Teleconverters (Multipliers) - A technical overview
Multipliers (also called 'TCs", "teleconverters" or just "converters") seem like a great way to get a long telephoto lens without spending a lot of money. You can take an inexpensive 70-210 zoom, add a 2x multiplier costing maybe $50 and instantly have a 140-400 zoom. So what's the catch? Well, in short you start out with a relatively slow but fairly sharp 70-210/3.5-4.5 and convert it into a much slower and significantly less sharp 140-400/7.0-9.0. While it's easy to say how much slower the 140-400 zoom will be (2 stops slower with a 2x multiplier), it's a lot harder to predict just how much sharpness you'll lose because so much depends on the specific lens and multiplier in question. Using a high quality 1.4x multiplier on a 300/2.8 APO lens is a very different situation from using a cheap 2x multiplier on a 28-200/5.6 zoom.
People often wonder if multipliers always reduce image quality. The short answer is "yes", but just how much quality will be lost is a lot harder to say. In some cases it may be small enough to be insignificant, while in other cases it may large enough to make the use of the multiplier a waste of time and film. Whether the image is acceptable or not depends on the quality of the lens used with the multiplier, as well as the quality of the multiplier itself, plus, of course, what you regard as acceptable. If the lens gives superbly sharp images on its own, then a small amount of image degradation may still leave you with excellent image quality. If the lens on its own gives an image which is just "OK" then adding a multiplier may result in unacceptable image quality.
You might wonder if the loss in sharpness is simply due to imperfections (aberrations) in the lens and the multiplier optics. If we could find a "perfect" lens and a "perfect" multiplier, i.e. diffraction limited optics with all aberrations as fully corrected as theoretically possible, would we still end up with a reduction in image quality when the two are used together? Well, the answer is still "yes". Before explaining why this is so, let's digress for a moment and discuss something about image sharpness, which is one of the most important factors in image quality.
(Technical Digression)Now back to multipliers. What a multiplier does is to effectively increase the focal length of the lens it is used with. Thus if we put a 2x multiplier on a 300mm lens we end up with a combination which has an effective focal length of 600mm. The definition of the f-stop of a lens is the ratio of the effective focal length to the physical size of the aperture (for the maximum speed of a lens let's call that the diameter of the front element for the sake of simplicity). Thus a 300mm f4 lens has a front element which is about 75mm in diameter (300/75 = 4). If we add a 2x multiplier to the lens, we increase the effective focal length to 600mm, but we do not change the size of the front element. thus now the lens has an f-stop of 600/75 or f8. We now see the first effect of adding a multiplier, that it decreases the speed, or f-stop, of the lens it is used with. A 2x multiplier always results in a 2 stop decrease in speed, e.g. a 300/4 lens becomes a 600/8. A 1.4x multiplier always results in a 1 stop decrease in speed, e.g. a 300/4 lens becomes a 420/5.6. All apertures are affected in the same way, not just maximum aperture, so if you add a 2x multiplier to a 300/4 lens used at f5.6, you end up with a 600mm lens operating at f11.
Image sharpness is often defined in terms of resolution, i.e. how much fine detail can be seen in an image. Though "sharpness" , i.e. the perceived clarity of an image, is related to resolution, it is also related to other factors such as contrast. When discussing "perfect" lenses, sharpness, contrast and resolution go hand in hand, i.e. when one is high, so are all the others (this is not quite true for real lenses, but let's not worry about that for now). Resolution is often measured in terms of "lines per millimeter" or l/mm. An image that can show 100 l/mm obviously can show more detail, and thus appears sharper, than one that only shows 10 l/mm. At this point we also need to define image. A lens produces what is called an "aerial image" in the film plane. This is a real image formed in space. We could examine it with a magnifier - in fact if we did, we would have a telescope! We can also see this image by recording it on film - in fact this is what we do every time we take a picture. Due to the limitations of film (grain size and light scattering), it can only record a limited number of "lines per millimeter". No matter how high the resolution is in the aerial image, the best slow films can resolve around 100-125 l/mm and the sensors in DSLRs currently have a resolution limit of between 70 and 80 lp/mm. Because of the way the resolution in the aerial image and the resolution of the film interact higher resolution in the aerial image always results in higher resolution in the final image on film. This can be mathematically approximated (at least for film) as follows:
1/(image l/mm)2 = 1/(lens l/mm)2 + 1/(film l/mm)2
Thus if the highest resolution visible in the aerial image was 50 l/mm and this was recorded on a film with a maximum resolution capability of 125 l/mm, the highest resolution visible in the recorded image would be 46 l/mm image. Similarly an aerial image capable of 125 l/mm would record no detail finer than 88 l/mm in the image on film and an aerial image capable of 500 l/mm resolution would record no more than 121 l/mm in the film image. The numbers would be slightly different for digital sensors, but would follow a similar trend. Finally, one last bit of mathematics. The maximum possible resolution in the aerial image formed by a perfect lens can be approximated by (1500/f-stop) l/mm. Thus a perfect f4 lens under optimum conditions could resolve about 375 l/mm in its aerial image, and a perfect f8 lens could resolve about 187 l/mm. We will come back to this later as it bears directly on how the use of multipliers affects image quality.
