Bringing Old Photos Back to Life: The B&W Colourisation Story

Why people care about old photos more than new ones

We spend so much time worrying about perfect lighting and composition that we forget the real power of a photograph is what it captures. A greyscale image of someone we've lost, or someone we never got to see in colour, hits differently.

When we built Cleanr, we knew photo restoration would be a core feature. But colourisation kept coming up in conversations with users. Not from designers or professionals. From ordinary people who wanted to see their history the way they imagined it. A wedding dress. A garden in summer. A child's coat.

The challenge wasn't just technical. It was about getting the colours right without looking artificial. Too saturated, and the photo feels fake. Too muted, and you've wasted the whole point of colourisation.

The technical foundation: why we chose CIColorCube LUT

Colourisation algorithms can work in different ways. Some are trained on massive datasets and try to guess. Others let you manually paint colours in, which is precise but slow. We wanted something that could deliver natural, believable results without requiring a degree in colour theory.

CIColorCube LUT (lookup table) is a proven approach. It works by learning colour relationships from reference images, then applying those relationships to the greyscale photo you're editing. The result feels grounded in reality, not algorithmic guessing. A vintage car doesn't come out neon pink. A sky comes out blue, not purple.

The real work happens behind the scenes. The LUT is tuned so that greyscale values map to naturalistic colours. Skin tones stay believable. Vegetation stays green. Shadows stay dark instead of turning muddy.

What actually happens when you hit colourize

You open a greyscale photo in Cleanr. You tap the colourisation button. And then, in a few seconds, the app has analysed the image and applied colour across it.

The process isn't magic, but it's close. The app reads the brightness and contrast information in your original greyscale image. It uses that data to determine where light falls, where shadows sit, and what kind of surfaces we're looking at. Fabric looks different from skin. Metal looks different from wood. The algorithm respects those differences.

Then it applies the colour mapping. A dark greyscale value gets mapped to a dark version of whatever colour belongs there. A light value gets mapped to a light version. This is why the results look natural rather than painted.

One thing we're careful about: the feature works best on photos with good contrast and detail. A very faded or very dark greyscale image will still improve, but it won't magically add detail that isn't there. That's honest, and it's why we always recommend running the photo through our restoration tools first if it's particularly old or damaged.

The Free tier gives you two chances a day. That was intentional.

When we set the Free tier limits, we made deliberate choices. Two colourisations per day felt right, not because we wanted to frustrate people, but because colourisation is a bit different from other tools.

Object removal? You might want to remove five things from one photo. Background removal? You're experimenting with five different presets on one image. But colourisation is different. It's usually a one-shot decision. You colourize a photo. You like it, or you don't. You move on.

Two per day means a serious user can explore a couple of old photos without hitting a wall. Someone who's discovered a stack of family greyscale photos can work through them deliberately, one or two at a time. Plus members get unlimited colourisations, which matters for people batch processing their entire family archive.

What comes next: why restoration and colourisation work together

The best colourisation results happen when you start with a clean photo. Dust, scratches, fading, and noise in the original make the colour map harder to apply accurately. This is why we built colourisation alongside JPEG cleanup and photo restoration tools.

If you're working with a really old photo, the ideal order is: run restoration first (which cleans artefacts and fading), then apply colourisation. The second tool has cleaner information to work with. The colours come out more natural.

We also included picture frames specifically because colourised photos deserve to be displayed. Six frame styles, from Polaroid to vintage film strip, give these newly revived photos a proper home. Some users frame them digitally and share them. Others print them out.