukiyo-E. We found out that this did not work though, and had to add Cyberpunk lighting to the list. At this time, we actually decided to allot three full day’s time to doing nothing but restructuring our schedule and figuring out if anything needed to be scrapped so that we wouldn’t be caught off guard as deadlines kept being missed. Luckily, we didn’t end up having to get rid of anything, because we were a little bit ahead of schedule as it was and were willing to add in an extra hour for the next few days to make sure it turned out well.
In general though, the lights and textures will have a profound impact on how the other looks (not to downplay the effect of the models), so it’s important that those two work together for this step. Once the 3D Artists have finished working on a piece, they can then replace the ones from the scene that the Lighting and Texture Artists have been working on (even if the replaced piece is not a 1-1 equivalent of what was there before). It’s very easy to get carried away here, but all you need to do is make certain they all look like they belong in the same style. Characters and architecture may look fine together, but the trees might look off. All of the artists should come together, figure out what the discrepancy is and fix it.
Then fix it again.
And again.
Until every type of asset works well together.
Seriously though, remember that up to a certain point, the more time you spend in these stages, the faster the others will go.
Step 5: Kill Your Darlings
I guess more like “ignore your darlings,” but don’t use any of the pieces you made before. Even if they look good, they’ll only hold you back. You deserve that all-expense-paid vacation around Europe, and they’re gonna try to stop you. Don’t let them.
At the same time, don’t full out delete them either. You want to pretend as if they don’t exist, because at step 8 (no peeking), you might find that one or two of them are actually salvageable. The team has learned a lot since the beginning of the process. The flaws from the start will still bleed through much longer than if you just create everything from scratch again, armed just with that knowledge.
Like parents always say “you’ll thank me later.”
Step 6: Asset List
Now that all of the artists have had a chance to try their hands at the style(s), create an asset list. As in any situation, really break it down into the most modular pieces you can. That way, if you find at a later point that something doesn’t work, you won’t have to delete as much. Many of these steps are simply in order to create redundancies, because when venturing out into the unknown, things will go wrong, you probably won’t have Captain Picard on your spaceship, and you’ll be wanting a save file.
Step 7: Basic 3D Modeling
The modeling, texturing, and lighting should be a cyclical process. Each of them influence each other to varying degrees, and how accurate to your style you get. Still, at this stage, have your 3D Artists start off with a good number of models. In the case of Etherian Sol (which was not huge), we aimed for 1/8 of what would eventually be in-game.
Note: These assets should now be what will appear in the final product, but they should not be finished at this stage. They should be able to be adapted to be the final asset, but not use up all of your time.
Step 8: Resurrect Your Darlings
With your whole team present, go back to the assets you “killed” before and check to see if any of them are in fact salvageable. Having had time away from them, you’ll all be able to make a far more accurate decision on that matter. Don’t be surprised if none of them do pan out, but even then, you might find that over the course of working on your project, you lost something that really could improve it.
Step 9: Color Blocking
Use only the colors from your palettes to block in the scene. This will keep you from going overboard and adding too many competing colors. Between the base colors and the silhouettes, you should be able to really see your style coming through, and be easily able to fix what doesn’t. It’s important not to put the textures on the assets yet, because if the colors are wrong, the normal maps, transparency maps, lighting, etc. will also look off.
For Etherian Sol, we were using a toon shader for this step, because in our case, we decided that we could get a better sense of what was right and wrong that way. Again, there are no steadfast rules.
Step 10: Basic Texturing and Light Blocking
Once the team is sure that the colors and silhouettes are good, begin adding the textures and lights to the scene.
Note: These should not be the lights from the diorama scene. Lighting Artists should use techniques they figured out in their sandbox before and implement them here.
Look at the settings differently than usual, because with a stylistic approach, it’s usually the stupid stuff that makes the scene.
We found that in Etherian Sol, the lower the light quality, the flatter the assets looked, which was perfect for us. With the shadow quality lowered as well, the imperfections gave the impression that it was made by a human, which also worked towards our goal. Plus, I can’t deny we were very happy about the added framerate. The light door in the basement was also discovered by accident. I had accidentally turned the volumetric fog up too high and when the light from the Oracle in the next room would hit it, it looked slightly like a door. Just a few slight tweaks later, and it became the most commented on part of our game.
