TC or not TC, that is the question...

by Bob Atkins

TCs (teleconverters, multipliers) seem to be a source of some confusion. There are questions about light loss, aperture change, depth of field and close focus distance that seem to be raised over and over again, as well as more subjective questions of how much sharpness is likely to be lost when a TC is used.

Lets deal with the technical stuff first:

1. A 1.4x TC will always cost you one stop of light. There's no way around this, it's basic optics, so an f4 lens becomes an f5.6.
2. A 2x TC will always cost you two stops of light. More basic optics - an f4 lens becomes an f8 lens.
3. A TC does not alter the close focus distance of a lens. If it focuses to 6ft with no TC, it will focus to 6ft with a TC added.
4. The Depth of Field of a lens + TC is exactly the same as the depth of field of an equivalent prime lens at the same aperture. So a 300/2.8 with a 2x TC (600/5.6) will have exactly the same depth of field wide open (f5,6) as a 600/5.6 lens at f5.6 (or a 600/4 shot at f5.6)
5. If you add a 2x TC to a lens and don't change your distance from a subject, your magnification will double and your depth of field will decrease - in fact for most telephoto subjects it will be approximately cut in half (see my DOF calculator)
6. If you stack TCs, you multiply their powers, not add them. So if you stack a 1.4x and a 2x, you get a 2.8x, not a 3.4x TC. You add the aperture losses, which in this case will be 1 + 2 = 3 stops.

Image Quality

The effect of TC use on mage quality is a bit more difficult to define because it depends on a number of factors, such as the exact prime lens being used, the exact TC being used and whether you're measuring using fine grain B&W film (e.g. Technical Pan), regular color film (e.g. Provia 100) or a digital sensor (and if so, which one and what the pixel spacing is). The images shown below were taken using a Canon EF 500/4.5L lens, Canon EF 1.4x and 2x TCs (Mark I TCs) and a Canon EOS 10D body.

It's not difficult to show that even a perfect TC on a perfect lens will decrease image quality. Since the maximum image quality (MTF) of a perfect lens depends on diffraction, which in turn depends on the relative aperture (f-stop), and adding a TC always decreases the maximum aperture, it clearly follows that even a perfect TC on a perfect lens will lower the MTF. For example, at 50m lp/mm the maximum possible MTF of an f2.8  lens is around 0.9. If you add a 2x TC (making it an f5.6 lens), the maximum possible MTF at 50 lp/mm drops to 0.8. If the TC also adds some optical aberrations (which is fairly likely, especially off axis), the MTF will drop even lower.

So here are some examples of the effect of TC use on image quality. The target magnification for each shot was maintained constant by moving to the appropriate distance from the target (i.e. with a 2x it was shot from twice the distance). The lens was mounted on a Bogen 3051 tripod with an Arca-Swiss B1 ballhead. Mirror lock up (MLU) was used.

Above: Full frame shot of test chart used

	This is a 100% crop from the center of the original digital image. 500/4.5L with no TC. 500mm at f4.5 (800mm effective focal length if 1.6x crop factor of 10D is taken into account.
	This is a 100% crop from the center of the original digital image. EF 500/4.5L plus EF 1.4x TC. 700mm at f6.3 (1120mm effective focal length if 1.6x crop factor of 10D is taken into account). Image quality is very close to the prime lens with no TC added. Resolution and contrast have not noticeably dropped.
	This is a 100% crop from the center of the original digital image. EF 500/4.5L plus EF 2x TC. 1000mm at f9.0 (1600mm effective focal length if 1.6x crop factor of 10D is taken into account). Image quality has noticeably dropped. Contrast and resolution have dropped and the image is now slightly soft.
	This is a 100% crop from the center of the original digital image. EF 500/4.5L plus EF 1.4x TC, 12mm extension tube and EF 2x TC. 1400mm at f13.0 (2240mm effective focal length if 1.6x crop factor of 10D is taken into account). Large drop in contrast and resolution. Clearly not a combination you'd want to use on any subject with fine detail.
	This is a 100% crop from the center of the original digital image. EF 500/4.5L plus EF 2x TC, 12mm extension tube and EF 1.4x TC. 1400mm at f13.0 (2240mm effective focal length if 1.6x crop factor of 10D is taken into account). Same as above, but with order of TCs reversed. Seems slightly worse with 2x next to lens.

