Canon EF70-200/2.8L IS USM with Canon Extender EF 2x II and EF 1.4x II
by Jeff Medkeff
The Extenders
I purchased Canon's current (as of early 2005) 1.4x and 2x teleconverters for three
purposes:
- To get a bit longer reach from my kit, so that I could (hopefully) dabble in wildlife
and bird photography in 2005;
- to learn longer-lens techniques inexpensively;
- to delay the purchase of a long telephoto until I could determine from experience what
specifications I would find best.
In other words, I wanted to cheat my way to a longer lens until I decided exactly what
I wanted from a long telephoto.
At the time, the longest lens in my kit was the Canon 70-200 f/2.8 IS L, which is a
workhorse for me. Using the 1.4x and 2x teleconverters would transform this lens into the
equivalent of a 98-280 f/4 and a 140-400 f/5.6, respectively.
The teleconverters are small and lightweight. The 1.4x converter is about the size of
Canon's EF 50mm f/1.8 lens. The 2x converter is a little over twice that length and a
little heavier. They are well-built and hold the heavy 70-200/2.8 IS rigidly. Because
optical elements and their mounting hardware protrude beyond the length of the
teleconverter's barrel, the teleconverters cannot physically be attached to all Canon
lenses; a list of compatible lenses is maintained by Canon.
Performance at similar Focal Lengths and Ratios
For the first test, I took a photo of the eagle on her nest with the 70-200 IS, set to
200mm. The lens was mounted by its collar to a Bogen 3221 tripod and 3029 pan/tilt head, a
support that is capable of holding a considerably heavier load. The Image Stabilizer was
turned off. The photos were taken with a Canon 20D digital SLR, using mirror lock-up, at
ISO 100. The attempt was to take all shots at the same focal length in order to compare
the effect that the teleconverters have on an image at the same pixel scale. Unfortunately
I missed the 200mm target by a hair when using the 1.4x converter. The "actual
reported focal length" is the focal length recorded by the camera in the headers of
the raw file.
The raw file was converted using Adobe Camera Raw, using identical settings for each
image (these settings are Adobe's defaults, except that I converted into the sRGB color
space). I then cropped the center of the picture to 640x480 pixels, and from within
Photoshop I saved for web at the same jpeg quality setting for each image. None of the
images were properly white balanced or corrected (daylight white balance, as shown here,
left the eagle's head cool; balancing for shade left it warm; I decided not to apply a
custom white balance to each image prepared for this test to save time). In all cases, the
eagle is near the center of the frame.
Above: 70-200 at 200mm, no teleconverter, at f/4.
Above: 70-200 zoomed to the (estimated) 142mm mark, 1.4x
teleconverter (for an actual reported focal length of 190mm), at f/4.
Above: 70-200 zoomed to the 100mm mark, 2x teleconverter (for an
actual reported focal length of 200mm) at f/5.6.
My conclusion is that the teleconverters significantly degrade the image compared to
the 70-200 alone, at comparable focal lengths. Both teleconverters are used wide open in
the above photo. The unconverted image is at f/4 to provide easy comparison to the 1.4x
converted image, but this lens appears to be equally sharp at all focal ratios until very
slow ratios introduce diffraction. Therefore the unconverted image's sharpness in the
focus zone is representative of the lens' sharpness at most focal ratios.
Interestingly, the teleconverter performance above is significantly worse than when the
lens is fully zoomed, as can be seen below.
Eagle photos at various f/stops:
The following shots were taken of the same nest and eagle, but from a slightly
different location and during different lighting conditions. The attempt here was to
determine whether the teleconverted lens had any "sweet spots" in its range of
focal ratios. The teleconverters were used with the lens zoomed to the 200mm mark (giving
280mm with the 1.4x converter, and 400mm with the 2x converter). A series of shots were
taken at each full stop from wide open to f/32.
1.4x Converter:
Above: 70-200 at 200mm, with 1.4x (280mm) at f/4.
Above: 70-200 at 200mm, with 1.4x (280mm) at f/5.6.
Above: 70-200 at 200mm, with 1.4x (280mm) at f/8.
Above: 70-200 at 200mm, with 1.4x (280mm) at f/11.
Above: 70-200 at 200mm, with 1.4x (280mm) at f/16.
Above: 70-200 at 200mm, with 1.4x (280mm) at f/22.
Above: 70-200 at 200mm, with 1.4x (280mm) at f/32.
2x Converter:
Above: 70-200 at 200mm, with 2x (400mm) at f/5.6.
Above: 70-200 at 200mm, with 2x (400mm) at f/8.
Above: 70-200 at 200mm, with 2x (400mm) at f/11.
Above: 70-200 at 200mm, with 2x (400mm) at f/16.
Above: 70-200 at 200mm, with 2x (400mm) at f/22.
Above: 70-200 at 200mm, with 2x (400mm) at f/32.
My conclusion is that the f/1.4 teleconverter performs best at f/8 to f/11, or a stop
on either side; the 2x converter appears to peak at f/11 to f/16. One of the significant
problems with both converters is the conspicuous blue shade at the border between the
eagle's white head and brown body (and sometimes between the head and background); also,
at some focal ratios the backgrounds may not be considered pleasing.
Teleconvert, or Resample?
The question of whether to use a teleconverter, or just resample a shorter focal length
image to a higher resolution, should be answered. Two of the following images were made
from the 200mm image above (without teleconverter). They were processed with an s-spline
algorithm with mild sharpening to a pixel scale comparable to that provided by the 1.4x
and 2x teleconverters. They are placed here adjacent to the best teleconverter images to
facilitate easy comparison.
1.4x Resampled vs. Teleconverted:
Above: Above: 70-200 at 200mm, no teleconverter, at f/4, resampled
using s-spline to 1.4x the original image scale.
Above: 70-200 at 200mm, with 1.4x (280mm) at f/8.
2x Resampled vs. Teleconverted:
Above: Above: 70-200 at 200mm, no teleconverter, at f/4, resampled
using s-spline to 2x the original image scale.
Above: 70-200 at 200mm, with 2x (400mm) at f/11.
In both pairs of images, the comparison is favorable in areas that are not very
detailed (such as the eagle's head and eye), but not favorable elsewhere. In particular,
the feather edges on the eagle's back do not fare well in the resampling - both
teleconverters are significantly better than the resampled image.
Crazy Talk
Sometimes you just can't resist stunts. A friend had his Canon Extender EF 2x (type I)
with him, and he handed it to me to add to my own 2x converter. The two converters went on
the 70-200, which was zoomed to 200mm, and I took some photos. Mathematically, this should
provide about 800mm at a widest aperture of f/11, but a design specification was clearly
being exceeded. The camera reports a widest available aperture of f/5.6, and the headers
indicate the available focal lengths are 140-400mm. TTL metering still worked fine, of
course. The image below was taken in aperture priority mode with the aperture set to f/8.
Focus was manual.
Above: 70-200 at 200mm, with two 2x (approximately 800mm) at a camera
setting of f/8 (actually f16).
Where to Buy
Purchasing from these vendors via these links helps support photo.net.
