Effects of focal length, F-number, and sensor size on background blur and facial features in portraiture
The first thing I want to say is that I'm writing this from a Micro Four-Thirds-centric view. It can be applied to any format, but the examples I use are MFT. The lens information is accurate as of 2013. The lens layout will obviously change in the future.
I'll start by saying that DoF (depth of field) and background blur aren't the same thing. It's entirely possible for two shots to have the same subject framing, same DoF, and one shot to have much more background blur than the other. Here is a site that illustrates this effect:
I'm going to post pics from various sites here, so that if the links go down for some reason, I still have the examples. If the owner of an image wants it removed, I have no problem with that.
|f/5.6 set 1|
|f/2.8 set 1|
|f/5.6 set 2|
|f/2.8 set 2|
You'll notice that she took a series of portraits at different focal lengths with the subject maintaining the same distance from the background in each set of shots. When two shots are framed identically with the lens at the same F-number on both shots, then the calculated DoF for both shots will be very close (though not identical), even when the focal lengths used for the shots are different. For all intents and purposes, we can consider all of her shots at F/5.6 to have the same calculated DoF and all of her shots at F/2.8 to have the same calculated DoF, with the F/5.6 shots having roughly double the DoF of the F/2.8 shots.
Note how much more background blur there is in the shots with the long focal lengths and how little difference there is between the f/2.8 sets and f/5.6 sets. While it would appear that focal length plays a much bigger role than F-number, the reality is that their roles are roughly equal (more on this later). It's just that there's a much bigger difference between focal lengths in her shots than there is between F-numbers, which is something that often happens in real life, especially with zoom lenses. In other words, there's not as much room to maneuver the F-number as there is to maneuver the focal length, which often makes improving background blur through adding focal length easier than through the F-number.
Something that may be of interest is that the difference in DoF between the f/2.8 shots and f/5.6 shots is the exact same difference you'd see between a FF and MFT sensor (when both have the same F-number). That is to say that the difference in DoF (as well as background blur) is there but not that great. Specifically, even though FF has 4x the sensor area and 4x the light gathering capability (assuming equal F-numbers), the DoF (and background blur) difference is only proportional to the sensor diagonal, which is to say a 2x difference. Not that much when you consider the size, weight, and cost differential.
So, what precisely defines the background blur if it has nothing to do with DoF? It's actually the entrance pupil diameter of the lens that determines which lens will provide greater background blur (provided you frame the subject identically with both lenses). The entrance pupil diameter is defined as the ratio of the focal length to the F-number, and can be roughly gauged (but not exactly measured) by the diameter of the front element glass.
|Oly 75mm front lens element|
Note that if we want to be 100% technically accurate, if the subject is very close to the background, then the shorter lens (with the same entrance pupil as the longer lens) will provide slightly better background blur. In other words, when the subject is very close to the background, a short, fast lens will be slightly better than a slow, long lens, even if both have the same entrance pupil. But for most portrait work, you'd want the background to be relatively far from the subject for better subject isolation, and there would be no difference between the long and short lens.
This brings me to perspective distortion with regard to portraiture work and the influence of focal length/distance. Obviously, the further away from the subject you are, the flatter the subject's facial features will appear. However, this isn't usually considered a bad thing, as it often reduces problem areas and offers a more flattering/flattening view. Here are a few more examples of the effect of focal length/distance on facial features.
You'll note that at the lower focal lengths, small changes in focal length/distance produce obvious changes in perspective. Once you reach a certain point (roughly 85 mm), when you increase the focal length further, the differences are greatly diminished for relatively large changes in focal length. In fact, I believe most people (myself included) would have a difficult time telling the difference between 100 mm and 300 mm (35mm eq) with regard to perspective distortion. This is a good thing, as it means there's great leeway with the range of focal lengths that can be used while providing flattering effects.
I've heard some people say that shorter focal lengths are probably better for people with small noses and other features, while longer focal lengths are better for people with more prominent features. In practice, I haven't really noticed much of an issue with using long focal lengths for anyone, as it's almost always a flattering look. This is one of the reasons why magazine photographers and other pros often use 300mm lenses to shoot head and shoulder portraits.
Another thing that's obvious by looking at the portraits shot at different focal lengths is that the closer shots have more obvious blemishes and skin texture, while the longer focal lengths have much less. The reason for this is also perspective distortion. When you're further away from the subject, everything appears more flattened, including the texture of the skin. You'll note that portrait shooters often do PP just to reduce blemishes and skin imperfections, when using a longer focal length will often achieve the same effect, without affecting the sharpness of the image.
With regard to sharpness, using a longer focal length to frame a shot identically to a shorter focal length will often reduce some of the fine 3D textures on clothing, etc. It doesn't reduce any of the 2D detail, such as on a painting. This fine 3D texture often adds to the perception of sharpness of an image, even when no real difference in sharpness exists. So, under some circumstances, using a long lens to achieve background blur can reduce some of this apparent detail.
So, which Micro Four Thirds lenses can be used effectively for portrait work? IMO, a surprising number of them and many that most people would never think of.
For instance, the 40-150, 45-150, 45-175, 45-200, 75-300, and 100-300 may not at first seem like portrait lenses, but they all have relatively large entrance pupils at max zoom and are fairly sharp at max zoom as well. They only require you to step a bit back from the subject to get that OOF background, but they'll all provide a flattering perspective due to the distance you have to get from the subject. The 35-100 is more obviously a capable portrait lens, due to its constant f/2.8.
Of course, there's the 42.5 f/0.95, which has the largest entrance pupil of any native, prime MFT lens (until the Panasonic 150 f/2.8 comes along in 2014). Then there's the 75mm f/1.8, which is currently second for entrance pupil among primes (and will be third in 2014). When the Panasonic 42.5 f/1.2 comes out, it will be just behind the Oly 75mm (in fourth place) for entrance pupil among primes. Behind the 42.5mm f/1.2 will be the Oly 45mm f/1.8. And so on.
Note that different people will have different preferences for focal length and some people may be constrained by studio environments. But if you're outdoors with plenty of space and light, any of these lenses are capable of producing great portraits, even the ones that you may not consider at first. You generally want to be at 85mm equivalent focal length or above, with the largest entrance pupil you can get your hands on, if background blur is a priority.
The views and opinions expressed in this article are those of the author and do not necessarily reflect the views and opinions held by dpreview.com or any affiliated companies.