[thoughton]

I guess the question I have is when you state Sigma 70-200/2.8 MK1 what version exactly do you have.

I have the APO 70-200/2.8 EX (non DG non Macro) This lens works very well when paired to the APO Teleconverter 2x EX DG, at least in my opinion. I also know that later releases of the sigma lens were reported not as sharp at 200mm.

That's the same lens I have, non DG non macro, just the plain old APO one.

I guess I should clarify that the Sigma 2x TC is good, for a 2x TC. I am a bit hesitant to actually use it though. I don't have any such reservations with the 1.4x TC.

---

Last edited by thoughton; 02-15-2012 at 09:05 AM.

[normhead]

Teleconverters are worth using if you are close to your subject. My birds are all within 20 feet and they look sharp, but if I'm doing landscape, sometimes I just go the the river steady my camera on the bridge and shoot the tree down at the curve in the river. If I shoot 300 feet away at 250 with the DA*60-250, then 180 with the TC on, the TC image looks like a water colour painting. Sometimes I like that, but usually not. I wish I'd kept those images now so I could show you... the thing is I've never shot 300 feet with a dedicate 400mm lens, so I don't know if it's worse than its could be, or if that's just the IQ you get shooting that image at that range... if anyone wants to send me 400 mm lens, I'll try out a few side by side shots and post them.

I think it's important not to have too high expectations if you're thinking of great pictures of girls on the nudie beach from your tree half a mile away.

[demp10]

I moved the tripod back and forth to keep the fields of view roughly similar. No PP apart from brightness adjustments so they all looked similarly bright.

I am not sure I understand the objective of such test.

One uses a teleconverter not because of its image qualities (although some can render interesting ones) but because of reach. If you simply can walk toward the subject to fill the frame, there is no reason to use a teleconverter.

A better test comparing apples to apples will be to keep the distance the same, then shoot a series with and without teleconverters, then crop the images without the teleconverter to cover the same area and then enlarge them digitally (use your favorite algorithm) so they have the same pixel count.

Once you have two images of the same size (pixel wise) one directly from a lens+teleconverter, and another cropped and enlarged, then you can make comparisons at the pixel level and select the option that is sharper. Keep in mind though that unless you are printing at the same size as you view on your display, pixel peeping will not tell you much about the final print if it is much smaller size.

For a better, real life comparison, you should print out photos of the same size covering the same area, taken from the same distance, cropped as needed and compare the results.

[thoughton]

The objective is simply to see what sort of degradation the 1.4x and 2x TCs produce when compared to the bare lens. IMO we can discount atmospheric haze, so moving the tripod a few feet back and forth should have no impact on the sharpness (or lack thereof) of the results.

I wanted photos that were the same size for easy comparison. Moving the tripod seemed like the most sensible way to achieve that.

I don't understand why you would want to enlarge a sample digitally, using your favourite algorithm. Surely different algorithms will produce different results? How is that a test of the TC? You are introducing a new factor for no good reason that I can see.

---

Last edited by thoughton; 02-15-2012 at 11:02 AM.

[Greyser]

I’m with demp10 here. Let’s say you have a bird, sitting 20ft. away. If you come closer it’ll fly away. So you have two choices here: to use a bare lens and aggressively crop the output or to use a TC and crop less. At relatively short real wild birding distances (15-30ft) TC works better for me, leaving more room for further printing enlargement and less distorted image. Of course, if your target is a backyard chicken, it’s better to get closer to it and use a regular bare lens.
I do some lens/TC tests of 2D targets for myself sometime in a way how demp10 described it. I call it “real life shooting test”. Changing distances technique doesn’t make sense for me.

[thoughton]

Are you being serious? It's not a 'technique'. It's a test to see the actual degradation you can expect from using a TC.

[demp10]

The objective is simply to see what sort of degradation the 1.4x and 2x TCs produce when compared to the bare lens. IMO we can discount atmospheric haze, so moving the tripod a few feet back and forth should have no impact on the sharpness (or lack thereof) of the results.

I wanted photos that were the same size for easy comparison. Moving the tripod seemed like the most sensible way to achieve that.

I don't understand why you would want to enlarge a sample digitally, using your favourite algorithm. Surely different algorithms will produce different results? How is that a test of the TC? You are introducing a new factor for no good reason that I can see.

Anything you add in front of a lens (e.g. filter) or behind the lens (e.g. teleconverter) will affect the overall quality to some degree so there is no debate if a teleconverter will degrade the image; it will. The question is how much and most importantly, is the result usable.

