I jokingly say that all lenses are bad. They are either unsharp, with poor aperture and flawed in all sorts of ways, or they are heavy and expensive.

To that’s the compromise, in a nutshell. You can design a lens that has great aperture and optical performance, but it’s going to be expensive and heavy. Alternatively, you can make it cheap and light, but that’s going to adversely influence performance.

not necessarily so; shot with “cheap and light” Olympus Pen with the collapsible 14-42mm kit lens

So, why is that? Well, first of all, you need to know how lenses are designed. Basically, they bend light in a certain way, and that introduces flaws. Then additional elements are introduced to correct for those flaws. Every additional element means additional point of light transition between air and glass, which means antireflective coating is super important to maintain contrast and reduce flaring. Also, more glass equals more weight. You can reduce lens size if you reduce the circle it’s drawing, but then you get vignetting and unsharp corners. You can remove those problems by enlarging the circle, but then you increase weight. Also, you can reduce the aperture and thus keep the lens smaller and less expensive, while retaining sharpness. So, you can have it light, sharp, fast and cheap, just not all at the same time.

Sharp and fast, while heavy and expensive

The zoom lenses increase all the problems, and at the same time remove lots of the benefits. For instance, a f/2.8 zoom is considered fast (fast not in a sense that it does anything quickly, but in a sense that it lets through lots of light, thus allowing for faster shutter speeds). For a prime lens, f/2.8 is considered slow, and “fast” really starts at f/1.8. Also, with zooms, the more range you have the worse the lens typically behaves, and adding a fast aperture to that makes things either much worse optically, or expensive and heavy. This is why I think trying for a “Swiss army knife” lens that would do it all is an exceptionally bad idea. Yes, the idea is to replace a bag of lenses with one, but the problem is, that one will often weigh and cost about as much as the aforementioned bag of lenses. You think I’m joking? Try looking at the prices of 24-70mm f/2.8 and 70-200mm f/2.8 lenses in top quality, and you’ll see what I mean. Then again, try shopping for f/4 zooms in high quality, and fast f/1.8 and f/1.4 primes, and you’ll see what I mean. This is the reason why I usually stick with f/4 zooms and use them where versatility of focal length ranges is paramount, and aperture is secondary or irrelevant – such as landscape photography from a tripod, and I go for primes where focal range is irrelevant, but speed and sharpness wide open are paramount, such as low-light photography handheld, or portraiture with shallow depth of field. You don’t need to have it all in one lens, which is why people invented system cameras with interchangeable lenses, so that you could avoid resorting to Swiss army knife solutions. Sure, I understand that there are conditions where you want to avoid changing lenses a lot, because you could either miss shots, or drop a lens, or allow dust/rain inside the camera. However, in most cases this is not actually a problem. Carrying around a huge and expensive lens on your camera, however, is very rarely welcome.

The tragedy of zoom lenses, and especially the “standard” zoom lenses, in 24-105 or 24-70mm range, is that people think they are versatile and they can get by with only one lens for most things. Sure, they are versatile in focal length range, but they are usually very limiting in aperture, and this is what you are more likely to need; in essence, a 50mm f/1.8 prime is more versatile than a 24-105mm f/4 zoom, and I’m more likely to take it with me as a walkaround lens. It is also likely to be very light, inexpensive and very sharp. The ability to open the aperture all the way to f/1.8 means you can shoot in the kind of dark where you see the stars in the night sky; you can blur the background for portraits and get close enough for almost-macro details, and once you can do that, the limitation of a single focal length is not that much of a big deal for most things you might wish to do. And yes, you can do a great deal of landscape shots with a 50mm, and if 50mm is too tight or too wide for you, then you will know what to get next.

50mm f/1.4 prime, wide open at closest focusing distance

A beginner will feel the need to cover all potential uses of a camera before they figure out what they actually want to shoot, and that’s usually both expensive and frustrating, because you might actually get limiting equipment thinking you’re giving yourself options. For instance, one of the most practical and versatile lenses is a 35mm f/1.8 with good closeup capabilities, for instance the RF 35mm f/1.8 macro for Canon, or FE 35mm f/1.8 for Sony. They are sharp, fast, light and not very expensive, especially considering what you’re getting. Get that as your “every day” lens, and then observe what you wish you had – a specialised lens for birds, or an ultrawide, or a longer macro, or a midrange zoom for landscapes from a tripod, and buy lenses that are specialised for those tasks, instead of trying to buy the most expensive and heaviest Swiss army knife.

35mm for nature/landscapes

35mm in town, late night

35mm for details

Also, if you want a portrait lens, someone might suggest a 70-200mm f/2.8, which is extremely popular, but unless you need the focal range for indoor sports and weddings, I would recommend a 85mm f/1.8, or a 135mm if you want things tighter. You can save tons of money, and get results that might actually look better.

