When I browse the popular photography web forums these days, one of the most common topics discussed is lens "sharpness" as if this element alone is the Holy Grail and most important factor, to obtaining great image quality. So, with that in mind, I thought I would offer my own personal take on this intriguing and rather complex subject.
First off, I think it's important to establish just what sharpness is. Technically, I believe it simply refers to "line/edge definition" and nothing more.
In a truism sense, sharpness per se, doesn't really exist because all objects in the universe have their own inherent, natural form and line/edge definition. They are simply "as is" and that will include a wide range of variants between the so called "soft" and "sharp" position. In fact, many things in the natural world are not sharp at all, but instead actually have a rather rounded edge, especially when viewed at a very close range, such as during macro 1:1, or closeup photography. Not everything is 'razor sharp' with clearly defined crisp lines and detail. Take a closer look at a leaf, blade of grass, or flower petal under a magnifying glass and you will see that in reality these objects are often quite fuzzy edged with little or no highly contrasting dark outline to separate the image from other surrounding parts of the photograph. This current concept that everything should be 'uber sharp' is merely the result of our modern day HD (high-definition) mentality, which now makes us come to expect that every single thing we view must have that big screen, IMAX type persona in order to be good.
There are two primary, inter-woven (but conceptually very different) core mechanisms that work together in tandem to create what WE commonly call - sharpness.
Glass as a material (in this scenario) is merely a 'facilitator' of light. It's a transparent medium that either allows 100% image DATA to pass through it onto the camera sensor, or it doesn't. Therefore it's most important quality is "optical clarity" (or visual acuity) and not sharpness. Great line/edge definition (ie: sharpness) is merely the result of optimal clarity that permits a 100% image capture, or close to it as possible.
Another factor that should be mentioned here is MTF or Modulation Transfer Function which is the “spatial frequency response” of an imaging system or component . I won’t go into this aspect here, simply because it is a rather complex, technical topic and beyond the intended scope of this blog. Anyone who is really interested in the mathematics and concept behind MTF can click here for detailed information. In lay terms, it represents the correlation between resolution and contrast which are inseperably linked.
Many folks seem to be confused between the traditional (manufactured) convex lens mechanism along with it's related distortion or aberrations which concentrates the central focus point, and the inherent glass material properties that enables the image DATA transfer to take place. These are two inter-twined, but very separate components within the same optical environment. Glass itself, as a vehicle, has no 'sharpening' abilities, it can ONLY allow a true flow of information, or limit/reduce that maximum potential, because of various existing factors, such as: distortion, etc.
What we (as photographers) really want/need...is not a sharp lens, but a 100% accurate or realistic lens. One that allows us to capture exactly the same image that is presented by the subject. However, a lense cannot form a perfect image. There is always some degree of distortion (aberration)introduced by the lens which causes the image to be an imperfect replica of the object. A poorly manufactured lens system will obviously produce a higher amount of optical aberrations and result in a compromised performance. However, even IF the transmitted ray convergence was absolutely perfect, image quality would still be diminished if the conduit material itself (ie: the glass) was also non optimal. Therefore, if aberrations will exist to some degree no matter what - then it is even more crucial that "clarity" of the transfer medium is maximized to offer the best possible image quality.
In the new digital realm the signal chain as a collective whole determines the final end result. The camera sensor, internal settings, parameters, focusing mechanism, lens optics, external shooting conditions, and not least - the post-processing software algorithm conversion & representation, all play a combined role in creating the finished image. The lens is merely one component in this chain, whereas all elements have their own individual influences, etc. It's no different than an "audio" signal created by an electric guitar...which passes from the strings (ie: the source)...along the chain via the various vehicles that make up the output sound or tone. Each step has it's own unique contribution to make, the strings, pickups, compressor, stomp-box, preamp, amplifier, speaker cabinet, and so on.
There is also another part of this equation to consider here. Take Nikon for example. All Nikon camera's by default produce a relatively 'soft' image straight out of the box (regardless of which lens is attached) due to the (AA) Anti-Aisling filters when no additional sharpening mode is applied. In other words, no Nikkor lens will produce a tack-sharp image all by itself, unless one either turns the in-camera sharpening setting up from zero, and/or adds sharpening later on (via unsharp mask, etc) during the processing stage in software. This is just as Nikon intended it to be. The image line/edge definition must be enhanced/adjusted/corrected through digital means, in order to obtain a sharp end result.
Please note - that “line-edge” definition is the primary element involved in our human perceptional experience of "sharpness”…so, that’s why USM was invented, to re-boot this specific component…and to compensate for diffaction when it starts to kick in. In short, one can take a shot at, say, f/22…but using USM, bring the line-edge definition quality back to an f/11 level. Obviously, we cannot re-capture any fine detail (data) that wasn’t actually recorded in the original take, but we can restore the line’edge definition of ALL detail that does exist. Detail is obviously important within the entire image context (especially with regard to object recognition and awareness) but it is secondary with regard to our perception of sharpness. More detail, simply means more information (ie: stuff)…however, it’s the “definition” (or optical clarity) of that information that is key.
Here is an example to demonstrate this: This shot was taken with the Nikon D300 body and pro grade Nikkor 17-55 2.8 lens. They are both 100% full size crops. The top picture is with a moderate amount of sharpening applied in post-processing software. The bottom photo is with no sharpening. As you can see, the photograph straight out of the camera without any sharpening added is noticeably "soft" in comparison.
Dahlia Flower 100% Crop - With Applied Sharpening
Dahlia Flower 100% Crop - No Sharpening
Please note - that I always use a tripod, shutter-release and make sure that the weather conditions are optimal when shooting flowers outside. IF I had taken this shot hand-held, it is obvious that the original raw NEF image (without any in-camera sharpening) would have been even less sharp than what I managed to capture here, despite using one of the best lenses that Nikon has to offer.
Click here Dahlia Flower to view the sharpened image at a normal viewing size.
In conclusion, I would say that a combination of sensor capability, lens structural performance (including glass-property data flow), software interpretation, and external shooting environmental factors - all play a role in creating image sharpness. Until we are able to produce a transfer-material with a full 100% accuracy, and a ray convergence mechanism with absolute zero aberrational distortion - this subject will always be of interest and highly contested.
And besides, does it really matter? Our human visual system is far from perfect. Most of us have a compromised acuity of sorts, various astigmatisms, myopia to some degree, and therefore we wouldn't actually be able to 'see' (ie: perceive) or appreciate that completely flawless image, anyways.
© Kev Vincent Photography. Copyright 2012. All Rights Reserved.