I am as guilty as any blogger of using the words sharpness, resolution and detail too many times in a week. Part of the obsession some of us have with quality, that is perceived, numerically or visually measurable quality, comes from the it's elusiveness in our early years.
Photography went from "weird science" to controllable hobby in a relatively short time (if you call a human generation short), but it went off like a bomb when it did take. All were in agreement; the technical side of photography was as much a part of the whole as any other part, allowing gear nerds, working pros, experimental or classical artists and enthusiasts to share the same journey, if taking different paths.
The second great wave of quality obsession came with the digital emergence. Starting from scratch, but with a pre determined quality bar to aim for and tools readily available for determining value, we jumped in with both feet, measuring micro differences between lens X and sensor Y, until the differences needed better equipment than our own eyes to even see those differences.
So, what is sharpness and is it that important?
Photographically, sharpness is the ability of the end product to convey a perception of clarity and definition using true edge and/or micro contrast (Canon's Lens Work publication explains this best, showing the differences between contrast and resolving power).
All sharpness is an illusion. The flat surfaces used to view images have no actual sharp edges, you cannot cut yourself on them (ok, a smashed screen or a nasty paper cut can disprove that point). The impression of sharpness comes from the clarity and strength of contrast used to show transition from one tone or colour end point to a neighbouring one.
In a dark room, this was called acutance and was often a chemically created phenomenon, so I guess it had actual substance, but even if digitally created the effect is the same. The sharpening tool in your processing suite is simply exaggerating the strength of the transition point between two areas, which is why over sharpening looks so bad as it is creating a clearly false (oversized) edge rather than a better defined edge.
How do we see sharpness?
Simply put, we are tricked into seeing it. Viewing distance is important, display media and viewer expectations also, but trickery is the name of the game. If you decide that "perfect' sharpness is your goal, that is sharpness so pure and definable, you cannot perceive the point where it is no longer sharp, them you have to stop looking before you exceed your point of maximum acceptable perception. That is, you must decide when you no longer want to look any further (closer), because eventually, if you get close enough, you will see the point of un sharpness.
Very big prints from 50mp full frame cameras will have a realistic resolution limit, that is not far past 16mp m43 sensors in real terms. Neither is perfect, but either is capable of producing fine art work to reasonable sizes. The reality is, M43 is not a revolutionary "catch up" leap to bigger sensor size quality through some exclusive, advanced technology. There is no magic juice. What has happened is a new wave of very sharp* sensors and supporting lenses has coincided with "already enough" quality and is acting more like a reality call than a real quality shift.
*Anti aliasing filters removed or reduced from the sensor path, increasing perceived sharpness.
Ming Thein did an exercise a couple of years ago where it tried to get unrealistic and impractical "loupe viewable" sharpness on a print. The process needed 36mp's, perfect technique and multiple passes of very fine ink droplets from a pro printing press onto carefully chosen paper to render that much detail and even then paper size was limited to small gallery sizes only. In his own observations he basically said no one will care except other photographers interested/obsessed by quality. It is all about viewing distance, expectations and delivery.
Example; look at your TV. look at it from across the room. Decide there how sharp the picture looks. Then get closer and closer, until the actual pixels making up the picture are more obvious than the picture itself. Some time in the middle, you will feel the image has become possibly sharper, but gradually unsharp. Next the actual pixels start to look sharp, but as you get even closer, those pixels loose their sharpness also. A couple of times on the journey you are happy with your perception of clarity and a couple of times not.
Another example, closer in relevance to the subject at hand.
The image below was taken with care a few years ago with a crop frame SLR, sporting plenty of pixels, a solid wood tripod and a better than average lens.
Eos 550D 15-85mm at 85 f8, at about 30 seconds. Important to remember here that in print, this looks sharper as the paper texture and physical ink work together.
Up to reasonable sizes (11x14 inch) and equally reasonable viewing distances the image looks sharp and contrasty in print, especially on more precious, heavy matt-fibre paper. Viewers have responded well to this series of colour and mono images, but the dirty secret is, none of the images were very sharp.