A portrait with an 85mm f/1.8 lens at f/2.8

85mm f/1.8 for details

Essentially, what I would recommend to a beginner who doesn’t want to overspend, and yet wants to be able to make photos of exceptional quality, would be to get a cheap standard kit lens that would be used stopped down to f/8-f/16 from a tripod, and a 50mm f/1.8 prime. Then, depending on the needs, get either a 16-35mm f/4 wide angle, or a 85mm f/1.8, or a 100mm f/2.8 macro, or 70-300mm f/4.5-5.6, or whatever you actually found out that you would need, based on actual experience and not guesswork. However, be advised that everything is a compromise – especially the “no compromise” optical solutions. So, the advice is to get in as cheaply as you can, possibly with used gear from a “dead” system that is heavily discounted, learn the ropes, figure out what you actually want, and then buy specific lenses for specific needs, avoiding the do-it-all solutions, because they are very often so impractical that they end up being not particularly good for anything.

I’ve been thinking how to explain photographic optics to beginners, because the explanation is usually something like this: the bigger the sensor, the lesser the depth of field, the closer the point of focus the lesser the depth of field, the wider the aperture the lesser the depth of field, the bigger the focal length the lesser the depth of field. So, depth of field is inversely proportional to focal length, sensor size, and aperture, and directly proportional to focal distance.

But what exactly is the depth of field?

To understand depth of field, you need to understand that only a single point in the image, the point of actual focus, is perfectly sharp. Everywhere else, the lens diffuses a ray of light into a circle, called the “circle of confusion”. If the circle of confusion is small enough to fit into a single pixel on the sensor, basically if its diameter is less than the pixel pitch of the sensor, we perceive it as “sharp”. However, what it actually looks like can be mathematically represented as a parabolic function:

Let’s say that the x axis represents the distance from the focal plane of the sensor. 0 is the point in focus. The y axis represents the diameter of the circle of confusion, and 0 means that it is in perfect focus, and y is a mathematical point. If it’s below 1, it means it is perceived as sharp because the diameter of the circle fits within one pixel. So, everything between 0 and 1 is perceived as sharp, and as the radius grows, the picture starts getting blurry. Let’s illustrate this photographically:

As you can see on the asphalt, there is a tiny strip of sharpness, and everything before and after is blurry, progressively with the distance from the point in focus. At the most distant parts you can clearly see the circles formed as the rays of light are defocused. On the strip in focus, it appears that more than just a single point is in focus, because for some time the defocused circles are small enough to fit within a pixel on this level of magnification. This strip that is perceived as sharp is called the depth of field.

The shape of the circles of confusion beyond the depth of field is called “bokeh”. The angle of the parabolic function defines the amount of bokeh. If the parabolic function is steep, it means that only the things very close to the point of focus are sharp, as everywhere else the circle quickly grows into huge “bokeh balls”:

If the function is “shallow”, it means that it is possible to create an illusion that the entire image is sharp, from the closest visible point all the way to the horizon, because the diameter of the circle of confusion at the desired magnification of the output is still less than 1, or close enough not to matter:

So, in effect, the perceived sharpness beyond the point of actual focus is an artifact of magnification and photographic trickery. If you magnify enough, you’ll see that almost nothing is ever in perfect focus.

Now we get to the useful parts. If you shorten the focal length, you reduce the angle of the parabolic function, making it almost parallel to the x-axis. The same happens when you reduce the sensor size, but the reason is trivial – you need lesser actual focal length to produce the same field of view on a smaller sensor, so it’s still merely a function of the focal length. When you stop down the aperture, you force light to go through a smaller circle, reducing the cone of light that produces a mathematical point when in focus, or disperses more strongly when not. In theory, only a picture formed by perfectly collimated light source, meaning parallel rays of light, would be in perfect focus, meaning no light cones, but laser rays; also, this means that a pinhole camera, or a camera with a lens stopped down so extremely that it becomes a pinhole, would form a perfectly sharp image with an infinite depth of field. However, as the aperture becomes smaller, the rays of light interfere with its edges, which blurs the image, which means that such a perfectly sharp image would also become perfectly blurry due to diffraction. So, since, the f-stop is a function of the focal length, the absolute aperture is merely a function of the focal length as well, which means that we managed to reduce three parameters to one – sensor size, aperture and focal length are all functions of the absolute focal length, which leaves distance from the focal plane. However, if you imagine the light cone formed by the object in focus, it is obvious that the angle of this cone becomes larger as the object is closer, and smaller as it is farther. This means that the rays of light become closer to parallel as we focus closer to infinity, which produces an effect similar to stopping down the aperture. This also means that we managed to reduce all our parameters to the angle of the light cone, and diffraction as the second limiting factor that literally interferes with sharpness.

Sure, there are other optical effects that play a role – for instance, the difference in sharpness away from the centre of the circle that the lens draws, and within which the sensor is inscribed, the difference in refraction of various wavelengths of light, which creates chromatic aberration, the difference in the amount of light that falls on the edge of the circle relative to the centre, which defines vignetting, and so on. There are many optical defects that detract from the ideal appearance of the image, and although sometimes those optical defects can look interesting or charming, I personally subscribe to the opinion that less is more when it comes to defects. The only optical defect that I actually intentionally introduce to images at times is vignetting, because I think it focuses attention to the centre of the image, which is usually useful.

