"From Light to Ink" featured the work of Canon Inspirers and contest winners, all printed using Canon's imagePROGRAF printers. The gallery show revolved around the discussion of printing photographs...
Getting photographs right in the camera is a combination of using your imagination, creativity, art, and technique. In Part 3 of this three part series, we focus on shooting strategy and the role of...
[Note: Photo.net wishes to thank
ADORAMA for the loan of a demo lens
for this review]
The EF-S 18-55/3.5-5.6 is an unusual Canon EOS lens. It was designed
only to be mountable on the Digital Rebel (EOS 300D) body since it has a reduced
image circle (enough for the Digital Rebel sensor, but not enough for full frame
35mm) and it sticks further into the camera body than a normal EF series lens.
[Editor's Note: A variant of this lens is also part of the Canon Rebel XT
when sold as a kit with a lens. That lens is called the Canon EF-S 18-55m
f/3.5-5.6 II. It has been reported that the newer model has the same
optics as the old model and the only differences are cosmetic.]
The "S" in EF-S stands for "Short back focus". Back focus is the distance from
the rearmost point of the last optical surface of the lens to the film or sensor.
There are certain advantages in lens design in having this distance small for
wide-angle lenses, but the limit in SLR cameras is determined by the requirement
that the SLR mirror must not hit the lens when it flips up. The mirror on the
Digital Rebel is smaller than that of full frame 35mm cameras because the sensor
is smaller and the mirror only needs to be large enough to cover the area of the
sensor. This allows the use of a lens with a shorter back focus distance.
Focal Length & Maximum Aperture:
11 elements in 9 groups (1 aspheric
Diagonal Angle of View:
75° 20' - 27° 50'
Inner focusing system with MM
Closest Focusing Distance:
0.28m / 0.92 ft. to infinity
Max. Diameter x Length, Weight:
2.7" x 2.6", 6.7oz. / 69mm x 66.2mm, 190g
The EF-S lens was greeted with what perhaps amounted to over expectations by
some people. They reasoned that since it didn't have to cover the full 35mm frame
and since it could be made to operate closer to the film than a standard lens,
it's performance would be much better than expected for a $100 lens. Well, maybe,
but what they failed to realize was that the reduced back focus distance is
only present between about 18mm and 22mm. From 22mm up, the back focus distance
of the lens isn't any shorter than can be used with a regular EF lens, and the
image circle (or at least the illumination circle) is actually large
enough to cover a full 35mm frame (though obviously not at high quality judging
from the MTF curves below). The difference in minimum back focus distance between
an EF lens and an EF-S lens is also pretty small. Maybe 1 or 2mm. So while
reduced image coverage and shorter backfocus may help a little at the wide end,
it's not really reasonable to expect them to make a $100 kit lens work like a
$1000 "L" series lens! If you really want all the highly technical stuff, here
are the MTF curves (from
Canon, not from
me!). The solid/dashed, blue/gray and thick/thin lines represent different
apertures and spatial frequencies. These are explained on the
website. If you compare them with other Canon MTF charts, be sure to note
that the horizontal axis only runs from 0 to 13mm, not 0 to 22mm because the lens
is only specified for use with 22.7 x 15.1mm sensors which have an image circle
of about 27mm (vs. 43mm for 35mm full frame)
Before anyone asks (and they will ask) "Why didn't you test it against
the 24/2.8 or the 28.2.8 or the 28-135IS or the 35/2 or the 24-85 or any or all
of the other Canon lenses covering between 18-55?", I'll tell you. I have a life
and there's only so much time for lens testing. Plus I didn't actually
have every other lens, plus - and this is the most important
point of all - it's a $100 lens. You don't
need to agonize over whether or not to buy it. If it's
halfway decent and you want a general purpose lightweight lens for a Digital
Rebel, it's good enough.
