At the 400mm level, the third-party lenses are clearly much cheaper than the Canon/Nikon lenses, both for prime and for zoom lenses.  In the field, I rarely ever see third-party 400mm zooms in use for birding, though; in the vast majority of cases, birders using a 400mm zoom were using a Canon or (more rarely) a Nikon.  The Sigma 500mm zooms are quite popular, however, since they’re very cheap compared to a 500mm prime, and are lighter and more compact, making them hand-holdable (unlike the Canon/Nikon 500mm lenses).  Above 400mm, the third-party primes become much more comparable to the name brands in price (rendering them less attractive), and above 500mm even the zooms become outrageously expensive (e.g., the Sigma 300-800mm zoom).  Note that the table above doesn’t include the ultra-cheap 500mm mirror lenses (which run about $150) nor the small-aperture (e.g., f/16), manual-focus, super-telephotos (e.g., 1000mm, give or take) that can be had for a few hundred dollars and are reputed to be optically horrendous.
    To address the question of relative optical quality between the top-tier and third-party lenses, we’ll consider a particular pair of highly popular birding lenses in great detail.  From this very specific example we’ll then try to (cautiously) draw some very general conclusions about what may be expected of third-party “glass”.  When considering a particular lens, it’s advisable to try to find a similar comparison to the one below, for the particular lens under consideration and a comparable name-brand lens, to see what you might be giving up by going with the third-party lens.
    The two lenses we’ll consider in detail here are the Canon EF 600mm f/4 L IS USM, and the Sigma 800mm f/5.6 EX DG APO HSM.  (We’ll dissect those lengthy names in just a moment).  Since these two lenses have different focal lengths and different apertures, we’ll make it a fairer fight by putting a 1.4× teleconverter (the Canon EF 1.4× Extender II) on the Canon, giving it an effective focal length of 840mm and an effective aperture of f/5.6.  Now we’re effectively comparing two lenses of the same aperture (f/5.6) and nearly the same focal length (840mm versus 800mm).  Since the Canon has a slight advantage in focal length (by 40mm), we’ll move the Canon lens back a few feet from the subject during imaging tests, so that both lenses will produce approximately the same magnification.  Note that the Canon is actually somewhat handicapped by this comparison, since the use of a teleconverter can be expected to reduce image quality.  Keep that in mind when we compare the resulting images: if the Canon, with TC, turns out to be as good as the Sigma (with no TC), then that would suggest that the underlying optics are better in the Canon, since with no TC the images should be even sharper still.
    In terms of the detailed specs for these lenses, the EF in the Canon name just refers to the lens mount, which is the same for the Sigma lens (this Sigma lens is available in Canon, Nikon, Sigma, and other mounts).  The L is just Canon’s designation for its most elite lens line (it apparently stands for “Luxury”).  The IS is of course Image Stabilization, which we covered in the previous section, and the USM stands for Ultrasonic Motor, which means that the focusing element is moved via subsonic vibrations rather than mechanical gears, making it both fast and quiet. 
    The Sigma’s acronyms are as follows.  The EX is a product-line designation, much like Canon’s L, and probably stands for Excellence or some similar platitude.  The DG means it was redesigned for use with digital cameras (Sigma’s product literature is a bit vague on this point, but refers to additional lens coatings that are somehow more suitable for use with digital sensors).  The APO of course stands for Apochromatic, as we discussed in the previous section, and the HSM is Sigma’s version of Canon’s USM.
    For the first test, we’ll compare the two lenses by photographing a US $20 bill affixed to a door at the end of a hallway.  For this test the lenses were mounted on the same camera (a Canon EOS 1D Mark III), on the same tripod, with all the same camera settings.  All images were shot at f/8, without mirror lock-up (MLU—see section 6.11), and with image stabilization on the Canon lens turned off.  The result is shown below, with the Canon image shown on the left and the Sigma image on the right.




