How to colourise old black and white photos

Old family photographs can feel distant when you have only ever seen them in black and white. Adding colour can make familiar faces and places easier to connect with, but it also raises questions: which tool should you use, how much preparation matters, and can you trust the colours? This guide explains how to colourise old black and white photos using AI or manual editing. You will learn how to prepare a clean scan, assess the result and keep the original safe. Whether you are making a family gift or preparing a heritage post for your business, the same principles apply.

How to colourise old black and white photos realistically

Colourisation adds an interpretation of colour to a monochrome photograph. It does not uncover the original colours hidden inside the file. The image records brightness and contrast, but those values alone cannot tell you whether a dark jacket was blue, brown or green.

AI colourisation tools use learned visual patterns to suggest colours for objects and surfaces. A human colourist does something similar, while also drawing on research, reference photographs and personal knowledge of the scene. Either method can produce a convincing image without producing a historically accurate one.

What AI can and cannot tell you

Recognisable features such as foliage and open sky can offer useful clues, but lighting, seasons and surroundings still affect their appearance. Skin tones also need careful review: a plausible-looking result is not proof that the tool has represented a person accurately.

Clothing, painted signs, cars and interiors often leave more room for guesswork. If accuracy matters, gather supporting evidence before editing. A surviving uniform, a colour photograph of the same building or a relative's recollection may help you identify choices that need correcting.

Aim for a believable interpretation that respects the source, rather than the brightest possible result. Keep the black and white version available so viewers can distinguish the original record from the later edit.

Prepare a clean scan before adding colour

Start with the best copy of the photograph you can access. A screenshot of a compressed family-chat image gives a colourisation tool less useful detail than a fresh scan of the print. Colour cannot compensate for a face that is already blurred or obscured by glare.

If you have a scanner, clean its glass and place the print flat without forcing fragile material. Choose a scan setting that captures fine detail, and save a high-quality master copy. Avoid repeatedly overwriting it as you experiment.

A phone can also work. Put the photograph on a flat surface in soft, even light, hold the camera parallel to it and check for reflections. Avoid direct flash, which can create a bright patch that hides facial features or clothing detail.

Preparation checklist

• Save an untouched master scan or photograph.
• Straighten the working copy and crop distracting borders.
• Check that faces are sharp and shadows retain detail.
• Address noticeable scratches, creases or fading before colourisation.
• Avoid aggressive smoothing that removes texture or changes faces.

Damage can confuse the colourisation process. A crease across a cheek, for example, may acquire a different tint from the surrounding skin. For a practical preparation workflow, read how to restore old family photos.

For valuable or delicate originals, consult a photographic conservator before attempting physical cleaning or repairs. Digital editing should happen on a copy, not at the expense of the original print.

Choose between manual editing and AI colourisation

Manual colourisation makes sense when you want control over specific details. In an editor such as Photoshop or GIMP, you can separate skin, hair, clothing and backgrounds into layers, then apply colour while retaining the underlying light and shade.

A common approach uses a Colour blend mode, soft brushes and masks to keep each adjustment inside its intended area. Lowering opacity can help prevent colours from looking painted on. Detailed boundaries, such as hair against a background, need particular care.

The trade-off is effort. A simple portrait is usually easier to manage than a busy wedding photograph with many faces, fabrics and flowers. Manual editing rewards patience and familiarity with masks, layers and subtle adjustments.

AI is a practical starting point when you want to colourise family photos without learning those techniques first. It can produce an initial interpretation quickly, leaving you to assess whether it suits your purpose.

You can also combine the approaches: use AI for the starting version, then make selective corrections in a separate editor. Choose manual work when documented colours matter; choose AI when convenience matters more than precise control. Neither approach removes the need to review the result.

Follow a simple colourisation workflow

Work on one photograph first rather than processing your whole collection immediately. This lets you judge the preparation needed and decide whether the style of result matches what you want.

1. Choose a suitable image. Start with a clear portrait or an uncomplicated scene.
2. Prepare a working copy. Straighten, crop and restore visible damage where needed.
3. Upload it to a colourisation tool. Check the tool's file requirements before uploading.
4. Inspect the result closely. Look at faces, hands, clothing edges and background objects.
5. Compare with known details. Use family references or other evidence where available.
6. Save separate versions. Keep the original scan, cleaned image and colourised result clearly labelled.

Example: a family shop photograph

Imagine you are preparing an old photograph of your grandparents outside their shop. You want a framed gift and a copy for the business's anniversary post. The scan has a crease across the doorway, while the shop sign appears as a dark grey rectangle.

First, address the crease on a working copy. After colourisation, check faces and the doorway for uneven colour. If the tool makes the sign red but a surviving sign shows it was green, correct it in an editor or make clear that the colours are illustrative.

This workflow helps keep a plausible AI choice from becoming an accidental historical claim. It also gives you a reusable process for the next photograph.

Where Pitanga Labs fits

The Pitanga Labs Colourise Black & White Photos tool adds natural, context-aware colour to scanned photographs for 3 credits. It accepts PNG, JPEG or WebP files up to 10 MB, with results available to download in minutes.

The practical benefit is a simpler starting point: you do not need editing software or layer skills to create a colour version. Prepare your image, upload it and review the output against the original and any reference information you have.

Context-aware colour is still an interpretation, not verification of the scene's original appearance. Check faces and meaningful details before using the result in a family archive, public post or printed gift.

Start with one representative photograph before committing to a larger collection. That helps you judge whether the result suits your needs and whether other images need more restoration first.

Check, print and share the finished image

Review the photograph at both normal viewing size and close up. A result that looks attractive on a small screen can still contain distracting colour patches around fingers, glasses, hair or clothing boundaries.

Watch for oversaturated skin, inconsistent colours across the same object and backgrounds that draw attention away from faces. If you edit further, make restrained changes and compare frequently with the monochrome version.

Before ordering an enlargement, check the printer's image-size requirements. If your file is too small, how to upscale a low-resolution image without losing quality explains the options and limitations. Inspect any enlarged version carefully rather than assuming extra pixels mean recovered detail.

Keep a full-resolution copy and backups in separate places. When sharing, add a short caption such as “Digitally colourised; colours are an interpretation.” For business use, confirm you have permission to publish the photograph and consider the privacy of anyone pictured.

Conclusion

Good colourisation starts with a clear source image and realistic expectations. Prepare a clean working copy, choose AI or manual editing to suit your needs, and review the result rather than accepting every colour as fact. Preserve the original and label the edited version when sharing it. To get started, choose one meaningful photograph and try Pitanga Labs' Colourise Black & White Photos tool. Check the result carefully before moving on to the rest of your collection.

FAQ

Can AI recover the original colours of a black and white photo?

No. AI predicts plausible colours from visual patterns. Use reference photographs, surviving objects or reliable records when particular colours need to be accurate.

Should I restore a photograph before colourising it?

Usually, if it has visible damage. Scratches, creases and heavy fading can interfere with colourisation. Restore a working copy and preserve the untouched scan.

Can I colourise an old photo taken with my phone?

Yes, a clear phone photograph can be a useful starting point. Keep the camera parallel to the print, use even lighting and avoid glare or direct flash.