I should add that I actually tested two different EF-S 18-55 lenses, but all
the images here are from one lens. While there were very minor differences in
image quality between the two lenses, just as there are likely to be between any
two lenses, they were small enough - certainly less than the differences between
different lenses - that they would not affect any of the conclusions drawn in
this article. I also took many more shots than you see here. The images
reproduced in this article are examples. The text refelects the results from the
full set of images.
So is it good enough....?
First let's take a look at the "lens testing range" that I used! It's actually
mostly my neighbor's garden, but it provides a convenient subject with detail in
the center of the image and tree branches against the sky in the top corners. The
areas outlined in yellow are those used in the test at 18mm. At other focal
lengths the center was the same, but different branches occupied the top corners
of the frame. Note that the images were shot over several days, so there maybe a
change from sunny with blue skies to cloudy with gray skies, however all the
shots at any particular focal length were taken under the same conditions.
Since I don't have any other lenses that cover 18mm, the test at 18mm was done
using only the 18-55 lens, shot at apertures from wide open (f3.5) to f16. The
first set of images below shows the center performance as a function of aperture.
It's pretty clear that it doesn't change much. I think there is a hint of
sharpening going from f3.5 to f5.6 and a hint of softening going from f5.6 to
f16, but you have to look hard to see a difference.
However, as shown below, the situation in the corner of the frame is quite
different. Wide open the image is pretty soft. Stopping down only 1/3 stop to f4
shows a perceptible improvement, but it's still soft. Stopping down one more stop
to f5.6 shows a large change with the image sharpening up significantly. Stopping
down another stop to f8 again shows an improvement, but further reduction of
aperture to f11 and f16 doesn't really buy you any more sharpness.
So overall, at 18mm the lens is good in the center but poor in the corner when
wide open. Best performance comes at f8. Stopping down to f16 doesn't get you any
more corner sharpness, but does soften the center of the image slightly.
Conclusion: At 18mm shoot at f8 or f11 for best image quality. Image quality
under those conditions is certainly quite acceptable.
I did have a 20-35/3.5-4.5 on hand, so at 20mm I was able to compare the two
lenses directly. Note the image magnification here is quite a bit higher than
that used for the 18mm test to better show the differences between these two
First let's look at the corner of the image:
It's pretty obvious here that the corner image quality is still pretty bad
wide open for the 18-55 lens. The 20-35 is a lot better. Stopping down to f8, the
18-55 gains a lot of sharpness and the 20-35 gains a little. The 20-35 is better
at both apertures.
Below are sections from the center of the image:
Here all the images are pretty similar, though the 20-35 is maybe a little
softer than the 18-55 at f4. Differences in the center of the frame are much
lower than at the image corners. I'd say the 18-55 maybe has a very slight edge
here, not bad for a $100 lens! Again the 18-55, when stopped down a bit, is a
Moving on now to 28mm and using a 28-105/3.5-4.5 as the comparison lens let's
first take a look at the center of the image:
Not a huge amount of difference. Again it looks as if the 18-55 may be
slightly sharper than the 28-105 at f4 but things are even at f8. Stopping down
from f4 to f8 doesn't change the 18-55 image much but it does slightly sharpen
the image from the 28-105.
Corner performance is shown below:
Not a huge amount of difference here. Both lenses are pretty similar and
stopping down doesn't really do a whole lot. Again a very creditable performance
from the 18-55, certainly more than acceptable at 28mm.
The final test was done at 50mm and the 18-55 was compared with both a 50mm
prime and a 28-105/3/5-4.5. First let's look at the edge of the image at f5.6
(wide open for the 18-55).
Not much doubt here that the 18-55 is the worst of the group. The 28-105 is
clearly sharper, and the 50mm prime sharper still. So what happens when we stop
down to f11?
Well, things now look more equal, however the 18-55 still seems less sharp
then the 28-105, though not by very much. The 50mm prime is best, which shouldn't
be a big surprise.
Below is the center performance of each lens at f5.6. Pretty similar quality,
but the 18-55 is maybe just a little softer than the other two lenses.