Bearing in mind what we said earlier about the resolution of a perfect lens, we can also see why the resolution decreases when a multiplier is used. If the best a perfect lens can do is (1500/f-stop) l/mm and adding a 2x multiplier doubles the f-stop, it is evident that it also reduces the maximum resolution by a factor of 2. There's another way to look at this if you are not mathematically inclined (or don't believe my numbers!). Take any lens, it doesn't have to be perfect, and look at the resolution in the aerial image. Say it's found to be a maximum of 100 l/mm. What the 2x multiplier really does is to take the center part of this image and make it 2x bigger before it reaches the film. It "stretches" the image by a factor of two. The 100 lines that could be counted in 1mm of the old image are now stretched out and cover 2mm, so the resolution in the final aerial image is 100 lines in 2mm, or 50 l/mm. Note that if film had infinite resolution and no grain, or if digital sensors had an infinite number of pixels, we could just as easily enlarge the image after it had been recorded on film instead of before! However, since film has finite resolution and grain and digital sensors have a limited number of pixels, enlarging the image after it has been recorded (by making a larger print and cropping) normally results in more image degradation than enlarging the image before it is recorded by using a multiplier in front of the film or digital sensor.
As we saw earlier, aerial image resolution is modified by the effect of the film, so that the resolution as recorded on the film or digital sensor doesn't drop by a factor of two with a 2x multiplier. It drops by a factor somewhat less than two, depending on just how good it was in the first place. The better the starting aerial image, the less drop you will see on the image produced with the multiplier. This is obvious if you think about it. Let's say your sensor resolves 80 lp/mm and your lens resolves 400 lp/mm. If you add a multiplier and drop your lens resolution to 200 lp/mm, you are still well "outresolving" the sensor and the change in observed resolution will be small. However if you start out at 100lp/mm and drop that to 50 lp/mm, your 80 lp/mm will clearly show such a resolution reduction.
For the very best lenses used with the very best 1.4x multipliers, the drop in image quality (measured by image resolution) may only be as small as around 10%, probably resulting a a barely noticable change in image quality. This might go up to 20% for the case of 2x multipliers combinations and that's usually noticable. Of course cheap, poor quality multipliers add their own aberrations to those of the lens resulting in a much larger loss of image quality and in that case you can sometimes get better results by simply enlarging the original image than by adding a cheap multiplier.
Conclusions
When choosing a multiplier, or deciding whether or not to use one at all, there are a number of points to think about:
(1) From the point of view of image quality, it's always better to get physically closer to the subject than use a multiplier from a greater distance. Sometimes, however, there is no choice - that's why we have multipliers.
(2) Even the very best multipliers used on the very best lenses will lower image quality, though that "lowered" quality may still be very high. With lower quality multipliers and/or lower quality lenses, significant image degradation may be noticed. There will be a loss of contrast and resolution, particularly at the edges of the image, and even at the center of the image there may be an unacceptable loss of quality. Some multipliers used with some lenses may also result in vignetting, i.e. slightly dark image corners.
(3) The lower the multiplier power, the better the image quality will be. All else being equal, a 1.4x multiplier will give sharper images than a 2x. Don't be tempted to buy an "off brand" 3x multiplier for any kind of serious work.
(4) If you can afford it, and if they are compatible with the lens you want to use them with, use the camera makers multipliers with the camera makers lenses. They will generally give better results, in part at least because they were specifically designed to work with a known set of lenses. For example the Canon EF multipliers are designed to work with (and in fact will physically fit on) only the Canon EF "L" series prime telephoto lenses and a few "L" series telephoto zooms.. Some companies, like Nikon, make several different multipliers, each one designed for use with a specific set of their lenses (e.g. lenses shorter than 300mm and lenses longer than 300mm). If the camera makers multipliers are too expensive, stick with ones made by the well known 3rd party lens makers (Tamron, Tokina, Sigma and Kenko).
