Deficient optics

Fixing it in post

The human eye is not a great imaging system.

Our astronomer is currently adjusting to a new set of eyeglasses.  Glasses by themselves constitute nothing new; he’s been myopic since his teenage years.  Over time his eyes have changed slowly, so that new prescriptions have been necessary now and then, though ordinarily shifting from one set to another has been unnoticeable.  This time, however, he has progressive lenses.  For the sake of young and clear-sighted readers, we should explain.

The human eye focuses its image on the retina as any simple lens does.  However, while cameras and telescopes adjust the distance between the lens and sensor to compensate for different distances to the object, the eye does it by changing the curvature of the lens.  As one gets older, the shift in curvature (“accommodation”) gets harder, and either distant objects or close objects get harder to see.  Our ancient astronomer needed extra help at both ends of the scale.  Tired of carrying one set of glasses for distances and another for reading, he ordered a set of progressive glasses: the top of the lenses are for distances, varying to reading glasses at the bottom.

They’re actually something of a technical achievement.  But mostly our astronomer has been noticing imperfections.  Looking at his vision field as a whole he sees astigmatism, spherical aberration, pincushion and barrel distortion, and things seem to shift their position as he turns his head.  It will take some getting used to.

However, we realize that any optical system that tries to cover such a range of distances and over such a wide field is bound to fall short somewhere; as the human eye does.  As an optical imaging system, the human eye not really very good.  It only sees fine detail and color over a limited field at the very center of vision, and the lens would be laughable if it were installed in a camera.  The only way our eyesight is at all useful lies in the software: our visual processing center somehow combines a series of poor images to produce the illusion of a fine landscape.  Photoshop and AI don’t even come close to this achievement.  You don’t notice the flaws because you only see the picture after post-processing.  In a very real sense, there are no raw visual data.  And this is before we start wondering how you recognize your sister at fifty yards, or react quickly enough to hit a baseball.

This is actually rather encouraging: by being clever, we can overcome grave limitations in our tools.  On the other hand, there are optical illusions.

Share Button