Again I think the 18-55 gets a decent rating at 50mm. Not as good
as a 50/1.8 (get one if you can find one!), not quite as good as a
28-105/3.5-4.5, but still very usable.
Just in case you're wondering how much difference an 18mm setting
makes makes, compared to say a 20-35, 24-85, 28-105 lens or even the old 22-55,
here's a shot with approximate frame lines. The gain over a 24 or 28mm lens is
significant. The gain over a 20mm lens (interpolate between 22mm and 18mm) isn't
Does the 18-55 exhibit the dreaded digital "purple fringing"? The
short answer is yes, it does. It's visible at all focal lengths and apertures.
Purple fringing is normally mostly due to chromatic aberration from the lens (we
won't get into debates about whether blooming or the sensor optics - microlenses
and filters - are involved). Most lenses - especially wide angle lenses -
especially zooms - show some degree of purple fringing and the18-55 is no
exception to this. There's nothing you can do to reduce chromatic aberration in a
lens if it's there to start with, Stopping down does not help. You can
play tricks in PhotoShop post-exposure which can reduce the effect.
Above are 200% crops from 3 frames shot at 50mm. On the left is a
shot from the 28-105 at 50mm. The "fringing" here is actually more red than
purple, but its visible. The center shot shows the 50/1.8 lens performance and
you can see that color fringing is slight. On the Right is the 18-55, which is
clearly showing a significant amount of purple fringing. As I said above,
stopping down doesn't help.
Note that such aberrations are MUCH more visible
when you shoot a dark line against a very light background. In the above images,
the white background is blown out and registers at or close to 255, 255, 255
R,G,B. The fringes are actually the result of the blue/purple component of the
intense white background being slightly displaced due to chromatic differences in
image magnifications (otherwise known as lateral chromatic aberration). On a
"normal" subject, even though it's there you may not even notice it. How much of
an issue it is depends on what you shoot and what your requirements are.
Below is a 100% crop of the corner of a shot taken with the 18-55 set to 50mm
(F11) on a more "normal" subject. There's not a lot of blue in this image and
there are no regions where there is an abrupt change in the blue content, hence
the result of a small shift in the blue component of the image is virtually
unnoticeable and chromatic aberration effects don't cause a significant loss of
If you routinely shoot images of tree branches against white
skies, then purple fringing will be a matter of great concern for you. If you
don't, then it may not matter all that much. It's never a desirable quality in a
lens, but sometimes it's not as much of a handicap as you might think.
Like every lens, there's some flare if you have a bright object
in or just outside the frame. The solution for this when the light source is
outside the frame is an efficient lens hood or other method of shading the lens.
I didn't do a lot of flare testing but I did take few shots with the sun just
outside the frame at 50mm and f5.6. The full frame view is shown below
Below is a comparison of the 18-55, 50/1.8 and 28-105/3.5-4.5.
Quite surprisingly, while all 3 lenses suffer from flare, the 18-55 seems to show
least effect. The top row of images are without any shading. The bottom row if
images are with the lens shaded (with my hand!). Note that a standard lens hood
would probably not have been very effective for the zoom lenses, since it has to
stay outside the frame at the widest lens setting, and an 18mm (or 28mm) lens
hood on a 50mm lens isn't very efficient and won't block any light source which
is only just outside the frame.
Finally here's a shot at 18mm (at f11) with the sun actually in
the frame. Yes there's flare, but it's no more than you might expect in this
Overall I'd say the flare performance of the 18-55 is above
average. It's still a good idea to use a lens hood though, and an even better
idea to use more efficient manual shading (hat, hand, card) under really tough
The 18-55 is actually pretty good. It gives 0.28x (a little over
1/4 life size) at the closest focus distance (11" from the sensor plane, about 5"
from the front of the lens) at 55mm and distortion is very low. In comparison,
the best a 50/1.8 II can do is 0.11x (1/9 life size). Barrel distortion is pretty
bad when close focused at 18mm and corner image quality isn't great, but who uses
an 18mm macro lens? (I should know better then to ask that if I don't want angry
email from the 18mm macros shooters club, but I'll risk it).