In any case, the difference between optical defects and optical laws is that the optical laws are something that will determine what happens with every single lens of a certain aperture and focal length, when projecting a circle of a certain size upon the focal plane. Optical defects, however, are what differentiates lenses of superior and inferior optical design and quality. This means that all telephoto lenses with wide aperture will create very shallow depth of field, but the good lenses will be tack sharp in the point of focus and devoid of significant defects, while bad lenses will be lacking in sharpness everywhere, and will introduce all kinds of optical defects, such as field curvature, low resolution, chromatic aberrations, colour cast and so on. Complexity of design and manufacturing tolerances that make a difference between bad, good and great lenses can be really extreme, which is reflected in the price; your tolerance for optical defects might vary. Unlike defects, the optical properties that are defined by focal length, circle size and aperture are universal, and don’t vary with lens design and price range.

I recently talked a lot about photographic equipment, and my annoyance with all kinds of nonsense I’ve recently seen on the Internet about how Leica or Fuji is “special” because it has “that something” or encourages you to take pictures with how cool it is and how you form a synergic relationship with it, unlike Sony or Canon which is “just a tool”. I get it, it’s a real phenomenon and I’ve personally experienced it, and what’s interesting is that it literally never happened with the kind of equipment you would expect. It didn’t happen with Leica, which is the first camera I shot with, or Minolta, which is probably the only camera brand I’m sentimental about because it was the one I used for, well, decades. It wasn’t Olympus E-1 which was justifiably praised as ergonomically great. It wasn’t the Canon 5d.

Fuji Astia 100F, EOS 3, EF 85mm f/1.8

It was Canon EOS 3, the camera that was widely criticised for being “too electronic”, with lights and whistles and what not. That was the first camera with which I experienced camera completely vanishing from the process of taking pictures. I didn’t have to thumb-wind film like I did with Minolta X-300, which made me remove my eye from the viewfinder, since I use the left eye to focus (I’m highly ambidextrous, but I use left parts exclusively for certain things, and right parts for others, for instance I use mouse with my right hand and touchpad with my left, and so on), I could do absolutely everything on the camera without breaking eye contact with the subject through the camera, and the autofocus was scanning my eye with a laser to decide what I’m looking at and choose the appropriate AF point. Also, everything was extremely fast to the point of being instantaneous and seamless – AF, film transport and so on. Unfortunately, I got it about the time when film was solidly on its way to history, in 2006, but it was absolutely the best camera I ever used at that point, film or digital. Returning to the 5d felt like going down at least one quality class, but it was digital and the pictures it produced were amazing, so I didn’t complain.

The worst experience I had with cameras recently, on the other hand, was with the Mamiya 645. I shot a few rolls of film with it, decided it’s heavy, clunky, required lots of adjustments on my part and acquiring muscle memory in order to be proficient, and it was a poor fit for my style. I returned to the Olympus E-PL1, which is also an ergonomically awful camera because it doesn’t have a viewfinder and autofocus is horrible, with a feeling of relief because it was so much better than the Mamiya, and the image quality of the Olympus was actually as good with my Epson 4990 flatbed scanner. The problem with those bad cameras is that they make the experience all about themselves. It’s no longer about you, what you see or feel, but about what the camera can do, what it’s good at, how heavy it is, how much of a nuisance it is, and I think it’s exactly that which eventually forms that Stockholm syndrome relationship with equipment, which is hard to use, makes you work hard to adjust to its quirks and nonsense, and is also expensive, so when you actually manage to jump through all the hoops, you get an endorphin rush as if you achieved something.

I recently took a picture with the Sony A7RV and the FE 35mm f/1.4 GM in sunset, and it’s one of my recent favourites because of how well it captures the feeling of clarity and peace, of a day coming to its end:

Sony A7RV, EF 35mm f/1.4 GM

I saw the light, the colours and the motive, framed the composition, pointed the AF to the plane with the branches and some leaves and made it somewhat parallel to the sensor in order to get the dept of field at f/1.4, and squeezed off the shutter. The camera was out of the way and just did its thing to a degree I haven’t experienced since EOS 3, although A7II was also similar, barring autofocus which is a nuisance. I think EOS 3 and A7RV actually do the things people ascribe to Leica and Fuji – they get out of the way and become an extension of your will and intent to the point that you lose awareness that you’re using a camera at all. It’s just that they don’t put you through boot camp and try to break you in in order for you to be allowed to just take the pictures you want to take without putting the damn camera on a shrine because you dedicated half your life to learning how to work it. I think that’s why the hipsters hate certain cameras and brands and say they are “just tools”. It’s because those cameras don’t lock them up in a basement and beat them until they call them “daddy”. They just do what you ask them to, and if the results aren’t good, you don’t get to say “ah, but you don’t know how many hoops I had to jump through in order to get that”. 🙂

Every single photo I took with EOS 3 was incredibly easy to take, and regardless of the fact that I got to shoot only a few rolls of slide through it, on those rolls was an inordinate percentage of my all-time favourites. Sony seems to be very much like that. It just gets out of the way and does its thing, and this allows me to get into my thing. And, believe it or not, “my thing” isn’t dealing with camera’s bullshit. It’s capturing the moment and the feeling.