Jernej LeticaPhotography Tools & Stories
Edit Photos

Turn a photo into pixel art

I turn a photo into a small colour grid, then save it at its real pixel size or enlarge it into crisp blocks.

Built and written by Jernej Letica

Browser onlyNo image uploadNo account

Choose an image to start.

The photo fits inside the grid without cropping. Empty space and transparency become white before palette matching.

How to convert an image to pixel art

I start with the shape of the subject, then reduce the colours. The preview is the grid that goes into the PNG, enlarged on screen so I can see each cell.

  1. Choose a photo from your device. I use a subject with a clear outline rather than a busy scene; the file opens in this browser.
  2. Set grid width and height. The whole image fits inside that rectangle, with white space around it when the shapes differ.
  3. Pick a photo palette or a console palette and set the maximum colour count. Turn dithering on to compare its scattered texture with flat colour areas.
  4. Read the pixel counts beneath the preview. Choose Grid size for the small original, or a larger PNG scale for visible square blocks, then download.

What I look for in a tiny image

I want the subject to read before I notice the effect. A silhouette, a bright jacket or a simple roofline gives the grid something to hold on to. Fine branches and distant faces depend on detail that a small grid cannot keep. I would crop those distractions before opening the image here.

I treat this converter as a starting point for pixel art. It samples the photo and reduces its colours; it does not decide which outline would make a better character or redraw an awkward eye. I judge the preview at a small size as well as enlarged. If the subject disappears, I increase the grid before adding more colours.

Grid size and export size

These dimensions come from the grid and export controls. Enlargement repeats each original pixel; it does not add detail or blend the edges.

GridPNG scaleSaved dimensions
32 × 321×32 × 32 pixels
32 × 324×128 × 128 pixels
32 × 328×256 × 256 pixels

Choose a grid before a palette

Grid width and height set the amount of spatial detail. Colour count sets how many tones can describe that detail. Those controls solve different problems: a larger palette cannot restore a feature that fell between sampled pixels. I use a wider grid for a landscape and a taller one for a standing figure.

The photo is fitted without stretching or cropping. Spare space is white before palette matching, so a dark console palette can turn that space green or grey. Transparent areas follow the same rule. If I want a tight square subject, I crop the source first rather than trying to remove its background with fewer colours.

Photo colours or a console look

From the photo builds a small palette from the sampled image. I choose it when the original lighting matters more than a particular game style. Game Boy uses a fixed set of greens, a familiar interpretation of its screen rather than a promise to reproduce a physical display. NES RGB gives another fixed set of colours; actual composite displays can look different, as the NESdev palette documentation explains.

PICO-8 uses the default palette described by Lexaloffle. With a fixed palette, the tool selects a subset up to the requested colour count and every output pixel uses an entry from that set. It does not enforce console sprite, tile or background restrictions. I can use the look for a picture without treating the PNG as a game asset ready for hardware.

Dithering and crisp PNG edges

Dithering distributes colour error into neighbouring cells. I switch it on when a sky becomes broad bands, then check whether the speckled texture helps. For an icon or a face, I often prefer the simpler patches with dithering off. The colour limit stays the same in either case; the arrangement of pixels changes.

The preview uses nearest-neighbour display. The export repeats the grid into solid squares at the selected scale and saves a PNG. I avoid resizing that file later with a smoothing filter, because that can soften its edges and introduce intermediate colours. For further manual editing, I keep the grid-size version as my starting file.

My pixel art checks

  • Keep the subject separate from the background before reducing the grid.
  • Compare flat colours with dithering instead of assuming texture always helps.
  • Use the photo palette when skin or natural light needs a closer colour match.
  • Save at grid size for editing and at a larger scale for sharing the block pattern.

Privacy

Your photo never leaves your device. I build the preview and PNG in your browser without uploading the image. No account is needed.

Frequently asked questions

How do I turn an image into pixel art?

Choose the image, set a small grid and pick a limited palette. I compare the outline first, then adjust colours and dithering. Download the grid-size PNG to keep its actual pixel dimensions for later editing.

Does the colour count limit every output pixel?

Yes. The count is a maximum, so a simple source may use fewer colours. With a named palette, every output colour belongs to that palette, including when dithering is enabled or the PNG is enlarged.

Should I turn dithering on for pixel art?

I compare both previews. Dithering can suggest smooth shading with scattered colours, but it can also distract from an outline. Turning it off keeps larger flat areas. Neither setting adds colours beyond the selected limit.

Why does my enlarged pixel art stay sharp?

The PNG scale repeats each grid pixel into an equally coloured square. It does not average neighbouring colours. I use this export when I want large visible blocks without the blurred edges from a smoothing resize.

Related tools and guides