Looking back at old files, some taken with cameras as many as four generations old (normal generations not Sony generations), basic lenses, a format some are suspicious of and sometimes “loose” technique to say the least, I am drawn once again to the question, “what is it we really need, and how do we get there?”
The first characteristsics we are often presented with when looking at cameras are pixel count and sensor size.
What does pixel count effect and how and where does sensor size come in?
Quantity of information.
The number of pixels mathematically defines the maximum resolution a sensor can produce which effects the maximum theoretical enlargement size (depending on viewing distance, image contrast, reproductuon limits etc) or the maximum realistic cropping of a file (or usually a combination of both) before pixels become visible.
I used “mathematical” and “theoretical” because to put it simply, this is only one form of measuring quality which can be affected by so many other factors. Printing has the ability to hide some degradation and has its own limits as do screens, ideal viewing distance and firmware/software is the future reality, so pixels are effectively de-throwned as the critical element. The big one is, very few viewers look as closely or a critically as photographers do.
The sensor size then determines how large each pixel is in relation to the sensor real estate. Larger sensors make pixels relatively bigger so allow the greater pixel density to be less destructive. The pixels on a 12mp full frame sensor (which some still consider to be ideal) are many times larger than those on a 20mp compact camera sized sensor, thus gathering more and cleaner light and better base image quality.
Ironically, the bigger the sensor and the greater the pixel count, the higher the bar for lenses, meaning that unless you get the best available (big, expensive, rare), you may loose some of what you gain.
The images below were taken with an older 16mp M43 sensor (EM10 mk2). They are plenty big enough to make decent sized gallery prints.
Noise.
The other feature (or benefit?) of pixel size is control of digital noise production. Larger pixels naturally create less noise as the light gatherings ability of the individual pixel is increased. Noise (grain) is created by light starved pixels failing to render information then bleeding into their neighbours, creating a “blotch” of colour or black pixels instead of harmonious colours and tones.
Think of each pixel like a little bucket. Big pixels are more hole than edge and the bottom of the bucket (the sensor) is less angled, letting in more information, smaller pixels are realtively more edge than hole (but same height) and tiny pixels are like a straw that only gathers light when the sun is directly over head. This is where the sensor size really pays in.
Again there are other factors like sensor design, processing electronics and the actual light available, but at the end of the day, all things being equal, bigger pixels gather more light.
There is a catch though. More pixels may indeed produce more digital “noise” reducing image quality, but they also produce smaller noise. This can be beneficial in certain circumstances, but is generally considerred to be less ideal than no noise at all.
What pixels do not effect are;
A file taken with a 12 year old, 16mp M43 sensor, hand held with a basic prime lens.
Visual Sharpness.
Sharpness is a product of contrast in colour, texture or light, the subject matter, focus, depth of field, clarity of glass/air/sensor, sound technique, viewing distance/light/expectations and more. Resolution, often caught up in that pool is actually not the same thing. Pixels are how much information is gathered, but not necessarily how well it is translated.
This has always been the case. With film, bigger negatives produced clean, grainless enlargements even at higher ISO ratings, but on close study, the lenses used were often less sharp inch for inch than smaller SLR lenses. What we were seeing was clarity, often shallower depth of field and the benefits of a relatively smaller enlargement size, not extreme sharpness, but it did the same thing.
Quantity can matter, but quality does also. From memory the Hassleblad 80mm for example was measured at around 65-70 line pairs per millimeter resolution (the measure we used before pixels), the Nikkor 55mm Micro managed 80-90.
Image power.
Perfect technique is irrelevant without image impact. Great photographic images have been created for well over a century. Some, especially early ones had obvious technical short comings, but regardless had the ability to hold our attention, to tell their story. Better technique was always the ideal, but a powerful image in its time was still iconic. Ironically, older, less technically perfect processes have one huge advantage, increased longevity.
The digital era heralded a period of technical improvement (much needed early on), that became infectious, addictive even. It became the habit to compare “X” to “Y”, empowering many web sites and a sub-hobby in its own right. That time has passed folks. All cameras are good enough, too good even, lenses are getting to point where their perfection is robbing them of character and the reality of end use seems to be forgotten.
Realistic needs.
Ok, lets be honest. Who among us ever prints fine art prints bigger than 16”x20” and if we do, do we leave our technical prowess at the door and throw ourselves to the mercy of just pixel counts and auto correct? Independent tests have proven that even this size cannot reproduce the full information of a 36mp sensor without unrealistic efforts being taken and if they do, we are down to viewing with a loupe. Who does that?
Maybe a book or two is your ambition, a decent web site, maybe displaying on a decent sized screen? The good news is, these all have their limits and in most cases, you have surpassed them.
Future Proofing?
I can guarantee you, your images will have to stand on their own two feet artistically or for their relevance before anyone cares a hoot about the technical issues. Issues likely only you see anyway. Things age and change, something you cannot stop. More future-proofing quality is a very small part of that picture.
I love the work of the early colourist photographers like Haas, Herzog or Leiter. I do not like the more modern takes some publishers have forced on their work, much preferring the originals for their gentle authenticity.
The truth is, the future will bring ever more powerful software to increase quality short falls, but you may actually choose to avoid that. Let them be what they are, representatives of their time and place.
This shot actually reminds me of colour images taken in the 1980’s except that the quality is too good! To capture that feel, something I like as it stirs memories of early discovery in a favourite time in my life, I would have to degrade the colour, the sharpness and brilliance of the image to increase its ‘70’s character.
We all do look too close at our images, which is our right and our bane, but we alone do that. Clients, friends, admirers look at our work on its aesthetic merits, not its technical ones. Take a bad photo perfectly and see how far you get.
There is a younger generation who seem to be split between the two opposed camps of more is better and those chasing that elusive something extra, the something that even those who lived through the many past transitions have forgotten.
So, what do we need?
Speed, accuracy, reliability, enough of everything else, which is less than the maximum possible and ourselves. Nothing more.