Why a crowded pattern makes the numbers harder to read
A digit you can read easily in open space becomes unreadable when packed between other digits. That is visual crowding, and it is why scanning a dense pattern out of the corner of your eye does not work.
Here is the thing that makes no sense at first. A 4 sitting alone in a wide background area is perfectly legible. The same 4, the same number of pixels tall, wedged into a lattice of tiny regions each with its own digit, is a smudge. Your screen has not changed, your eyes have not changed, and the character is exactly as large as it was. Something else is going on, and it has a name.
Visual crowding
Crowding is one of the sturdier findings in vision science: an object that you can identify perfectly well on its own becomes much harder to identify when other objects sit close beside it. It is not a matter of the target being too small — you can demonstrate it with a letter that is comfortably above your acuity limit — and it is not blur. The target is resolved; it just cannot be told apart from its neighbours as a distinct thing.
The usual description is that features from the flankers get pooled in with features from the target, so what reaches your recognition machinery is a jumble with the right ingredients in the wrong arrangement. People report exactly that: not that they see nothing, but something with the right texture which refuses to resolve into a particular digit.
Two properties matter here. Crowding depends on the spacing between the target and its flankers, not on how big anything is — push the neighbours away and it releases. And it gets dramatically worse the further the target sits from the centre of your gaze. In the very centre it is almost absent. A few degrees out it is severe.
Why you cannot scan with the corner of your eye
That second property collides with a fact about the eye that most people know abstractly and few apply. Only a very small patch at the centre of the retina has high acuity. Away from that patch, resolution falls off fast, and it falls off much faster than most people expect, because the visual system papers over the gap so convincingly that peripheral vision feels detailed when it is not.
The practical consequence is that looking at something means moving your eye to it. Having it on the screen is not looking at it. When you try to scan a dense pattern for the next region marked 5, what you are doing is a rapid series of jumps, sampling one small patch at a time, and any digit not currently in that patch is both low-resolution and crowded. Both effects stack in the same direction.
This is why the search feels unaccountably slow, and why people conclude their eyes are getting worse when they are simply working outside the range where the technique they are using can function. Peripheral scanning genuinely does not work in dense pattern work. It is not a matter of effort or practice.
What actually helps
Zoom, because it changes the spacing
Zooming in is the direct fix, and it is worth understanding why it works so well, because the reason is not the obvious one. Making the digits bigger helps a little. Making the gaps between them bigger helps a lot, because crowding is governed by spacing. At twice the magnification, a neighbour that was two millimetres away is four millimetres away, and the interference falls off sharply with that distance.
Cover the neighbours
The crude version of the same idea: put a fingertip or the edge of your other hand over the digits next to the one you are trying to read. It looks silly and it works immediately, because you have removed the flankers rather than moved them. It is a useful diagnostic too — if occluding the neighbours makes a digit readable, the problem was crowding and not your vision, and zoom will solve it properly.
Stop reading digits altogether where you can
The best answer to a crowded field of numbers is often to not read numbers. Most apps mark every unfilled region belonging to the colour you have selected, which converts the task from reading small text to spotting a shape — and shape detection in the periphery is far more robust than character recognition. You get to use the peripheral vision you actually have rather than the peripheral vision you assumed you had.
Where an app offers a control that jumps to the next unfilled region of the current colour, that removes the search entirely for the awkward last few, which are usually the smallest and worst-crowded ones in the picture.
Choosing artwork with this in mind
If dense work is a struggle rather than a pleasure, the fix upstream of all of this is to pick pictures with larger regions. Density is a property of the artwork, not a difficulty setting, and no amount of technique makes a lattice of hair-width regions comfortable to read at fitted size.
- Look at the thumbnail and ask whether you can see individual regions in it. If it reads as texture rather than shapes, every digit in it will be crowded at fitted size.
- Geometric and mandala artwork varies enormously — some has a few hundred large regions, some has thousands of tiny ones, and the thumbnails look similar.
- A bigger screen is a real fix rather than a workaround: more physical millimetres per region means more spacing, and spacing is the whole game.
One last note, since this page is about eyes. Crowding is normal vision working as designed, not a defect. But if digits have recently become harder to read in ordinary conditions, or the effect is much stronger in one eye than the other, that is a question for an optometrist rather than for a colouring app.
Frequently asked questions
Why can I read a number alone but not the same number in a dense area?
Because of visual crowding: an object that is identifiable on its own becomes much harder to identify when close neighbours flank it. The effect depends on the spacing to those neighbours rather than on the size of the target, which is why the same digit at the same size can be legible in one place and not another.
Does zooming help because it makes the numbers bigger?
Partly, but the larger effect is that it increases the spacing between a digit and its neighbours. Crowding releases with spacing, so doubling the magnification doubles the gaps and the interference falls away sharply. That is why zoom often works when leaning closer did not.
Why does scanning a pattern out of the corner of my eye not work?
Only a small central patch of the retina has high acuity, and crowding gets much worse away from the centre of gaze. Both effects push the same way, so a digit you have not directed your eyes at is both low-resolution and jumbled. Searching by colour highlight rather than by reading digits sidesteps this.
Is this a sign my eyesight is getting worse?
Not by itself — crowding affects everyone and is normal vision working as designed. If reading has become harder recently in ordinary conditions, or one eye is clearly worse than the other, that is worth an optometrist appointment rather than a change of app.
Try it on a real canvas
Numbrush is colour by number for adults: mandalas, rose windows and geometric pattern work, twelve colours a picture, a magnifier for the tight spots, and a wrong tap that costs you nothing. Free, offline, no ads and no account.