These results pretty much reflect my experience shooting with both film and digital bodies, and with various "L" series prime lenses including the EF 600/4L, EF 300/2.8L. EF 500/4.5L and EF 300/4L. My experience has been that using a Canon EF 1.4x TC results in excellent image quality. Often it's hard to see a difference between the prime lens and prime lens + 1.4x TC, though in detailed tests using fine grain, high resolution film a small resolution drop can be seen.

However with the 2x I have typically seen the image go slightly soft. This isn't a focus error (I focus bracketed these tests), nor is it camera shake (I used a 12.5lb Bogen 3051 tripod with an Arca Swiss B1 head. plus mirror lockup). Images made with the 2x are usable, in fact I have made some excellent 5x7 prints from images shot with the 500/4.5L + 2x TC and an EOS 10D. Sharpness and contrast can be boosted in an image editing program to partially compensate for the degradation induced by the 2x TC and pretty good image quality is possible with care and a little effort. An example is shown at the end of this article.

Though the Mark I Canon TCs cannot be directly stacked (the Mark II versions can), on many of the long Canon prime lenses, TCs can be stacked with a 12mm extension tube between them and distant focus is still possible (you can't quite get to infinity, but you can focus up to a few hundred yards away). The last two images in the sequence above show the results of stacked TCs on a 500/4.5L. Clearly not very good. Now they might be usable in some cases where fine image detail was not required (maybe shots in fog or silhouettes or sunsets), but if fine detail is important, this isn't the way to go for large images. For small web images, or for shots without fine detail it could be OK as shown below:

Above is a full frame shot with 1.4x and 2x TCs stacked on a 500/4.5L. Doesn't look too bad! On the right is a crop at 50%. Though soft, since there isn't much fine detail the image is acceptable.

Edge Quality

Here are some 100% crops of a test pattern about 80% of the way to the corner of the image.

	500/4/5L at f4.5, no TC
	500/4.5L plus 1.4x TC (f6.3). There's a little more degradation in image quality here at the edge than was seen at the center. Resolution is slightly down and there's a little extra chromatic aberration.
	500/4.5L plus 2x TC (f9). Again image quality takes a bit of a hit with the 2x. Contrast is slightly down, resolution is down and there's more chromatic aberration visible.
	500/4.5L plus 1.4x plus 12mm extension tube plus 2x TC. Not a pretty sight. Resolution and contrast are down. Chromatic aberration is up.
	500/4.5L plus 2x plus 12mm extension tube plus 1.4x TC. As above, resolution and contrast is down, Chromatic aberration is up

Edge quality suffers more from the effects of a TC than center quality. One reason is that chromatic aberration (at least lateral chromatic aberration) doesn't affect the center of the image. Another reason is that aberrations such as astigmatism and coma also don't affect the center of the image, so it's easier to keep center image quality high than edge image quality.

Digital vs. Film

There are two differences here. The first is that film is capable of higher resolution than digital. I say capable because it depends on the film (which needs to be slow speed, high resolution) and how you measure the resolution (from a scan or using a microscope). So on film, under the right conditions,  you may well see more difference between the prime lens and the prime lens + 1.4x TC. In the past I've done film tests which showed maybe a 10% drop in resolution, say from 80 lp/mm to 72 lp/mm. Digital can't resolve more than around 60-65 lp/mm, so this difference will be less obvious on digital than film (though the effect may still be seen since overall MTF drops along with "resolution").

The other difference is that most current digital sensors are smaller than 35mm film, so "edge" quality on digital may well look better than "edge" quality on film, since on film the "edge" is further from the optical axis and so off-axis aberrations will be stronger.

TCs or more Enlargement?

Since results using a 2x TC are always inferior to those from the prime lens with no TC added, the question is sometimes raised as to whether it's better to use a 2x TC or to simply enlarge the image from the prime lens by a factor of 2. Below is an example that gives a clear answer. The image made using the 2x TC is clearly much better then the 2x enlargement of the image from the prime lens.

The bottom line on TCs is that the lower the power, the better the image quality, so a 1.4x will generally yield better results than a 2x. Stacked TCs don't work well. There may be some situations in which they are usable, but not many.

