How to Choose the Perfect Photo for a Custom Paint by Number
Light, contrast, subject, and resolution decide how a custom paint-by-numbers turns out. A practical guide to picking — and fixing — the right photo before you convert.
The single biggest factor in how a custom paint-by-numbers turns out is the photo you start with. A good source picture has clear light, an obvious subject, and enough resolution to hold detail — get those three right and almost any photo converts beautifully. This guide walks through exactly what to look for, why it matters, and how to fix a borderline photo in your phone before you convert it.
Start with light and contrast
A paint-by-numbers works by grouping your photo into flat regions of colour. That grouping is only as good as the separation between light and dark in the original. Soft, even light with a clear difference between the subject and everything around it gives the converter crisp edges to trace. Flat, grey, backlit, or heavily shadowed photos give it mush — the algorithm can’t tell where the face ends and the wall begins, so it guesses, and the guess is what you end up painting.
- Soft daylight — near a window, or outdoors in shade
- The subject clearly brighter or darker than its background
- Colours that are actually different from each other
- Gentle, directional light that shapes the face or object
- Harsh midday sun with blown-out highlights
- Strong backlight (a bright window behind the subject)
- Deep shadows that swallow half the detail
- A subject the same tone as the wall behind it
If a photo is a little flat, a quick nudge to contrast and brightness in your phone’s built-in editor often rescues it. You’re not trying to make it pretty — you’re trying to make the subject stand apart from the background so the converter can find the edges.
Give your subject room to breathe
The clearest paintings have one obvious hero: a face, a pet, a building, a single tree against the sky. When the frame is busy — a crowd, a cluttered room, a landscape with a dozen competing elements — the available colours get spread thin across everything, and nothing gets enough to look sharp. Before you upload, ask: what is this picture of? If you can’t answer in one word, crop until you can.
Cropping in tight is the most under-used trick here. A half-body portrait that fills the frame will always beat the same person as a small figure in a wide scene, because every colour the converter spends now lands on the part you care about. Busy backgrounds aren’t fatal — the converter simplifies them into calm blocks of colour — but they read best when they’re clearly secondary to the subject.
Resolution: how big is big enough
Detail can only survive if the pixels are there to begin with. Any photo from a modern phone camera is comfortably large enough. The ones that struggle are screenshots, images saved from social media (which compress hard), tiny thumbnails, and heavily zoomed crops. As a rule of thumb, aim for at least 1000 pixels on the long side; below that, fine features like eyes, whiskers, and text start to fall apart no matter how good the light is.
Match the amount of detail to the photo
High-detail scenes — a landscape full of foliage, a portrait with intricate hair, a city skyline — genuinely need more colours to look right, and more colours means a longer, more advanced paint. Simple subjects on clean backgrounds look great at low colour counts and make a relaxing, quick paint. There’s no single “best” setting; there’s the setting that fits both the photo and how much time you want to spend.
We cover this trade-off in depth in How color count changes difficulty, but the short version: let the photo lead. Give a detailed photo the colours it needs, and keep a simple photo simple.
The colours are pulled from your photo
One thing that surprises people: you don’t pick a palette. Wabihana reads the actual colours in your image and builds the palette from them, so a sunset painting gets sunset colours and a snowy one gets cool greys and blues. That’s also why a well-lit photo with real colour separation produces a nicer palette — there’s more genuine colour for it to find.
If you’re curious how the conversion actually works under the hood — how a photo becomes numbered regions without any AI guessing or repainting — we wrote a plain-language explainer in How a photo becomes a numbered canvas.
A 30-second pre-flight checklist
Before you convert, run through these. Most take one tap to fix:
- Light: is the subject clearly brighter or darker than its background? If not, bump contrast.
- Subject: is there one obvious hero? If not, crop in.
- Sharpness: is it in focus at full size? Blur can’t be recovered.
- Resolution: at least ~1000 px on the long side, and not a re-saved screenshot.
- Faces: are they big enough in the frame to hold eyes and features? Tiny faces vanish.
Pets, portraits, landscapes and old family photos each have their own quirks on top of this — the pet portrait guide is the best next read if an animal is your subject, since dark fur and catching the eyes are their own small art.
Frequently asked questions
Turn your own photo into a paint-by-numbers
Free to start, private, and ad-free while you paint — on web and Android.
Start painting