Pekka Buttler, January 2022
The conundrum…
Sharpness is … well it depends on who you ask. It can either be the single most valuable characteristic of a lens, or it can be the most overvalued characteristic of a lens. I’ll add to the confusion, by stating that sharpness is often one of the most misunderstood characteristics of a lens.
The problem lies in that many of us (both photographers and bloggers) use terms like ‘sharpness’ with way too little nuance. For this, I blame those technologically perfect, quantifiable and real-world irrelevant lens tests many sites use.
Don’t get me wrong: Seeing a bar chart of a lens’ measured ‘goodness’ at various apertures, and being able to see one value being 3552 and another being 3701 (and thus better) offers a level of reassurance even I feel tempted by (Full disclosure: I used to be addicted to opticallimits (back when they were photozone.de) and photodo).
But a problem with such an approach to testing is that it is based on test charts of black and white bars or Siemens stars, and while terms such as ‘sharpness’ in photographing such subjects is pretty straightforward (basically an application of an MTF function), real world photography is different.
There are a lot of forms of ‘renditional quality’, such as definition, contrast, colour, etc. The first grievous injustice bar-chart/Siemens star -based testing commits is to discard colour altogether. Basically, these tests are only interested in luminosity (grayscale). The second folly is that these tests tend to conflate (or ‘lump’ together) ‘definition’ with ‘contrast’ into one overall metric they refer to as ‘sharpness’. So too, when I say ‘sharpness’, I am using the phrase as shorthand for ‘definition and contrast’, but I will also make sure to distinguish between the two.
Sharpness = the combination of definition and contrast
So, what are definition and contrast. Definition is a description of how cleanly a line is rendered, while contrast is a description of how well the luminosity differences are retained. In an ironic twist, allow me to illustrate it with some black and white bars:
As I said:
• definition describes how cleanly and precisely a line or border is rendered, while
• contrast describes retention of luminosity (differences).
Therefore, the middle sample has bad definition because the lines/borders are blurred, but as the center of the black bars is still totally black and the centre of the white bars is still perfectly white, it retains good contrast. The right sample has the opposite situation (clean lines, low luminosity retention).
Typically, either of the centre or right hand samples would not score high on MTF-calculated ‘sharpness’, but I think it is important to distinguish between them, partially as the causes for the ailments are different, and partially because the remedies also differ.
Obviously, in real-world pictures this all gets mightily more complex, because we’re not dealing with only one colour of light, and I deplore your photography if all you ever photograph has a lot in common with back and white bars…
But the basic principle of what contrast and definition are holds even there.
What causes lack of definition and lack of contrast?
There’s another significant point to consider: lack of definition and lack of contrast can be caused by very different mechanisms.
Lack of definition can be caused by a great number of ailments, but the most typical are:
• Not optimal lens design (meaning that the curvatures of the lens elements do not fully support rendering a point light source onto a point of the sensor/film). When shot wide open, this is mostly due to spherical aberration, whereas when stopped down also lateral chromatic aberrations play a role.
• Imperfect lens element manufacture (meaning that the machining of the lens elements either does not perfectly follow the designed curvature, or has resulted in uneven surfaces).
• Imperfect lens assembly or trauma (meaning that the lens elements are not exactly where they should be or are at an angle (decentering))
Lack of contrast is most typically caused by light from bright areas being refracted poorly (such as due to spherical aberrations, coma or astigmatism) or light scattering (meaning that some of the light that hits an air/glass interface goes somewhere else then it should), thus ending up where they should not. In effect, in such situations light from brighter areas pollute the entire image. Lack of contrast is most typically caused by any of the following (most typically a combination of) factors:
• spherical aberration, coma, or astigmatism
• longitudinal chromatic aberrations (when shot wide open)
• flaring/ghosting (in shooting against strong light sources),
• less than effective or damaged coatings,
• lens-internal reflections, typically caused by design.
• a severely dusty, scratched or surface damaged lens (see a case example here)
• sub-par glass mixtures used in element manufacture.
So which is more important – definition or contrast?
I cannot answer that for you. Rather, that depends on your photography. Mind you, a mild lack of contrast can often be compensated for by adding contrast in post, and some level of blurriness can be overcome by sharpening in post, but as with everything you do in post-production, there is a limit to how much blur you can sharpen and how much contrast you can ‘recover’, without introducing myriad artefacts.
Even so, and on the whole, I’d maybe say that contrast is somewhat more malleable than definition (especially, as while you can sharpen up some blurred lines, you cannot recover fine detail. That is basically lost for good).