An Image

Finally an image shot with a 2x TC on the 500/4.5L in the field (1/250s at f9, ISO 800 on EOS 10D) . It's an American Bittern. This is actually a crop of a larger image.This is just to show that although the strict test using a resolution chart makes the use of a 2x TC look undesirable (and it is if you have a better alternative), when you need the 2x, it's possible to get pretty decent images, especially if you have them in the digital domain and so can optimize them.

And here's a 100% crop of the Bittern's head. Not bad for a 2x, though both images shown here did have a little help from my image editor in boosting contrast and unsharp masking. A 4x6 print looks excellent.

Recommendations

TCs are very useful accessories but they work best when low power TCs are used with high quality prime lenses. If you're looking to preserve the maximum amount of image quality from your prime lens, use a 1.4x TC and preferably one made by the lens manufacturer. 3rd party TCs can be quite good, especially when used with digital cameras with small sensors since in that case the image quality at the edge of the full 43mm image circle used in full frame 35mm format isn't seen by the sensor. See my article here on photo.net comparing a Canon 1.4x with a Tamron 1.4x TC.

2x TCs will drop image quality but can yield good results when used with care an in conjunction with a quality prime lens. Putting a 2x TC on a 100-300 consumer zoom will probably drop you into the "unacceptable image" region so bear that in mind if you're looking for a really cheap way to get a long lens. "Cheap", "good quality" and "telephoto" aren't words that go together!

Stacking TCs is a last resort. Maybe you can get away with it sometimes on the right subject, but even with quality lenses and quality TCs your chance of getting usuable images isn't good.

Where to buy

Most of the photo.net affiliate stores have a selection of manufacturers and 3rd party TCs. Using these affiliates via these links helps support photo.net.

Readers' Comments

Add a comment

Jonathan Swank , April 30, 2004; 02:37 A.M.

good article. I had actually wondered about the practicality of stacking teleconverters... a question tho...the bokeh in your "no parking" photo is really ugly! is this due to the stacked TCs, the effective aperture with stacked TCs, or something else? I'm sure the bokeh of the lens without TC looks much better! Obviously different TCs may influence bokeh differently...maybe this just isn't a good combination to do it with...

Pradeep Raghunathan , April 30, 2004; 04:11 A.M.

Another wonderful article by Bob, this is such a timely article, with the number of questions on TC's in the forums. Way to go Bob.

Anton Browne , April 30, 2004; 05:27 P.M.

A very useful article. I'd just like to say that I have found a Pentax 2X converter used with the standard 104 f2.4 lens on the Pentax 67 excellent for portraits. It enables closer focussing and I haven't noticed any softening but I figure this would probably be an 'advantage' anyway. The only danger is remembering the 2X exposure factor when transferring readings from a hand held meter in a hurry.

Anton

A. Thomas Burke, Jr. , April 30, 2004; 11:54 P.M.

I presume you do mean teleconverters for the slr that fit between lens and camera. Teleconverters placed on the far end of the lens may increase the F:Stop more or less depending how they are designed. Olympus has some excellent glass TCs for its IS series and digital that go on the end of the lens and do NOT add to the F:Stop. Others have some that have a higher F:stop than the behind the lens converters of an SLR.

Additionally, depending on the design and lens coatings of the SLR TCs there can be more than a 1:1 ratio of T:stop to F:Stop.

Tom

Bob Atkins , May 01, 2004; 12:16 A.M.

Correct. This article only addresses TC that fit between the lens and an SLR camera body (film or digital).

There are no "front of the lens" TCs for SLRs as far as I know - at least none of any quality made by any well known lens maker. Generally TCs (high quality TCs) are intended for use with large aperture telephoto lenses. Even if you could make a high quality "front of the lens" TC, the cost (and weight) would be prohibitive, since for a 600/4 lens it would have to be at least 6" in diameter!

The major application of the "front of the lens" TCs is for small diameter, non-interchangable lens applications on P&S type digicams, where there is no alternative (since you can't put one between the lens and camera body).

Mark Overton , May 01, 2004; 08:48 P.M.

