Got inspired by these little guys. Spent a few hours pounding out a small mockup (~60 unique tiles)

Not happy with it, but it was definitely fun.



Anyway, back to proper work. Cheers.

Here's my full sketchpad if anybody wants to play with these. Autotile format is RMXP 47-state autotile.

Zecronious wrote:

By the way my 'owl' is actually meant to be a run down square style fort.

Yeah, I thought so. It took me a minute to recognize what it was, but the flag and the door clued me in. My girlfriend walked up to talk to me before she went to work this morning while I was fiddling with the first post and she asked me why the owl was so big compared to the little guys. I threw it in as a bit of a joke, but sarcasm, text, and programmers don't mix.

http://i.imgur.com/lXi4wGm.png

^This is three years worth of unsorted pixel art (This is only about 2/3rds of what I've made, as the other third was paid for work with exclusivity agreements.)

Very little of that work uses a custom palette. You can go an awful long way with a limited palette.

I really started to improve when I stopped relying on photoshop like a scrub and started trying to understand the medium.

I eventually have come to the realization that 2-6 colors per sprite is really the ideal for the style that I want to portray.

Of all the work I've done, the biggest improvements have been from coming to understand Dawnbringer32, 16, and gameboy 4-color palettes.

I strongly advise against making your own palette at first.

As for how to use a palette, you basically can take a palette can create color ramps for specific materials. 2-4 colors per material is often enough.

Dawnbringer was kind enough to create some neat tools for selecting your color ramps:



There's a TON of information in this graph, but it should more or less help you in creating color ramps for your art.

Usually, I start out by getting the form right. A good silhouette rarely turns into bad art. Just use one or two colors to get the silhouette right. After you've got the silhouette, start working in color areas in a midline color. After you have the areas done, you can start shading and detailing. At every stage you should tweak and zoom in and out to ensure that your graphics aren't too busy, and aren't too plain.

If you are ever tempted to add random noise: DO NOT. Random noise is something that should be avoided.

As for shading, try to avoid what I did with those mountains. That's called pillow shading, and it's ugly as all sin.

Banding is another thing you want to avoid. Try to avoid pixels sort of grouping together in different shades in sequences. The more mechanical your actions feel, the worse the outcome will be.

As for colors, use medium saturation for your baseline colors, and low saturation for shadows. Tinting your color's hue toward blue will really help with shadows.

There's some interesting tricks I've learned in recent months that have been a big help. Don't always use like colors. The greens in dawncaster shouldn't be used in sequence. Instead, mix in the browns, teals, and greys and maybe even some of the blues to create good vegetation color ramps.

Do not be afraid to leave huge areas untextured. Plain colors look great in pixel art if you use them right.

Too much detail can be really bad. It can make the art look super blurry.

Learn tropes. Look at what other artists are doing. Some genres have specific patterns in how they draw certain things. Pay attention to the tropes. Your art looking like something else is good, not something to be afraid of.

Stay away from dithering as much as possible. It's good if you know how to use it, but unless you are trying to create a smooth gradient and eliminate banding, it can look super bad. See my art for examples of when dithering and noise don't work.

Also, shading isn't as simple as "light on left, dark on right.". Try using a flat, light color for the tops of objects, rather than doing any shading at all. Try to capture the geometric shape, not just adding color.

Yut Put is also a wonderful resource for learning to pixel. He's taught me a lot through criticism and showing me his work. Luckily for me, our two styles and tastes have been converging toward one another for some time, though he is far more skilled than I am.


If it seems overwhelming, that's okay. It takes time. Take a step away. Draw a happy face, write "orc" on it, and put it in your game. Do some zen programming or just take a break. It takes time and dedication, but nobody gets anywhere but the psych ward by burning themselves out all day every day. If you never finish anything, no big deal unless it was a paid job for a client. Just try to enjoy what you are doing and find what works for you and also tickles your fancy.

Don't try to measure up against other artists. Try to measure your STYLE against other styles you want to imitate. If someone is better than you are at your chosen style, examine, study, imitate. If someone is technically a better artist than you, but their style isn't what you are going for, learn what you can but don't get hung up on a style you can't imitate.

Understand that very detailed/large art takes a long time. The more small variations you make on themes, the longer it's going to take. Be realistic. You aren't going to be able to do everything you want and that's okay. Just get the basics done, and come back to the details later when you have time and nothing priority going.

Don't bother with tutorials unless you are going to draw. If it takes time away from your practice of art, it's not going to help you. The number one thing you can do is draw. Draw a lot. Don't worry too much about feedback. Learn to self-appraise and find areas where you could improve. Look at tons of reference images and get feedback from people you trust. Feedback from strangers is worthless about 80% of the time, because most people don't actually understand the subtleties of the medium, or lack a common point of reference.

The very last point I want to make about pixel art is that you shouldn't think about it as realism. Pixel art is capable of carrying realism, but the time cost is super high. Focus on impressionism. When you get really good, you can mix in some surrealism and get away with it. The best pixel artists (in my opinion) tend to fall somewhere between Dali and Van Gogh in inspiration and form. Don't be Michelangelo. Be god damn Van Gogh.

Believe me, if you could have seen my art 15 years ago when I first started... It was bad. I only got better by forcing myself to learn and by looking at what others were doing. Tutorials didn't do me a damn bit of good. Looking at other artists' work and figuring out what the hell they were doing was the biggest boost. But yeah, from one programmer-who-forced-himself-to-art to another, good luck and godspeed.

Man even after just applying a few of the simple rules you proposed my work is looking much better I think. It's not amazing and still crappy but not super-mega-deluxe crappy.





Lots more to go but I felt like this looked promising. I'm on limited time right now so I'll keep going slowly.

Have to agree, those are much better. Especially the owl ;) The more you practice and study the theory, the better you'll get.

So this afternoon I kicked back and did some work on this Viking guy. Obviously I don't want to just straight up copy you Ter so I took what I had and reformed it, applying the new rules.

I focused on form because of how important I'm realising that is. He was originally supposed to be muscular but I'm happy with where this took him.



~ Using Dawnbringer

Oh and big thanks, I owe you. When the student is ready the teacher will appear I guess.

I'm late to this thread, but good gads is that palette inspiring. Thanks for sharing that.