A favourite topic of mine is looking at the realistic end point of our work, the tools needed to get there and their application. This was an obsession for a long time, then slowly shifted to a more intellectual exercise as my own processes and gear settled, allowing me to stop chasing rainbows and take a hard look at technical realities.
The photo and video industries are very quick to tell is that more is better. More pixels, higher resolution, faster operation, more options, more endurance, greater potential, but the reality is, we maybe do not need as much as we are led to believe.
This particular re-post was inspired by a visit to a friend about to go on a trip with a decent compact camera and a good eye at a time when some more Japan files have been stumbled acros. I keep seeing good images coming from gear considered so very sub-par by the current thinking.
Photographers and videographers are the worst judges of this.
We are generally pixel count and resolution obsessed, something the industry as a whole is pushing to stay relevant and alive, so always chasing better at the cost of more than good enough. This has always been the case, but like most things at the moment, it seems to have been turbo charged (although slowing thanks in part to COVID, which has allowed us to take stock).
In the film era, much desired improvements happened once a decade at most, bigger formats ruled, enlargement and viewing tools were limited, so the technology itself limited our choices and in turn these limitations forced better technique.
We as a group still turned out good work for over 100 years.
Firstly lets address the viewer.
Most of your viewers should not be practitioners. If they are, you need to get out more. The average viewer will be at first attracted to your subject matter, then their emotional contention to same. This connection may be driven by relevance or simply content, but without it, you have nothing. The average viewer will only be aware of the technical stuff when they are aware they fall short of the norm, showing faults of some type. With video, this is ironically most often related to poor sound.
Beauty is in the eye of the beholder and that beholder is rarely concerned or even aware of the finer technical details.
So, what is the norm?
The current viewing base-line is 1080 to 4k screen viewing, often on small screens at that. The occasional need to “explore” details rarely happens on any but a purely technical level and must be balanced with applying the rule of the proper viewing distance.
Large fine art prints, billboards and huge, high resolution screens are usually viewed at the correct viewing distance and they all fall apart on closest inspection (a 4k screen is only 8mp resolution and printing comes eventually down to ink dots and paper texture).
Contrasting subjects, strong light, deep depth of field and clean lines all contributed to this images perceived “quality”. It could however have been taken by any decent compact camera.
Often the weakest link in the quality chain is outside of your control or irrelevant in real terms.
Ming Thein on his blog experimented with ultra high resolution printing and although successful, found that the ability to mimic on paper what a high resolution (36mp) file on a screen could potentially show (loupe-level detail of the moon on a landscape print), was possible, but ultimately irrelevant and cost prohibitive. Basically he came to the ultimate question of “yes, but who cares?”.
Ctein, a master printer, offered for collectors a cheap (printing cost and freight) 13”x19” print from an older 12mp EPM-1 Olympus M43 sensor image, a sensor known to be softer than later models, just to show that with good technique, pixels and sensor size are ultimately quite low on the quality control check list. The image of a large portion of a bridge showed individual rivets clearly. All his work has a thread of the practical application of sensible gear. His take is good technique beats just tech alone, or alternatively, no amount of tech will save bad technique.
So, what is important?
Separation.
Not always the primary quality concern of an image maker, separation adds to the perception of sharpness or lack of it.
Separation (sharpness) comes from many sources, edge contrast being the main element. Contrast is a combination of light, subject matter (texture and tone), control of noise or grain, depth of field, colour, clarity, lens sharpness and image “stillness”, all contributing to the visual elements of clear separation.
The two images below are identical except the left hand one has added contrast. No sharpness, just contrast in a variety of forms (gentle exposure+vignetting, clarity, dehaze, contrast). The right hand image is sharp, but lacks obvious tells for the viewer.
You can lose sharpness easily enough by compromising any one of these elements, but you do not add any more visual sharpness by throwing more pixels at the problem. This only matters when enlargement size or micro viewing become the main consideration and only after the rest have been addressed. Pixels are only the quantity, not the quality of the image.
The two images below were taken identically. The one on the left looks less sharp, because the right hand image has more contrast and cleaner elements. They are identical down to “chicken wire” level.
More pixels can actually detract from sharpness, if their higher needs are not met. A high resolution full frame sensor needs some pretty tight technique and premium glass to feed it. No amount of post processing will fix resolution lost to blurr, lens aberrations or soft corners, all more likely the bigger the sesnsor and higher the pixel count.
This image is cropped slightly to square from an older generation 16mp, M43 camera file, hand held with a medium grade 45 f1.8 prime lens. It has separation, contrast, shallow depth, colour and clarity. It would not be visually sharper with more pixels, but if enlarged enough, it will fall apart when the maximum enlargement size of the relatively small pixel count is surpassed. If image sharpness is retained with more pixels added then you can print it bigger, but all the other rules are at play here first.
As much as we all get a little thrill when we can see minute details like the small hairs on the leaf edge, who apart from a field scientist, obsessed photog or the occassional overly curious observer, looks this closely at a wider image. Go macro if this really interests you or accept it for what it is.
Pushed in as close as the 16mp sensor will happily go, there is still detail separation to be found, probably lost in a print or a screen. Notice also how the visible sharpness plane has shifted. The rear leaf looks sharp above, but has become a blurred background element at this magnification. Printing, although it can lose fine detail, can also organically smooth out digital artefacts, a useful trade off.
Sometimes, even with the best gear and technique an image by its very nature resists impressing us with its detail retention. A lack of separation at one level can however change context at another. In the image below, a lack of colour contrast and the finest detail being too small to see makes the wider image look high quality and textural but not “snappy”.
The crop reveals micro detail and adds contrast with simpler and cleaner elements.
Quality is dependant on so many things, pixels are just one and often not the biggie.
