In this section we’ll deal with two commonly cited aspects of image quality for artistic bird photos: sharpness, and bokeh.  The former, sharpness, should be a fairly intuitive concept, though there are some aspects of lens design (and lens operation) that can potentially lead to variations in sharpness, and we’ll discuss these first.  The issue of bokeh—the quality of having a smooth, pleasing, and non-distracting background—will be addressed later in this section.


Sharpness

Assessing the relative sharpness of two images, when they’re shown side-by-side, is usually pretty easy as long as they’re images of the same subject or scene.  Assessing whether a lens is “sharp”, or “sharp enough”, based on images seen in isolation can be much more difficult.  One problem is that what really needs to be measured is not the sharpness of the lens, but the sharpenability of images produced through the lens under ideal conditions.  In other words, if we were to take some bird photos using a given lens, would we be able to sharpen the photos enough in Photoshop (or similar software) so as to produce usable images?  Of course, the answer depends on the intended use of the image: whether for a web page, or for an 8"×10" print, or for a 40-foot billboard.  The larger renderings (i.e., the billboard) will obviously require more fine details in the image file, in order to still appear sharp after enlarging.
    To illustrate the potential for differences in sharpenability between lenses, consider the two hawk photos taken below.  The left image was taken with a 1200mm lens (actually, a 600mm lens with a 2× teleconverter attached—but that’s not really relevant at the moment), while the right image was taken with a 600mm lens with no teleconverters attached.  Both images are shown at 100% crop, which means that you’re seeing the native resolution of the camera (in other words, each pixel on your screen represents exactly one pixel in the image file; only a tiny portion of each image is shown, however, as the word “crop” should suggest). Because the left image was taken using a 2× TC, it should be much less sharp than the image on the right—meaning that fewer fine details of the bird should be visible, and any that are visible should be less distinct and more “fuzzy”.   In this case, since we haven’t applied any sharpening yet, the difference in sharpness between these two images isn’t terribly striking, though the left image may appear a little “fuzzier” than the image on the right.


    Which crop level is best for comparing two lenses depends on how often you’re likely to use each crop level when “publishing” your images (i.e., when making prints, or when posting photos on a web page).  If you typically photograph large birds close-up and are using a large-focal-length lens, then you probably won’t be cropping at all, whereas if your passion is tiny birds like warblers and hummingbirds, and you’ll be viewing them at a distance or through a moderate focal-length lens, then you may indeed find yourself aggressively cropping your images in order to make the bird appear large in frame.  In the latter case, assessing the sharpness of the lens via “pixel peeping” makes more sense.
    In terms of assessing image sharpness digitally—i.e., on the computer screen—there are a number of important issues to consider.  First, most cameras employ an antialiasing filter which sits over the imaging sensor, and which will to some degree decrease apparent image sharpness.  For this reason, some sharpening in post-process is always recommended for cameras having such a filter, to counteract the slight blurring effect that the antialiasing filter imposes on the RAW images.  That’s one reason why we introduced the notion of sharpenability above.  Second, if you’re not shooting in RAW (and you probably should be), then it’s important to realize that your camera may be applying some sharpening behind the scenes to the JPEG file that it’s exporting.  This is more of an issue when comparing the resolution of different cameras, but it can become relevant for lens comparisons if the amount of in-camera sharpening is such that it ends of masking slight differences in lens sharpness.
    Another very important issue is that apparent sharpness tends to increase with smaller apertures (i.e., larger f-numbers)—at least, up to a point.  As we mentioned in section 3.1, the depth of field (DOF) increases with decreasing aperture.  Thus, for smaller apertures (like f/11), a wider margin in front of the bird and behind the bird will appear to be in focus than at smaller apertures (like f/4).  To the extent that autofocus doesn’t always perfectly set the focal point exactly on the bird, a larger depth of field can help to mask these slight focusing errors, since a slightly mis-focused bird that is wholly contained within a large DOF will still look mostly in focus.  This is illustrated in the figure below.




