“Soft” and “sharp” photos

Christmas decorations, photo taken with f/1.2, 50mm lens.

As holidays are traditionally time to be lazy and just rest, I have not undertaken any major photography projects either. One thing that I have been wondering though, has been the distinction between “soft” and “sharp” photos. There are actually many things intermingling here. In old times, the lenses I used were not capable of delivering optically sharp images, and due to long exposure times, unsensitive film (later: sensors), the images were also often blurry: I had not got the subject in focus and/or there was blur caused by movement (of target and/or the camera shaking). Sometimes the blurry outcomes were visually or artistically interesting, but this was mostly due to pure luck, rather than any skill and planning.

Later, it became feasible to get images that were technically controlled and good-looking according to the standard measurements of image quality. Particularly the smartphone photos have changed the situation in major ways. It should be noted that the small sensor and small lenses in early mobile phone cameras did not even need to have any sort of focus mechanisms – they were called ‘hyperfocal lenses’, meaning that everything from very close distance to infinity would always be “in focus” (at least theoretically). As long as you’d have enough light and not too much movement in the image, you would get “sharp” photos.

Non-optimal “soft” photo: a mobile phone (iPhone) photo, taken with 10x “digital zoom”, which is actually just a cropped detail from the image optically created in the small sensor.

However, sharpness in this sense is not always what a photographer wants. Yes, you might want to have your main subject to be sharp (have a lot of details, and be in perfect focus), but if everything in the image background shows such detail and focus as well, that might be distracting, and aesthetically displeasing.

Thus, the expensive professional cameras and lenses (full frame bodies, and “fast”, wide-aperture lenses) are actually particularly good in producing “soft” rather than “sharp” images. Or, to put it slightly better, they will provide the photographer larger creative space: those systems can be used to produce both sharp and soft looking effects, and the photographer has better control on where both will appear in the image. The smartphone manufacturers have also added algorithmic techniques that are used to make the uniformly-sharp mobile photos softer, or blurry, in selected areas (typically e.g. in the background areas of portrait photos).

Sharpness in photos is both a question of information, and how it is visually expressed. For example, a camera with very low resolution sensor cannot be used to produce large, sharp images, as there is not enough information to start with. A small-size version of the same photo might look acceptably sharp, though. On the other hand, a camera with massively high-resolution sensor does not automatically procude sharp looking images. There are multiple other factors in play, and the visual acuity and contrast are perhaps the most crucial ones. The ray of light that comes through the lens and falls on the sensor produces what is called a “circle of confusion”, and a single spot of the subject should ideally be focused on so small spot in the sensor that it would look like a nice, sharp spot also in the finished image (note that this is also dependent on the visual acuity, the eyes of the person looking at it – meaning that discussions of “sharpness” are also in certain ways always subjective). Good quality optics have little diffraction effects that would optically produce visual blur to the photo.

Daytime photo of naakka (jackdaw) in winter, taken with a 600mm telephoto lens (SIGMA), f/7.1, exposure 1/400 seconds, ISO value at 6400 – with some EOS 550D body/sensor’s visual noise removed in postproduction at Lightroom. Note how the sharp subject is isolated with the blurry bacground even with the f/7+ aperture value, courtesy of long focal-range optics.

Similarly, the sharp and soft images may be affected by “visual noise”, which generally is created in the image sensor. In film days, the “grain” of photography was due to the actual small grains of the photosensitive particles that were used to capture the light and dark areas in the image. There were “low ISO” (less light-sensitive) film materials that had very fine-grained particles, and “high ISO” (highly light-sensitive) films that had larger and coarser particles. Thus, it was possible to take photos in low-light conditions (or e.g. with fast shutter speeds) with the sensitive film, but the downside was that there was more grain (i.e. less sharp details, and more visual noise) in the final developed and enlarged photographs. The same physical principles apply also today, in the case of photosensitive, semiconductive camera sensors: when the amplification of light signal is boosted, the ISO values go up, faster shots or images in darker conditions can be captured, but there will be more visual noise in the finished photos. Thus, the perfectly sharp, noise-free image cannot always be achieved.

