Before you start: pick a size and a purpose
Decide the final display size first. A character for a small game might be around 16 pixels tall, while a larger one might be 32 or 48. Starting at 16x16 or 32x32 is common for beginners because the scope stays manageable.
Write down three things before you draw: the canvas size, the palette limit (for example, 16 colors), and the outline style. The outlines guide explains the options. Having these fixed up front makes each later decision faster.
Step 1: Gather references
Choose a pose, a costume, and an expression. Use references to understand proportions and clothing, not to copy one image line by line. Look at finished sprites of a similar size to see how other artists handled small details.
Step 2: Thumbnail the silhouette
Draw three to five small versions of the character at the actual size, in one solid color. Pick the one that reads best. A good silhouette shows the pose and the main props at a glance. If the shape is unclear as a single flat color, the details will not fix it.
Step 3: Block in the base shapes
Build the figure from simple shapes. Use a circle or rounded square for the head, a rectangle or trapezoid for the torso, and tapered rectangles for the arms and legs. Keep the head large enough for a readable face. Put both feet on a shared ground line so the stance is clear.
Step 4: Clean up the outline
Refine the blocked shapes into the final outline using single-pixel lines. Keep the outline style consistent across the figure. If you use colored outlines, pick a darker shade of each fill color rather than black. For the rules behind clean lines, see the pixel art outlines guide.
Step 5: Choose a small palette
Pick a palette with a short ramp for each main material: skin, hair, cloth, metal, and an accent. A basic character often fits in 8 to 16 colors. Lay out the ramps before you fill anything. The palette guide covers how to build and test them.
Step 6: Fill the base colors
Fill each region with its base color and skip the shading for now. If your editor supports layers, keep the outline, base colors, and shading on separate layers. That makes it easy to adjust one part without disturbing the rest.
Step 7: Add shading
Choose one light direction and add shadow and highlight using the ramps you built. Shade the form, not just the edges. Avoid pillow shading, where every edge gets a dark outline and the highlight sits in the middle, because it makes flat shapes look puffy. The full method is in the shading guide.
Step 8: Add details
Add the face, hair, belt, buckles, and props. Keep the face the highest-contrast area, since the eye goes there first. Eyes can be one or two pixels, and fewer details often read better at small sizes. Stop when the next pixel adds nothing to the read.
Step 9: Check at 1x and with a squint test
View the sprite at actual size, then squint or view it blurred. The main shapes and the face should still read. Flip the image horizontally to check whether the proportions look balanced. Look at a grayscale version too, to confirm the value steps separate the forms.
Step 10: Export
Save a PNG with a transparent background at its native size. For why PNG suits this, see PNG or JPG for digital art. For previews, scale copies with nearest-neighbor so the edges stay sharp. Keep the layered source file as well, so you can make changes later.
Common mistakes
- Adding details before the silhouette works.
- Using more colors than the palette you planned.
- Uneven outlines and stray pixels at corners.
- Overdoing detail on a small sprite.
- Skipping the check at 1x until the end.
Once the static figure reads well, the next step is motion. See pixel art animation for beginners. For the sizing decisions behind step 1, see pixel art sizes, scale, and proportions.