Bokeh is sometimes maligned by photographers who have not been exposed to its true meaning. Bokeh is not how much blurring a lens can render, which is mostly down to depth of field, but the quality or character of that blurring. This can effect perceived subject separation.
An example of smooth Bokeh and shallow depth of field can play for clean separation. One of the reasons photographers are often drawn to smoothly rendered Bokeh is it adds to the impression of separation by cleanly contrasting the sharp with the unsharp. This was one of the “Leica magic” characteristics so loved in the film era.
Sometimes, even when your sharpness ducks are not in a neat row, you can still get that “snappy” rendering. The Olympus 17mm, unique to me in that it renders background detail very coherently even at wide apertures, but still manages to separate middle distance subjects out.
Cleanliness.
I am not talking about the photographers hygiene or even their subjects, but image cleanliness.
This has layers so bare with.
A lack of actual image cleanliness, such as noise (grain), chromatic aberration or any other obvious technical flaw and poor techniques like missed focus or poorly arresting subject motion will reduce the objective quality of an image. The common offender is often flat, poor light mixed with high ISO Grain and mild subject movement, making poorly lit indoor sport the king of pain.
Contrary to this, sometimes grain or texture can actually add to separation and define edges as it has enough contrast and is clean in its own right. I had a friend, sadly passed away over a decade ago, who was very successful on the international fine art circuit with his ultra grainy black and white images. His methodology was to use super sharply defined film (later digital) grain on an otherwise clean base, rather than try to avoid it.
Sometimes when most of the elements of clean separation are missing, you have to take what you can. The soldier in the background is an ideal example of lost separation due to low sharpness, blended colours and other aberrations, but these in turn add to the feeling of clean separation of the bugler.
When things go your way, it can seem too easy.
Brilliant sun adding contrast, shallow depth of field and no noise or overly busy Bokeh all add to the quality of this image.
This may not be what you want. Blur, movement, texture, grain, other colour weirdness may be the desired elements of an image, in which case a designed-to-be-flawed Holga, relic SLR or home made pin-hole film camera will likely be the tools of choice/ If not, cleanliness will help raise your perceived image quality to the next level.
Ironically, the video industry is mostly trying to avoid overly sharp “digital” looking footage, often employing softening filters to “take the edge off” or in other words, “dirty up” the image. Separation in movies is often desirable, but the hyper-realistic hard-sharp look is not.
*
The other form of image cleanliness is compositional clarity.
Have a clear message and communcate it cleanly. This comes with experience and a decent dose of inherent talent (I rely on experience mostly).
This has two benefits. On the one hand it makes comprehending the image easier for the viewer so they look at it instinctively, not critically. On the other, it reduces visual distractions and increases the feeling of subject separation.
A worker at Hangar 17 distillery. The elements are kept simple, the only “clutter” is relevant and additive.
OK, so from the above, where am I going (again)?
Take good, compelling, relevant photos or nothing else matters.
Use good technique and learn your gears strengths and limits to unlock its potential (which few ever do fully).
Do not get distracted by arguments about sensor size, pixel count of resolution until you undeniably come across a situation you cannot solve any other way.
Learn to appreciate your own work on its visual strength, not its technical short comings.
Look at the best of what others do, but avoid the obsessive need to find out what gear they used, because ultimately, it does not matter. Looking at photos is good, obsessing over gear is pointless. I did for years and the take away is, it is wasted time.
*
Pixels/Resolution is not all important, only technically relevant at the extremes of reproduction. End viewing needs rarely require more than the current base line (20-26mp). Software empowerment aside, only large printing fine artists or scientists genuinely need more than this. Sharpness and clarity on the other hand, is important.
For video, 1080p is what most clients actually need, even if they think they need 4k or more. Good 1080p trumps poorly realised 4k every time. Often 4k is reduced in size for common use, with no inherent advantage. Again, quality trumps quantity.
Sensor size has some relevance for image cleanliness, especially in low light, especially for video, but often good technique can fix any issues. As a smaller M43 sensor stills user, I use fast glass (also with a depth of field advantage), good processing, exposure tricks and lighting to solve all the issues I come across and if these do not work, only a small handful of bigger sensor cameras would offer any real benefit before they also fall over (we are talking about shooting in near complete darkness with very demanding expectations). Use the kit that best suits your needs, the math evens out.
Embrace your camera, but if the images it makes fall short, look to technique before buying a new one. Maybe a new lens if you only have the kit one, something fast for shallow depth, maybe a prime to help train your eye and guarantee top quality. A (good) tripod can help, then migrate away from jpegs to RAW or do a workshop in a field that interests you, but try to avoid more pixels just for their own sake. It is the path of least reward.
*Studies have shown that the quality difference of video resolutions over 4k at this point in time, even if viewed at an uncomfortably close distance, is effectively invisible to the eye. The same goes for printing. A billboard at the correct distance can look razor sharp, even though at closer distances it is made up of golf ball sized blotches. Quality at the correct viewing distance is paramount, not resolution.
Another study at a major technical university (RMIT I think?) showed two prints the same size side by side in a thoroughfare with passers by asked if they noticed anything. One was printed at 72 dpi, the other at 300 dpi or maybe higher. Very few viewers noticed any difference, even though the faculty were experts in the field.
A further example is a pair of older articles on The Luminous Landscape website. One is the story of a 50mp medium format landscape shoooter who used a 12mp G9 Canon compact camera for record keeping and found he could print to the same standard size as the MF camera before even his experienced friends could not see a difference.
The other showed prints made from 8, 24 and 50mp cameras of the same street scene. At very high enlargements there was more detail in very small image elements like the fine print on a parking sign, but even the higher MP cameras could not resolve details well enough to be seen on a print. At normal sizes and viewing distances the 8mp image was effectively enough.