Why it exists at all
A lens has to bring different wavelengths to the same place, and doing that precisely is expensive. One reply put the cause as the gap between an idealised custom design and one assembled from widely available standard elements: precision glass can be made for a specific design at great cost, and every part swapped for a near-enough catalogue equivalent increases the aberration. Another added the reassuring half, which is that every lens has it and the amount is irrelevant when it is close to the size of a sensor pixel.
The two kinds, and which one post-processing can fix
- Transverse aberration is straightforwardly correctable afterwards, and raw processing tools have dedicated modules that handle it well.
- Axial aberration is harder. Most algorithms deal with it by smudging, so removing it costs detail. That is a trade-off rather than a fix, which is why lens designers work to minimise this kind in the first place.
Why it shares a solution with distortion
One reply drew the connection that explains why these two corrections usually appear together. Geometric distortion and chromatic aberration are separate optical faults, one being the image not laid onto the sensor at the proportions of the scene and the other being wavelengths not bent symmetrically, but both are radially symmetric around the optical centre. That shared geometry is why a single software correction can address both.
Where you will see it
- At transitions from light to dark: the edge of a building, leaves, or a ridgeline against a bright overcast sky.
- More at the long end of a zoom than in the middle of its range, which matches what the person asking had observed.
- Not from the sensor and not from a processing mistake, which were the two other causes the question proposed.
One further point about modern cameras: since mirrorless became common, manufacturers have leaned on software correction rather than on optics, because the photographer no longer looks at the lens's own image through an optical viewfinder.