But like many photographers seek for the soft “bokeh” effect into the backgrounds (or foregrounds) of their carefully composed photos, some photographers do not shy away from the grainy effects of visual noise, or high ISO values. Similar to the control of sharpness and softness in focus, the use of grain is also a question of control and planning: if all and everything one can produce has noise and grain, there is no real creative choice. Understanding the limitations of photographic equipment (with a lot of training and experimentation) will eventually allow one to utilize also visual “imperfections” to achieve desired atmospheres and artistic effects.

Chocolates were shot with f/1.4 value (50mm lens) – the ‘dreamy’ look was desired here, but note how even the second piece of chocolate is already blurred, as the “zone of acceptable sharpness” (also known as the “depth of field”) is very narrow here.

Testing Sigma 150-600mm/5.0-6.3 DG OS HSM Contemporary

I have long been thinking about a longer, telephoto range zoom lens, as this is perhaps the main technical bottleneck in my topic selection currently. After finding a nice offer, I made the jump and invested into Sigma 150-600mm/5.0-6.3 DG OS HSM Contemporary lens for Canon. It is not a true “professional” level wildlife lens (those are in 10 000+ euros/dollars price range in this focal length). But his has got some nice reviews on its image quality and portability. Though, by my standards this is a pretty heavy piece of glass (1,930 g).

The 150-600 mm focal range is in itself highly useful, but when you add this into a “crop sensor” body as I do (Canon has 1.6x crop multiplier), the effective focal range becomes 240-960mm, which is even more into the long end of telephoto lenses. The question is, whether there is still enough light left in the cropped setting at the sensor to allow autofocus to work reliably, and to let me shoot with apertures that allow using pretty noise-free ISO sensitivity settings.

I have only made one photo walk with my new setup yet, but my feelings are clearly at the positive side at this point. I could get decent images with my old 550D DSLR body with this lens, even in a dark, cloudy winter’s day. The situation improved yet lightly when I attached the Sigma into a Viltrox EF-M Speed Booster adapter and EOS M50 body. In this setup I lost the crop multiplier (speedboosters effectively operate as inverted teleconverters), but gained 1.4x multiplier in larger aperture. In a dark day more light was more important than getting that extra crop multiplier. There is nevertheless clear vignetting when Sigma 150-600 mm is used with Viltrox speedbooster. As I was typically cropping this kind of telephoto images in Lightroom in any case, that was not an issue for me.

The ergonomics of using the tiny M50 with a heavy lens are not that good, of course, but I am using a lens this heavy with a monopod or tripod (attached into the tripod collar/handle), in any case. The small body can just comfortably “hang about”, while one concentrates on handling the big lens and monopod/tripod.

In daylight, the autofocus operation was good, both with 550D and M50 bodies. Neither is a really solid wildlife camera, though, so the slow speed of setting the scene and focusing on a moving subject is somewhat of a challenge. I probably need to study the camera behaviour and optimal settings still a bit more, and also actually start learning the art of “wildlife photography”, if I intend to use this lens into its full potential.

My SIGMA 150-600 mm / Canon EOS M50 setup.

Lens trumps the camera?

It is sort of interesting to think that maybe cameras have already got “good enough”? By this I mean that the capabilities of the camera body are no longer the real bottleneck in photography. Following the field, it is easy to find anecdotal stories about professional photographers relying on their 10-year-old, even much older equipment, with no need to update or upgrade. And this does not count in the “retro” photographers who for various reasons prefer the film cameras and vintage equipment.

