How to Turn Any Photo into Pixel Art

By ImgScale Team  ·  June 2026  ·  8 min read · by kizura

Pixel art has a unique visual charm that no filter can quite replicate — the deliberate reduction of color, the chunky grid, the way a handful of shades can suggest an entire scene. But converting a photo to pixel art isn't just "blur and posterize." The result is only good if you make a few key decisions correctly: resolution, palette, and dithering. This guide walks through each one.

1. Choosing the right pixel resolution

The resolution slider controls how many columns of pixels wide the output grid is. It doesn't set the output image size — it sets the pixel grid size before upscaling. Think of it as the "blockiness" knob.

• 16–40 px — Very chunky. Good for icons, game sprites, or when you want a strongly abstracted result.
• 60–100 px — The sweet spot for most photos. Enough detail to read faces and shapes, still clearly pixelated.
• 120–180 px — Fine grain. Portraits look recognizable. Still has the pixel art aesthetic at 4× or 8× export scale.
• 180–220 px — High-res pixel art. Barely looks pixelated at 1× but produces beautiful results when exported at 4× or 8×.

2. What makes a good source image

Pixel art converters work best when the source image has clear edges, high contrast, and simple backgrounds. Here's what to look for:

• Sharp photos beat blurry ones. Blurry source images produce muddy pixel clusters, not clean pixels.
• Simple backgrounds help. A portrait against a plain wall converts better than one against a busy street scene.
• Strong color contrast helps the palette work. A flat blue sky and green grass will use the palette colors cleanly. A gradient of fifty similar greens will fight you.
• Square crops work best for sprites. If you're making a game sprite or avatar, crop to square first.

The converter uses area-average downscaling (progressive halving) before mapping colors. This means a 4000 px photo will be cleanly reduced without the aliasing you'd get from a single-step resize.

3. Understanding the built-in palettes

The palette is the most impactful setting. Each palette restricts the output to a specific set of colors, which defines the "era" feel of the result.

4. Dithering: when and how to use it

Dithering is a technique that simulates color gradients using patterns of different-colored pixels. Without it, color transitions in your image become harsh bands. With it, they stay smooth.

None

No dithering. Each pixel is mapped to its nearest palette color, full stop. Best for clean icons, logos, and art where you want crisp, flat areas of color.

Floyd–Steinberg (Diffusion)

The most popular dithering algorithm. Quantization error from each pixel is distributed to neighboring pixels, which makes gradients look smooth even with very few colors. Use this for portraits and photos where skin, sky, and gradients need to look natural.

Ordered (Bayer)

Uses a fixed repeating pattern (a Bayer matrix) to add structured noise before palette matching. Produces a regular, almost screen-print look that many artists prefer. Good for landscapes and anything with large flat areas of gradual color change.

5. Converting animated GIFs

The Pixel Art Converter handles animated GIFs fully. Drop a GIF and every frame is pixelated individually, then re-encoded using a single unified palette so the animation doesn't flicker between frames.

A few things to know about GIF conversion:

• The unified global palette is built by sampling colors across all frames, not just frame one. This means a frame with a unique color might look slightly off, but the animation as a whole stays consistent.
• GIF re-encoding takes a moment for long animations. A 30-frame GIF at 100 px resolution takes about 2–4 seconds on a modern laptop.
• If your GIF has transparency, enable the Keep transparency option before exporting.
• For smoother results on animated content, try the Auto palette with 32–48 colors instead of a 4-color retro palette.

6. Exporting a sharp PNG

This is where most people make a mistake. If you save the pixel art at the same size as the pixel grid (say, 100×75 px), it'll look tiny. The export scale options solve this.

• Crisp (auto) — Picks the largest scale that fits 1100 px. Good default for sharing.
• 1× — Exact grid size. Use only if you're embedding the image in a tool that handles its own scaling (like a game engine).
• 4× — Four pixels per grid square. Good for medium-sized sharing (social, Discord).
• 8× — Eight pixels per grid square. Ideal for print and large display. A 100-column grid becomes 800 px wide.

7. Quick tips for better results

• Increase contrast before converting. A high-contrast source image maps to the palette more cleanly. Run your photo through any contrast tool first.
• Use the sprite outline option for game art. It adds a 1-pixel border around non-transparent areas, giving sprites the classic outlined look.
• Enable "Clean up noise" for busy photos. It applies a subtle pre-blur that reduces isolated noise pixels before palette mapping.
• Try the before/after compare. Drag the slider to see exactly what's being changed — useful for calibrating the resolution slider.
• For portraits, use NES or Auto with 24–32 colors. NES has better skin tone coverage than PICO-8. Auto is most accurate but loses the retro feel.
• Game Boy works best with faces. Four shades of green sounds limiting, but the luminance mapping means faces read clearly — which is why the original Pokémon sprites looked so good.

Ready to convert?

Open the free Pixel Art Converter, drop in any image or GIF, and start playing with the settings. No account, no upload, no limit.

Related: How to Extract a Color Palette from Any Image  ·  How to Reduce GIF File Size