In the top part of the figure (above), the lens’ iris is opened to its widest setting (f/1.4 in this case), corresponding to the shallowest depth of field.  In this example, the lens is slightly front-focused—notice that the black dot (what we’ll informally call the “focus point”) to which the red arrows converge is slightly in front of the bird.  The orange gradient surrounding this black dot represents the depth of field; the fact that the orange fades out gradually represents the fact that the “in-focusness” of the points around the focus point also gradually fades out at increasing distances from the focus point. 
    In the bottom part of the figure, we’ve stopped down the lens (meaning that we’ve reduced the aperture) by closing the iris a bit (to f/2.8), resulting in a wider depth of field.  Now the orange gradient (representing points that appear more-or-less in focus) extends further to either side (forward and backward) of the focus point.  Since the bird is now more completely contained within the darkest part of the orange gradient, it should appear much more in-focus than in the top part of the figure.  Thus, by decreasing the aperture, we’ve reduced the effect of the focusing error (i.e., fact that the black dot wasn’t positioned exactly on the bird).
    The reason depth-of-field increases with smaller apertures (and also with distance to subject) is explained in Box 3.7.1, below.  Feel free to skip the box if you’re not interested in understanding why DOF works the way it does.  Just note one additional thing: at some point, reducing the aperture can actually decrease sharpness rather than increase it, via diffraction.   Diffraction limits typically don’t become extreme until the aperture drops below f/16 or f/22 on most lens/camera combinations, so it’s usually not an issue for practical bird photography.



    There’s another good reason to “stop down” a lens to a smaller aperture.  It turns out that most lenses are not their sharpest when shooting wide open (i.e., at maximum aperture), and this isn’t simply an artifact of the depth-of-field effect discussed above.  In the figure below are three images of a Brown-headed Cowbird (Molothrus ater), the two leftmost images taken at f/5.6 (the maximum aperture of the lens that was used), and the rightmost image taken at f/8.  Since the focus point was directly on the bird’s cheek or neck area in all three photos, we can largely ignore issues of DOF.  As you can see, fine feather details appear more distinct in the cheek/neck area of the image taken at f/8.  For this particular lens (the Sigma 800mm f/5.6 prime lens), I’ve found that maximum sharpness occurs at about f/11, which is two full stops down from maximum aperture (f/5.6). 




    The conventional wisdom is that while most lenses need to be stopped down a bit to achieve maximal sharpness, not all lenses need to be stopped down by the same amount.  Thus, a higher-quality lens might require less stopping-down than a lower-quality lens, to achieve maximal sharpness, and that would allow you to get the sharpest images without sacrificing as much light by stopping down.  Keep this in mind as you read lens reviews.  Good reviewers will typically test a lens at various apertures.  As a general rule, cheaper lenses are cheaper for a reason, and a lack of sharpness—particularly sharpness at maximum aperture—is often one of those reasons.  Whereas my Sigma 800mm f/5.6 lens typically needs to be stopped down to f/11 (2 full stops) for maximal sharpness, my Canon 600 + 1.4× TC (effective: 840mm, f/5.6) combination is generally at its best at f/7.1—only 2/3 of a stop down from wide open.  In this case I’d chalk that up to the R&D advantages of the larger company (Canon).


Bokeh

A much more subtle differentiator between birding lenses is what is known as the bokeh.  Bokeh is a Japanese term used to describe the background of an image.  Today, many people prefer images in which the main subject stands out well from the background.  Photographers thus try to get photos in which the background is smooth and indistinct, with few details to distract the eye from the main subject.  An example of poor bokeh is given by the Belted Kingfisher (Megaceryle alcyon) photo shown below.




The background in the image above contains a lot of distracting elements.  In this case, the bird was perched in front of a bush with many bare twigs showing.  Because the lens was a fixed-aperture lens, I had no choice but to shoot at f/12, which resulted in a large depth-of-field that wasn’t shallow enough to render the bush totally out-of-focus.  Another problem evident in this photo is the lens’ tendency to make doughnut-shaped patterns in the background, which some people find distracting.  In this case, both problems (the large depth-of-field and the doughnut patterns) were caused by my use of a mirror lens—actually, a Maksutov-style astronomical telescope.  Mirror lenses are often criticized for the poor bokeh they impart to their images.
    Even expensive lenses can produce distracting background artifacts, however.  In the Louisiana Waterthrush (Seiurus motacilla) image below, you can see that the background also contains many round or polygonal shapes which can be somewhat distracting.  This image was chosen for inclusion in a nature calendar, so apparently the bokeh was good enough for the calendar’s editor.  But the fact that the lens, an $8000 (US) prime lens made by Canon, produced background artifacts shows that no lens is perfect in this regard.