55mm macro shot. Coverage = 5mm x 7.5mm
Summary and conclusions
The 18-55/3.5-5.6 is a very good lens for $100. It's weaknesses show up at the
ends of its zoom range (18mm and 55mm), especially in the corners of the frame
when used wide open where resolution and contrast can be rather poor. To get
decent sharpness in the corners requires stopping to f8 at 18mm or f11 at 55mm -
but you still won't get the same image quality as you would from a a prime or "L"
series zoom lens. In the mid range (28mm), performance is pretty good at both
center and edge. For a cheap, plastic mount, low end consumer lens, the 18-55 is
certainly better than I would have expected. It's not a rival for an "L" series
lens, but much of the time it can hold it's own against Canon's full frame
coverage mid-range consumer lenses, especially in the center of the frame.
Considering the price ($100) and the zoom range (18-55mm), this lens doesn't
really have a lot of competition. I'll rule out the 16-35/2.8L since it's over
$1300. I'm sure it's better. If it's not heads will role at Canon... Yes, there's
the 17-40/4L, also certainly better based on published reports, but it's around
$650 and it covers less range. The 20-35/3.5-4.5 is a bit closer in price ($340)
and sharper wide open at the wide end (20mm), but it's not quite so wide and the
zoom range is smaller. The 28-105/3.5-4.5 isn't really an alternative if you're
looking for a wide-angle, though it only costs around $200 and performs
better in the 50mm region. It also zooms to almost 2x the focal length of the
18-55 - but again, it doesn't give you the wide-angle end. The 50/1.8 II is
cheaper ($70) and better, but it doesn't zoom and it's not a wide-angle. Of
course ALL of the alternatives can also be used on a full frame digital or film
body, so that's an important factor if you're shooting with both media.
There is a low cost, small sensor coverage Sigma 18-50mm lens, but it's
normally only sold as a pair with their low cost 55-200. There may be a few
places selling it on it's own (gray market imports). I haven't seen any reports
of image or build quality on this lens. It's possible it may be a viable
alternatve to the Canon lens, but my experience with low end Sigma lenses hasn't
been positive in the past.
So for $100 you can't really go wrong with the Canon EF-S 18-55/3.5-5.6. Other
lenses will certainly outperform it over various parts of its range and if you
ever intend to shoot film again (and some of us do), the 18-55 will be useless to
you since it won't mount on any film camera and even if you modify it with a
hacksaw, the mirror of a film SLR will hit the rear of the lens at 18mm - and
even if it didn't there would be severe vignetting. So if you want to shoot
digital AND film, the other choices will allow you to do that and the 18-55
To me it seems like buying the Digital Rebel Kit which includes the 18-55 for
$999 (vs. Body only at $899) seems like a no brainer. It's a great "walking
around" lens with the same angular coverage as a 28-90 lens on a film camera.
It's small, it's light, it's cheap and as long as you know it's limits, its a
good performer. Of course there might be times when a 20-35 or a 50 or a
28-105 would yield a technically better image, but lenses are only useful if you
carry them with you. A lightweight 18-55 on the camera is better than a bag full
of larger, heavier "L" lenses left back in the car or at home!
Photo.net again wishes to acknowledge
ADORAMA for the loan of a demo lens
for this review.
Where to buy
This lens is not sold separately and is only available bundled as a kit with a
Canon Digital Rebel or Canon Rebel XT. The Canon Digital Rebel has been replaced
by the Rebel XT and new models of the original Rebel are now difficult to find.
You can buy the Rebel XT kit which includes the EF-S 18-55mm f/3.5-5.6 II and the
Canon Digital Rebel XT body. The kit is available from the
following photo.net affiliate stores. Buying via these links helps to
support the photo.net website.