There are several things that can be concluded from the figure above.  First, the image on the right appears darker, or more saturated, than the image on the left.  As further comparisons could show, the Sigma lens in fact has a bit of a yellowish cast that it imparts to all photos taken through it, possibly due to one of the antireflective coatings applied to the lens by Sigma.  I consider this to be a very, very minor issue, since it’s often not noticeable (except in direct comparisons like this one), and can be corrected easily in post-processing.
    Second, the Canon image has more constrast than the Sigma image.  Notice how the blacks in the Canon image look black, while the same regions in the Sigma image look brown instead.  This is particularly apparent in Andrew Jackson’s coat, and in the word “SERIES” on the lower right corner of the image (this becomes clearer in the next figure, below).  As a result, fine details such as those in the former president’s hair or facial lines, or in the eagle emblem near the lower right corner, seem sharper at this crop level in the Canon image than in the Sigma.  Look also at the letters T and U along the rightmost edge of the two images.  They stand out much more noticeably in the Canon image.  Looking up then at the U and the “RESE” above that, the Canon image seems to have greater definition, due to the increased contrast.
    In order to compare true sharpness, we need to zoom in to the pixel level, as is done below.  The Canon image is again on the left, while the Sigma image is on the right.  (You can again see the deeper yellow saturation in the Sigma image).




    Note first that the black again seems blacker in the Canon image than in the Sigma, while the light colors are lighter, indicating greater contrast.  Comparing the “SERIES 2006” lettering and the eagle emblem, you can see that both lenses do appear to have been focused properly for these images.  But in the Canon image you can see that the paper has a fine hexagonal pattern embedded in it.  (Look just above and slightly left of the eagle emblem).  Seeing this hexagonal pattern in the Sigma image requires more effort, and perhaps even some imagination.  To be completely fair, this level of detail might, conceivably, be brought out in the Sigma image by first correcting the color and contrast issues.
    What’s most impressive, however, is the fact that the Canon setup involves the use of a teleconverter, whereas the Sigma setup for this experiment didn’t.  As discussed in section 3.4, all teleconverters degrade image quality—even $400 ones like the Canon 1.4× Extender II used in this comparison.  The fact that the Canon lens with a teleconverter attached was able to out-resolve the Sigma lens with no teleconverter attached (while, nevertheless, achieving equivalent magnifications) strongly suggests that the native optical quality of the Canon lens is strictly greater than that of the Sigma. 
    But photos of $20 bills aren’t photos of birds, and what really matters is how well a lens performs in the field.  In the figure below are shown a pair of female Northern Cardinals (Cardinalis cardinalis).  The left image was again taken with the Canon 600mm f/4 lens + 1.4× TC (effective: 840mm f/5.6), while the right image was taken with the Sigma 800mm f/5.6 prime lens.  (For all of the bird photos in the remainder of this section, the Canon image was taken with a 1D Mark III, 10MP camera, while the Sigma image was taken approximately a year prior with a Canon 30D, 8MP camera).





     To summarize these three, highly ad hoc comparisons, I’d cautiously suggest that the Sigma 800mm lens is providing bird images with roughly the same amount of detail as the Canon 600mm lens + 1.4× TC combination.  That says something good about both lenses: for the Sigma, it says that the 800mm prime is a very good birding lens, while for the Canon it says that the 600mm + 1.4× TC combination is a good birding “lens” and that Canon’s own 800mm prime lens should be even better, since it provides 800mm of focal length without the need for any teleconverter.  Were we to compare equal focal lengths (without the use of teleconverters), I’d predict that the various Canon (or Nikon) lenses would prove optically superior to their Sigma counterparts, but that the Sigma lenses would prove “good enough” for most birders.
     The 800mm f/5.6 is one of Sigma’s most prestigious (and expensive) lenses, so generalizations from the comparison above need to be very cautious.  Certainly, the 800mm f/5.6 Sigma lenses—both the prime and zoom versions—are becoming increasingly popular among birders, as are the two current Sigma 500mm zooms.  It’s likely that many birders would be very satisfied with either of these two Sigma lenses, and as Sigma continues to improve its lens line, Canon and Nikon may at some point start to notice a significant loss of market share in the long telephoto segment.
    In summary, it appears that at least one third-party company—Sigma in particular—has the capability of producing lenses with optical quality rivaling that of Canon and Nikon.  Sigma lenses do seem to have a better reputation than those of the other third-party manufacturers, so if you’re thinking of going third-party, I’d recommend sticking with Sigma if possible.  Be sure to check all the designations of the lens you’re considering: the Sigma lens considered above was an EX DG APO HSM, and was purchased in early 2007.  Recent Sigma models having all of these designations might compare as favorably as the model considered above to the brand-name lenses.  As always, it’s best to do your homework by searching on the internet for actual comparisons of the lens under consideration (or for sample bird photos taken with that lens, at photo-hosting sites such as Flickr), and to buy from a merchant with a long return period (like Amazon, which typically offers 30-days, or Adorama/B&H which typically offer 14 days) and to test the lens thoroughly yourself during the return period.  Keep your receipt, and if possible avoid filling out the warranty card until after the return period has elapsed.