Comparing the images without a reference to an objective is rather academic. For all practical purposes you could also included a short macro lens at much shorter distance and got even sharper images. Then one may concluded that the shorter the focal length the sharper the image, which is largely true. For a more focused test you must set a goal and see how each lens combination satisfies that goal. The simplest test will be to make the same picture of a small bird using both combination.


If you need to make a 8x10 print at a given dpi (e.g. 250 dpi) of a small bird covering a certain percentage of the frame (e.g. 50%) from some fixed distance (because you cannot go closer), you have two basic options.

1. You use your lens without a teleconverter, crop the image accordingly to frame the bird and then print. The printing software will do all the enlarging as needed (and if needed) to get the desired dpi.
2. You use a teleconverter to fill the frame so no cropping (or a lot less) is needed. The printing software will not need to enlarge that much (or even may have to reduce the resolution) to meet the desired dpi.

Then you compare the prints. What you expect to find is that at smaller print sizes, when the cropped image has enough pixels to meet the desired dpi without any digital enlargement, the lens without the teleconverter will always outperform. As the print size gets bigger, the demand for more pixels to meet the dpi requirement increases and the driver must digitally enlarge the original image to make for it.

A sharp original image can be digitally enlarged and still be sharper than a fuzzy larger image up to a point. It is that point that will determine the largest print that can be made from the lens without a teleconverter and still be better. Above that, the lens+teleconverter combo will be better.

Being better though does not mean that it is acceptable also. It can be sharper but both can really look terrible at great sizes.

By digitally enlarging the cropped image, you simulate the effect of printing at larger magnification to achieve the same size so the comparisons are at the same level.

[Greyser]

Are you being serious? It's not a 'technique'. It's a test to see the actual degradation you can expect from using a TC.

It's very academic. It does not seem too practical to me (please see my post above).

And I totally agree with very well put demp10's explanation.

[newarts]

My motivation for using a teleconverter is almost always to get a larger image of a subject that's too far away. Therefore the test that applies to me is if the cropped photo without TC is as good or better than with TC.

[Greyser]

My motivation for using a teleconverter is almost always to get a larger image of a subject that's too far away. Therefore the test that applies to me is if the cropped photo without TC is as good or better than with TC.

Exactly.

[Lowell Goudge]

One other thing to consider, aside from the comfortable distance to shoot small little wild critters is the limitations of a lens itself. You can be limited by the close focus distance of your lens. You have3 options, we have discussed 2 of those here, a TC or crop, but the third choice is an extension tube.

Keep in mind however, a TC does not alter the MFD of a lens, so a shorter tele with a TC might just have a closer focus distance than a long lens, and at the end of the day the biggest native image on a sensor still is important

[pacerr]

. . . to get a larger image of a subject that's too far away.

In the "good ol' days" with only MF prime lenses (35, 50, 135 & 200) my thinking was "how can I position myself to format and frame the image with the len(es) I have available?" That was a composition-based mindset and sometimes the choice of vertical or landscape orientation to support final framing/cropping/printing was as controversial as the use of a TC today. (There was also that controversy over the "horrible" image degradation produced by those silly new zoom lenses despite the fact that some very good PJ work was being done with 'em.)

My transition to APS-C with a an Oly Pen-F kit and more recently the DSLR bodies included adopting some of those 'modern' zoom lenses and an unnoticed, subtle but significant shift in my thinking about lenses occurred. Focal length became a handle for distance and zooms a lazy excuse to manipulate a scene by image size rather than perspective.

Thinking through some of the discussions in the earlier photo forums re-aligned my thinking to the field-of-view / angle-of-view and point-of-origin perspective again.

If you plant yourself in the position needed to get the desired composition and perspective, lens choice including TC use becomes obvious and easy when restricted to what's available at the moment.

A TC then becomes a force-multiplier rather than an IQ-destroyer allowing you to make the best possible image with what you have in hand. Unless you have a stack of test-bench comparison shots, you either like the TC results or you don't -- as with any other shot -- and the TC itself becomes a non-issue.

Just use the best tool you have at hand. If you don't know which one that is (yet?) then the whole point of arguing over IQ degradation is mute anyway.

An' don't forget to have fun too.

H2

[thoughton]

Anything you add in front of a lens (e.g. filter) or behind the lens (e.g. teleconverter) will affect the overall quality to some degree so there is no debate if a teleconverter will degrade the image; it will.

Thank you for sharing that, although I don't recall debating about whether a TC will degrade the image. May I suggest that you read my post again? The relevant sentence was "The objective is simply to see what sort of degradation the 1.4x and 2x TCs produce when compared to the bare lens".

