On real image quality needs

A lot is written about camera gear and the ever increasing reality envelope they offer, but have we lost sight of what we really need and that we probably already have it.

Recently my business (a book shop) hosted a camera release night for the OMD EM1 mk 2. As you may already know, I have a clutch of trusty EM5 mk1 cameras and a new Pen F, so it was with a fair amount of forced denial that I hosted, but did not get too caught up in the event. The reality is though, new cameraitis set in a bit. I know the new camera will not take a noticeably better image than the ones I use and that is because I also know (not too deep down either) that what I need is easily fulfilled, but it is undeniably a slicker machine.

Lets help me and possibly you overcome this most recent GAS attack (Gear acquisition syndrome) and continue on happily with our own, owned gear.

First up, a little story that has been shared previously, but I will go into more detail.

While testing a new lens (23 f1.4 Fuji) on an old camera (XE 1) and comparing it to the 17mm on my OMD, I stumbled upon an old friend, the image quality reality check, purely by mistake.

The OMD was set up as usual for RAW, but the XE 1, a less used camera, had been set to low res JPEG for some posting images and I had not noticed. The whole morning I shot side by side comparison shots with both cameras, came home and loaded them (went off for a coffee while they loaded and did not notice that the Fuji images fairly zipped into upload). Then I started to process them side by side. The simple test was to determine two things. The first was to see if the 23mm fit my needs (I did not buy it in the end as it was too big and expensive, no sharper than the excellent 27mm and indeed I got out of Fuji not long after). The second was to see how close the simpler, more literal images of the Olympus could get to the more delicate Fuji jpegs in colour and "pop" (pretty close and as I learn more, even closer).

While processing I immediately noticed that the 100% views of the Fuji images were only jumping up about half again from the screen image, roughly enough to fill the 29" screen on my Mac, but they were delicate, super sharp and full of quality. On checking, I discovered the size issue and would have been mightily pissed is it were not for the quality I was seeing. How quickly we forget! These images were 2400x1600 pixels, about 4 mp and had more than enough for a glorious 11x14 inch test print, which I did for many.

There was something humble and pleasant about an image that was perfect to the eye, to a reasonable size, but denies an unnecessary close an examination. When looking at the Olympus images at 100%, they became far too technical and intrusive an examination. The lesson here is don't look at your images at 100% as it has little relevance to the end product.

A small jpeg from my XE1 with the 23mm f1.4. This printed perfectly to 11x14"

A small jpeg from my XE1 with the 23mm f1.4. This printed perfectly to 11x14"

Ok, so what do we really need;

Pixels (resolution).

For galley quality images up to 12x18 prints, carefully used 6mp's will do. More can give excellent colour depth (more a camera generational thing) and render some more detail at 100% on a screen, but no one will complain that your images are sub par due to quality issues, they will not be able to tell (plenty of blind tests have been tried including huge differences in print resolution etc, so check the interwebs if you want). Landscape shooters will have issue with this no doubt, but these same image makers used lower res cameras for many years successfully. If you look at this the other way, doubling your pixel count or even tripling it will not change the way your prints are perceived unless maximum size at closest viewing distance is your only criteria. Why do we look happily at an image on a screen bigger than the print size we intend, like what we see, then look at 100% and get depressed by the slight fall away in quality when we cannot even print it out? Do we look at our food under a microscope when preparing it and if we did would change what we tasted?


If sufficient in quantity for the viewing size, pixels have nothing much to do with sharpness. Image sharpness is determined by many other factors, pixels only add resolution, not sharpness. Like contrast, they are different animals. Sharpness comes from good technique, good sharp lenses at usable apertures, correct focus and depth of field and file quality enough to hide any unnecessary, introduced issues. Over sharpening and over saturation are the single biggest culprits for modern image quality short comings, not a lack of pixels or unsharp lenses (we have plenty of these). Mike Johnson on the Online photographer blog shows us a great examples in his posts "The Color Disease" and "Are you Real".


Like pixels, this one is in the top few "must fix" categories for the tech heads and has become apparently unacceptable on any level recently, but like pixels, it is not that big of an issue. Like a perfectly made blanket, we are now picking on every very minor imperfection or flaw in the fabric and risk missing the cosy comfort it offers. As an example, OMD cameras are known to have some noise at base ISO, but if you print TO SEE IT, it can be hard to find. Much like film grain, tight, clean noise can become a part of the texture and character of an image, with as much right to be there as colour or tone. Indeed black and white film shooters used grain as an element in their imagery deliberately, which is why grain engines are deliberate ADD IN options in digital film simulations. I do not mean to mushy, smudgy poorly exposed/processed looking faux grain common in some film simulations, but fine, textural grain that enhances the perception of edge acutance and tonal transition.

If we think like a printer (i.e a realistic large image maker), the camera we have will usually exceed our needs as indeed cameras have always somehow provided for, but if we spend our days splitting hairs over the slight differences between camera "X" or camera "Y", we will never be satisfied. This continual quest for better is a cycle that will never stop and hides the fact that we have usually past sufficiency for most needs. 

Just for fun, see how good an image you can produce with your worst camera-file size/lens combo, but best technique. The though crossed my mind that I could shoot 4mp jpegs, use 2Gb cards and a 100GB hard drive for all of my images.