As digital cameras include microprocessors, and the light-sensitive sensors are based on semiconductor technologies, the development of new cameras has gained a lot from the “Moore’s Law”, and quick progress in manufacturing faster and faster silicon chips. It is today particularly in the design and marketing of smartphones where this “speedrun” is obvious, with the next generation following the previous one in every six months or so. But even in smartphones, the sales are slowing down, and one reason appears to be that the existing phones are already – good enough.

The brains of a digital camera are its processor, the system chip. This is where sensor information gets processed, operations such as AF (automatic focus systems) are coming from, and where any in-camera postprocessing of photos takes place. I have been mostly following the evolution of DIGIC series of image processors by Canon, and it is obvious that many genuinely useful features for photographers have come from the new processor generations. In addition to being able to fit in data from lens and light sensors to produce more-or-less optimally exposed photos, the newer generations have e.g. introduced face-detection autofocus, which can automatically find faces in a group photo, and set the depth of field so that all of them are sharp. Mostly the new generation usually just provides incremental improvements in the some fundamental areas such as speed of image processing, noise reduction in low-light conditions, or speed and preciseness of autofocus.

It is nice to have a fast-shooting, fast-focusing camera that does all sorts of intelligent things like scene detection, and is able to apply many settings automatically. On the other hand, much of the art and craft of photography is in learning to think about the key dimensions of photographs, and about developing the ability to make use of technology to produce a certain kind of creation. The “smart” processor might be useful in removing the danger of technically failed shots, but it might also slow down a bit the ability to experiment, and learn from mistakes? I know from my own experience how easy it is just to give the “Program” (the ‘semi-auto’ mode in Canon) the reigns, and then end up living in somewhat smaller creative sandbox, as the result.

Putting over-emphasis on the latest features in cameras has also the danger of missing out other important dimensions of cameras as physical tools. The mechanical construction of a camera, the size and shape of it, how the physical dials and control buttons work – all of this have a very significant effect on the handling and ergonomics that matter a lot while taking photographs. Consider the latest smartphones, for example. In many cases the wide-angle and normal focal length photos can be shot with a smartphone with technically excellent results. However, most professionals still prefer to have a tool that is designed to be a camera also in ergonomic terms, while taking photographs all day long. The slippery smartphone with virtual, on-screen buttons just does not provide same kind of experience and sense of control.

Thus, in many cases one can actually save some money by settling for an older-generation model in the camera body, and investing into lenses instead. This can be a bit tricky, of course, as new camera and lens generations sometimes also come with new lens mounts; the autofocus and metering systems, for example, might rely on new pins for exchanging information between the lens and the body in new ways, or -as in the case of mirrorless cameras – the lenses are redesigned to take advantage from the smaller shape of mirrorless body (that is, moving the lenses physically closer to the image sensor). In many cases, however, the manufacturer standard lens mount still applies, or there is a perfectly working adapter available, to fit new lenses to older generation bodies, or the other way around.

Thus, one way for an enthusiast photographer to move forward in the actual image quality and range of photos one can achieve, is to stick with a bit older camera technology, but put the available savings into updating the lenses. In interchangeable lens cameras there are different basic options for the lens selection, and this relates to the style of photography one is working on. A street photographer, or one that mostly shoots people and events, can do nicely with a “normal” lens – or in portraiture with a short telephoto. In this lens range, the maximum aperture, sharpness and absence of various distortions what one is paying for, in a good quality (or “professional”) lens versions.

I think that I have pretty decent situation in wide angle and normal focal lenght photography at the moment, but there is much to improve in the longer telephoto lenses. Particularly my growing interest in nature photography translates into need for long-range, bit-aperture and sharp lenses. And unfortunately those things do not come cheap. Below are a couple of interesting alternatives for a Canon EF mount – I’d be interested to hear any comments or experiences you might have of these, or other EF mount telephoto lenses!

Canon EF 100-400mm f/4.5-5.6L IS II USM (Photo credit: Canon.)
Sigma 150-600mm F5-6.3 DG OS HSM | S. (Photo credit: Sigma.)