Their response was to the colours and tones, the location and the composition. If I had managed better technical quality would it have really mattered to the average viewer or would that simply have made me happier with the image?
I have narrowed the culprits down to vibration from the quite heavy sea crashing on the beach at the feet of the tripod (amazing what a long exposure can hide), possibly combined with the tripod very slowly sinking into the soft sand and maybe a less than brilliant lens (some obvious CA issues), but either way, the whole evening was less satisfying than desired. At first I thought I had left the stabiliser on. Never a good idea when using a tripod, but that would have been too easy.
A crop of the above.
Technique was tried and tested, gear should have been ok, but the images lacked any real "bite" at 100% on a screen, so most of the perceived sharpness came from added contrast, print stock selection and by limiting size.
Another example of perceived sharpness compared to actual sharpness below.
Taken with an ancient Pen 25mm from the 1960's, shot wide open, where it's not at it's best, the image looks soft and muddy.
The second image has had contrast added to it globally (clarity and darkened blacks) and specifically to the camera body (the brush tool with slightly increase clarity, de hazing and contrast). All of the clarity you see comes from applied contrast. Reasonable sharpness was there, just hidden. The lens actually showed a little more 3D pop than the newer lens it was compared to and slightly unusual (old school) bokeh.
The lesson learned here is; actual sharpness is only part of the story.
Olympus and Panasonic M43 cameras are very satisfying in the sharpness department. What I used to think was pretty good is now base line and exceptional has become the norm. So what? The elusive smoothness I loved from my Canon cameras and the glassiness of the Fuji's, turned out to be less fine detail resolution with contrast increasing perceived sharpness.
The newer raft of sensors are very clear, but Fuji and Canon in particular are aiming for a more holistic quality. What the Olympus cameras can supply is a choice between ridiculous sharpness for it's own sake or a deliberate scaling back, allowing other factors to come forward. The choice is great, it just took me a while to get over the "sharpness at all costs" thinking and use it.
An example of lush smoothness, brilliance and biting sharpness from with an EM5 and 75mm. Yes, you can have it all if you are careful.
M43 users have a tendency to be a bit defensive of their little sensor, revelling in the many comparisons done, often between full frame and M43 cameras. They usually show little practical difference between the formats in real terms, but we need to stop obsessing and concentrate on the whole process more. It is true, that the gap has been bridged, to the point that M43 users need to look over their shoulder at the surprising 1" sensor pack nipping at their heels, so move on and create with what ever camera is your chosen poison, it does not matter.
What does it mean to me?
When image quality was a far less controllable thing, image makers used all of the short comings of the medium to their advantage. A little movement blur, well crafted grain patterns, controlled colour failure, deliberate cross processing or processing errors and many other stretched technical realities were embraced (I remember first looking at Robert Capa's processing ruined images of the D-Day landings, thinking how he really got emotion and chaos across. I wonder if more perfect images would have had the same impact?).
EPM2 45mm f1.8 1/20th f8
It is a shame that I feel I am running the risk of losing my feel for image imperfectness. The image above was a mistake (a street grab using city landscape settings). It is also my favourite of hundreds taken on my last trip to Melbourne. Ironically, if I had managed to capture it as I intended, it most likely would have become another of my many so-so street images.
On one level, I am grateful for the wonderful quality available now to all of us for relatively little cost, but I am personally at a cross roads. Do I want to loose what is left of my experimental, curious photographer brain, concentration on perfection always or do I need to stop upgrading, settling for my almost perfect cameras, but allowing some small hint of the wonder of miss-takes in. Maybe a compact pushed too hard, a film camera (oh the expense!), use the old manual focus Pen 25mm lens moreor even use a down graded model sometimes, just to keep up the hunger and awareness of how lucky we are.
I just sold my EPM2 and kit lens. They are excess to needs and produce much the same images as my other gear. They were basically too good and did not add anything different to my kit. Even cheaper alternatives can lack a point of difference. Maybe I should ?
Sharp images are a thing no doubt.
Sharp images are not the only thing.