The question is how much and most importantly, is the result usable.

I'm only interested in "how much". I don't agree with your "most importantly" - whether the result is usable or not is highly subjective.

Comparing the images without a reference to an objective is rather academic. For all practical purposes you could also included a short macro lens at much shorter distance and got even sharper images. Then one may concluded that the shorter the focal length the sharper the image, which is largely true..

Macro lens? Why are you talking about another lens? All my photos were taken with one lens. You appear to be talking about someone else's comparison.

For a more focused test you must set a goal and see how each lens combination satisfies that goal. The simplest test will be to make the same picture of a small bird using both combination.

I look forward to your comparison photos of the same bird, both with and without a TC. To make a sharpness comparison as easy as possible it would be convenient if the bird was in the same pose in each frame.

If you need to make a 8x10 print at a given dpi (e.g. 250 dpi) of a small bird covering a certain percentage of the frame (e.g. 50%) from some fixed distance (because you cannot go closer), you have two basic options.

1. You use your lens without a teleconverter, crop the image accordingly to frame the bird and then print. The printing software will do all the enlarging as needed (and if needed) to get the desired dpi.
2. You use a teleconverter to fill the frame so no cropping (or a lot less) is needed. The printing software will not need to enlarge that much (or even may have to reduce the resolution) to meet the desired dpi.

Then you compare the prints. What you expect to find is that at smaller print sizes, when the cropped image has enough pixels to meet the desired dpi without any digital enlargement, the lens without the teleconverter will always outperform. As the print size gets bigger, the demand for more pixels to meet the dpi requirement increases and the driver must digitally enlarge the original image to make for it.

A sharp original image can be digitally enlarged and still be sharper than a fuzzy larger image up to a point. It is that point that will determine the largest print that can be made from the lens without a teleconverter and still be better. Above that, the lens+teleconverter combo will be better.

Being better though does not mean that it is acceptable also. It can be sharper but both can really look terrible at great sizes.

By digitally enlarging the cropped image, you simulate the effect of printing at larger magnification to achieve the same size so the comparisons are at the same level.

This is all very interesting for people who print images. All of mine are displayed digitally, either on websites, in digital photo frames, or via the screensaver on my 40" TV.

[newmikey]

One issue often overlooked is exposure accuracy. As a TC effectively multiplies the focal range of a lens by 1.4X or 2X (or 1.7X with the Pentax), the camera's matrix metering will compute its optimum exposure on the basis of the image crop presented to the sensor. If you do not use a TC and crop in PP to accomplish the same FOV, the metering will have been done on the full image possibly leading to different exposure results.

As to optical degradation: it often is a toss-up between losing pixels due to cropping or losing resolution due to glass.

[Lowell Goudge]

One issue often overlooked is exposure accuracy. As a TC effectively multiplies the focal range of a lens by 1.4X or 2X (or 1.7X with the Pentax), the camera's matrix metering will compute its optimum exposure on the basis of the image crop presented to the sensor. If you do not use a TC and crop in PP to accomplish the same FOV, the metering will have been done on the full image possibly leading to different exposure results.

this is a valid point, but most people shooting wild life shoot spot metering or otherwise compensate for the subject exposure to get it right. only someone who either makes a mistake or lacks experience, would use full matrix metering on a subject they know or should know they are going to crop down.

but there is another issue with teleconverters and exposure that should be discussed. all cameras to some extent, have exposure errors that are compensated by the camera body because the body knows the open aperture of the lens. This is evident when using legacy (manual aperture) lenses, and was quite a topic for discussion with specifically the K10D and K20D which had a really bad non linearity of metering with legacy lenses. Most teleconverters do not correct for the magnification and F stop reduction that results from adding the TC, and therefore the camera makes all its metering decisions based upon the native aperture opf the lens. On the K10D this results in a .7stop error with my SIGMA 70-200/2.8 and a 1.4x TC and a 1.5 stop error when using the 2x TC. This also screws up P-TTL flash, which also wants to know the true aperture. The only TC I have that corrects for aperture is the SMC-F 1.7x AF TC.

As to optical degradation: it often is a toss-up between losing pixels due to cropping or losing resolution due to glass.

Unfortunately it is not just this trade off. The TC by making the lens longer and slower also impacts the selection of shutter speeds and aperture. It is virtually impossible to properly evaluate addint a TC to a prime, and compare the images directly, withoput also comapring the TC plus prime to an equivelent prime at the new focal length.

Many of the sharpness issues are due to either subject movement or camera shake that are highly related to the use of a longer lens irrespective of whether it is a prime or prime plus TC.