Step 11: Final playtesting
As I mentioned before, your team should not be in a vacuum and should be running the style by outside sources as much as possible, but at this point – with basic models, textures, and lighting – you should be able to provide a full lens of what the final product should look like. Polish a few of the assets to a finished level (but again, always be ready to change them if you receive feedback that says they are hindering your project overall). Create a list of questions that push for more than just “yes” or “no” answers. Don’t lead them towards a conclusion by asking questions like “how bad is the flower pot?” Ask something more along the lines of “What did you think of the flower pot?” They may have liked the it. In which case, let them. Even if it’s the bane of your existence, and you want to 3D print it solely for the sake of smashing it.
On the more likely side of things though, as with anything anybody’s ever poured their hearts and souls into, everyone is probably very attached to how the style(s) look at this point. If others say that it doesn’t read well, listen to them and focus on what it is that’s causing it.
Step 12: Cleanup, Cleanup. Everybody, Everywhere.
This step is kind of a combination of all the previous steps. Use the remainder of your time to clean everything up as best you can, jumping around to whatever is most important at that moment. Prior to this step, the majority of the assets probably still won’t look the way you want. At this point though, every team member should basically be an “expert” in the same way medical students end up thinking they have every ailment known to mankind. Really though, you know the styles a whole lot better than when you first dissected them. With whatever time you have left, rework your schedule, because the hierarchy of importance has almost definitely shifted. As before, work your way down from the most to least important. Try to leave as much time for this step as you can, because everything you do now will only help in the end. Even by the end, you’ll recognize parts that don’t look like the style(s) should, but at this point, even people who are experts with the style(s) in question probably won’t even notice it.
Finally
Once the project has made it out into audience’s hands, take a little bit of time away from it, and then come back together as a team. Create a post-mortem document detailing what worked, what didn’t, and what could have been better. Just like anything else, you’ll never be perfect at it, but by acknowledging your achievements and mistakes, you will do exponentially better each time until it becomes second nature.
Even in writing this article, after having done multiple projects that didn’t work, then some that did, and getting better each time, I know that I will look back 5 years from now and say “Really? Why did I put it in that order?” That is simply the nature of trying to lock creativity down with a systematic approach. It can’t be done, and there is a lot that will only be able to be done through experience and instinct.
Conclusion
Mood Boards and Color Palettes
Make ‘em and spend a lot of time making sure they’re right.
Dissecting Your Style(s)
Look at your mood boards and color palettes. Then go back and edit them to make sure they’re even more right. Let them affect each other.
Concept Art and Texture Pipeline Prototype
Test out the 2D aspects of the project
Set up a Diorama
Use a scene from a previous project to allow the lighting and texture artists to test their setups. 3D Artists should add basic assets to the scene to make it more accurate to how the final product should look.
Kill Your Darlings
Get rid of all your assets, and textures, and such, and use your now somewhat knowledgeable brain to start from scratch. FOR EVERYTHING GOOD IN THIS WORLD THOUGH, DO NOT DELETE YOUR STUFF.
Asset List
Create an asset list of everything that needs to be made.
Basic 3D Modeling
The 3D Artist should begin making models to appear in the final game. These should not be considered final once this step is done though.
Resurrect Your Darlings
Go back over the assets that were discarded before. Check to see if any of them are salvageable or if anything could be gleaned from them that had been lost over the course of the project.
Basic Color Blocking
Add only the colors from your palette, to make sure that everything reads.
Basic Textures and Light Blocking
Begin filling out the details with textures and lights. Do not go overboard with either yet though.
Final Playtesting
This is likely the last time you will get a full playtesting opportunity. Create questionnaires and Polish a few of the assets to check that your audience will like the end result.
Cleanup
Remake your asset lists to pinpoint what is most important. Do whatever must be done to complete the project. Show it. Be proud.