Interesting article! My one suggestion would be to re-do the comparison of 50% cropping versus a 2x TC. In the photo shown, the cropped version is inferior due to pixelation, which means you've become limited by the sensor's resolution. I'd be interested in seeing the comparison done with a fine grained film. I suspect the comparison would then show roughly equal quality between cropping and a 2x TC. Anyway, thanks for writing this article.

Bob Atkins , May 01, 2004; 09:42 P.M.

With any decent lens and quality TC, a 2x will give better results than enlarging the original image, even on film. For one thing grain will be twice as bad on the 2x film blow-up. As far as resolution goes I've done tests on fine grain film in the past. A lens that showed 80 lp/mm on film dropped to about 60 lp/mm with a 2x TC. For a 2x enlargement to be just as good (from a resolution viewpoint), the 2x TC would have to drop resolution to 40 lp/mm.

Sulka Haro , May 03, 2004; 08:28 A.M.

Regarding the already commented bad bokeh issue in the "no parking" sample - did you have a UV filter on your lens when you took that image? There was a discussion on similar bad bokeh on DPReview.com a couple days ago and the problem turned out to be the UV filter which was interfering with the sensor of the 300D used to shoot the images (which has essentially the same sensor as the 10D used by Bob). The discussion is here:

http://forums.dpreview.com/forums/read.asp?forum=1029&message=8538576

I found it quite odd that a UV filter can cause such a unfocused area aberration but I guess that's possible. :)

Bob Atkins , May 04, 2004; 09:42 P.M.

No front mounted filters. I never shoot lens tests with filters (except for the rear mounted filters which are required in the long "L" series telephoto lenses).

The "No Parking" shot was made with a 500/4.5L, which does have a rear filter of course, but that filter is part of the optical design of the lens.

Dan Barthel , May 05, 2004; 07:07 A.M.

In general my results are the same as Bob's. I almost always have the 1.4x mark II mounted on my 500f4. I do have the 2x mark II extender, and in tests it performs much better than Bob's mark I results. I've also found that the 1.4x stacked behind the 2x mark II produces results no worse than the 2x alone. The Canon 1.4x extender is a jewel.

YC _ , May 05, 2004; 08:01 A.M.

the Mark II 2x I have also performs better than the results shown here and consequently, so does stacking the 1.4x and 2x. However, I agree with the overall conclusion that you should avoid stacking TCs whenever possible.

Bob Atkins , May 06, 2004; 09:45 P.M.

It's helpful if people state which lens thay are using a particular TC on. A TC may match one lens better than another. As I said earlier, the performance of the 2x Mark I and the 2x Mark II were very similar indeed on a 300/4L and 500/4.5L. It's possible that results on a 500/4L would have been different.

Bob Atkins , May 13, 2004; 05:27 P.M.

Ospreys, EOS 10D, 500/4.5L + 2xTC + 1.4xTC

This is just a followup on the stacked TCs issue. I had occasion to stack TCs to get decent shots of an Osprey nesting platform. All I had with me was a Canon 2x and a Tamron 1.4x. so I stacked them on my 500/4.5L and took the shot seen here. This image is, of course, optimized, downsized and sharpened in an image editor, but even so it makes a very acceptable 4x6 print and a good web image. Better than I would have predicted!

Miha Hadl , May 29, 2004; 05:09 A.M.

I just want to add that all written facts (so far) are considering the use of lenses with built-in focusing mechanism. If you use lenses + TCs without this mechanism (like ones on Mamiya RB67 / RZ67) which focus with a help of bellows, some of the statements do not apply. The close focus distance of a lens does change since the extension distance changes. For example: the close focus distance of a 90 mm lens + TC equals the focusing distance of the 180 mm lens, while with the 180 mm + TC you get the same close focusing distance as with the 360 mm lens. As I have tried, while doubling the focal length the close focusing distance multilpies approximately about 2,8x.

The close focusing distance and also the magnification ratio changes differently when using TCs in connection with bellows or extension tubes (rings) for macrowork. It is crucial to know, where to put the TC (behind the lens or behind the extension) since the TC itself represents a bit of extension. Further details you can read in the John Shaws book about the Close-up technique in the field.

Otherwise this is a very useful article often to be referred to. Thanks, Bob.