(5) Multipliers work best with high quality fixed focal length lenses, especially telephoto lenses. As a general rule they don't work well with inexpensive zoom lenses, mirror lenses or short focal length (wideangle) lenses.
(6) In the case of auto focus systems, for reliable AF operation (sometimes for any AF operation at all), you have to ensure that the maximum aperture of the final lens+multiplier combination is f5.6 or faster except for a few cameras (suvh as the EOS-3 and EOS-1 series which will Af at f8 with the center AF point). For a 1.4x multiplier this means staring with a lens which is f4 or faster. For a 2x you need to start with a lens which is f2.8 or faster. Some multipliers on some cameras with an f5.6 cutoff will give AF operation with a combination having a maximum aperture of f8, but usually only in good light with high contrast targets and AF speed and accuracy may still be compromised. Don't depend on it.
Finally, a couple of additional technical points about multipliers.
(1) They work no "magic" with depth of field. The depth of field of a 300/4 lens plus a 2x multiplier (= 600/8) is exactly the same as the depth of field of any other 600/8 lens. For a lens system of any given focal length and aperture, the depth of field is essentially the same, whether or not a multiplier is used.
(2) To a close approximation, the close focus distance of the original lens is retained. If you have a 300/4 lens capable of focusing down to 6ft, when you add a 2x multiplier you end up with a 600/8 lens which is also capable of focusing down to 6ft.
Multipliers are valuable tools and they are used with great success by many amateur and professional wildlife photographers. However, they can't work miracles and the wise photographer knows the limits of his or her equipment! Don't depend on the use of a multiplier for your photo safari to Africa unless you have thoroughly tested it on all the lenses you intend to use it on and you are happy with the quality of the results. The time to judge image quality is before, not after, you have taken those "once in a lifetime" shots!
What do I use?
I have three multipliers, a Canon 1.4x, a Canon 2x and a Tamron 1.4x. I normally use the Canon 1.4x with a Canon EF 300/4L lens, making a very sharp 420mm f5.6 lens with excellent AF. This is a very good combination. I will occasionally use the 2x with the 300/4L and results can be quite good, especially with the lens stopped down by one stop. However that gives me a 600mm f11 manual focus lens, so the light needs to be quite good before I use the 2x. It's the smallest, lightest, way to get a decent (if not great) 600mm lens when weight is an issue.
I use the Tamron-F 1.4x on several lenses. One is the EF70-300/4-5.6IS USM. This isn't a lens that the Canon TC will fit on due to a protruding front element. The Tamron will mount on any EF series lens and actually works OK with the 70-300IS (it wasn't so hot with the older 75-300IS). AF attempts focus but hunts and eventually fails, so I use this combination in manual focus mode - however the green dot focus indicator still works, so there is some electronic confirmation of focus. I've also used this TC with the EF50/1.8 to get a fairly fast 70mm f2.5 lens for portrait work.
I also use the Tamron-F 1.4x with an EF500/4.5L because in this case it gives me good AF performance. With the Canon 1.4x, AF is disabled by the camera since the resulting aperture (f6.3) is 1/3 stop slower than the f5.6 minimum. Image quality and AF performance are excellent with an APS-C sensor EOS 20D, but I suspect that the image quality at the edges and corners of the frame would be worse on a full frame camera using the Tamron 1.4x than the Canon 1.4x. The Tamron would also likely produce slight vignetting at the corners of a full frame camera (I know it did with a 300/2.8L and 600/4L when I tested it with them).
When I need maximum reach I use the Canon 2x on the Canon EF 500/4.5L, giving me a 1000mm f9 lens (which on a 20D with the 1.6x multiplier gives me the reach of a 1600mm f9 lens on a full frame SLR). Image quality can be OK (if not spectacular), but focus is manual and that can slow things down a bit.
Here's an example with the 2x on the Canon EF500/4.5L. This was shot in Acadia National Park last year and shows two fledgling Peregrin Falcons way off on a distant cliff face. First here is the full frame shot:
And here is the center section as a 100% crop after a little digital adjustment.
I have stacked The Canon 2x and Tamron 1.4x converters together to give me 2.8x, but usually the results aren't any better than shooting with the 2x and enlarging the image a little more.
Further Reading