by Guy Lerner
April 7, 2003
(Note: This is NOT a review of the Canon EOS 10D. For my full review of this camera, go here. This article takes an in-depth look at some of the issues reported with the new 10D, and looks for solutions to common problems reported by 10D owners.)
Ever since the much-anticipated release of Canon's new digital SLR camera, the EOS 10D, reports have abounded on problems linked with the camera. Some of these “problems” (such as overexposure) I mentioned in my initial review, and others (such as focus) I didn’t test for at the time (mostly because I didn’t find a problem to test for in the first place).
I decided to conduct further testing based on these reports, more to determine if there was a demonstrable problem (with my own camera and lens combinations at least) than to prove or disprove any of the reports. Irrespective of my findings, I do believe some users have genuine recourse for complaint with faulty cameras and/or lenses, simply because, as with any large production run of a consumer product, there will be a margin of manufacturing error. It may be small to the point of insignificance in the bigger picture, but that doesn’t help the handful of individuals unlucky to be caught up in it.
All About Focus
I received several e-mails from readers asking to know my
view on the reported focus “issues” with the 10D. These centered primarily on
focus miscalibrations resulting in either front- or back-focus errors with the
10D and certain lenses. For those of you not familiar with the terminology,
front-focus means the camera is focusing slightly in front of a subject (having
confirmed a focus lock); back-focus is the inverse, the camera focuses slightly
behind the subject.
Before I ran the tests for front- and back-focus, I noted feedback from other reviewers and users that the 10D seems to light up its active focus points in the viewfinder independently of focus lock. In other words, the focus points light up to display which points are active in finding the target, but do not indicate that focus is achieved. That’s done with a small green circle in the bottom right hand corner of the viewfinder. If that’s the case, users of cameras other than the 10D, which confirm focus lock with the focus point indicators, could be tripping the shutter before focus is actually achieved, resulting in misfocused pictures.
I can’t say for sure if this is a feature of the 10D, or a bug that needs squashing. It does, however, depend on the focus mode the camera is in. For example, in one-shot AF mode, both the active focus point and focus confirmation light will light up. However, switch to AI Servo mode, and page 66 of the manual states:
When focus is achieved, the focus confirmation light in the viewfinder will not light and the beeper will not sound. If the focus confirmation light in the viewfinder blinks, it means focus has not been achieved.
This can lead to confusion or worse, poorly focused pictures. Of course many users have reported poor focus using the focus confirmation light, so it’s not a general caveat in this case. What you do need to ensure, if you’re running tests of your own or recording a high percentage of poorly focused pictures, is that you’re always in one-shot AF mode and are paying attention to the focus confirmation light and audible “beep” to indicate focus lock.
One other “issue” that was singled out as a potential 10D problem has to do with the “soft” images the camera tends to produce. I dealt with this at length in my review, and concluded that, if set correctly, the 10D produces perfectly acceptable (that is, sharp and defined) pictures out of camera, and can be set to produce even sharper pictures with the slider controls available.
Poorly focused pictures will appear very “soft” on computer, and no amount of sharpening will bring them back. This leads me to think that some “soft picture” problems with the 10D have more to do with poor focus (either on the part of the camera, or the user) than in-camera sharpening. Hopefully the results of my tests will shed some light on that assumption.
The Tests
For the focus tests I used a very simple (read unscientific)
setup to maximise accuracy. I carefully pasted a sheet of newspaper on a wall
about two metres from my tripod. The tripod was levelled (using three level
indicators on the tripod itself), and the camera fixed to the tripod at a
height that allowed me to zoom in on one well-defined point on the target. This
meant the target was on a flat surface exactly (but perhaps not to the
micro-millimetre) parallel to the lens plane. (I would later run another test
(see below) with a target set at 45 degrees to the lens plane).
Two zoom lenses were used for the test: a professional-grade Sigma 70-200 EX f/2.8 HSM lens, regarded by many as the equivalent to Canon’s highly-prized 70-200L f/2.8 lens; and a consumer-grade Canon 28-135 f/3.5-5.6 IS lens, a very popular lens for its high zoom range, image stabilisation and generally excellent image quality.
Pictures were taken at different focal lengths with different apertures. Since poor focus is typically noted with a shallow depth of field (i.e. with the lens open to its largest aperture), shot sequences were taken starting from the largest aperture (f/2.8 on the Sigma, f/5.6 on the Canon) and working down to around f/8. Two shots were taken at each aperture, using the centre focus point only in one-shot AF mode; the first relied on the camera’s AF confirmation, and the second on manual focus (I threw the lens out of focus and refocused manually to take the second shot in each set).
The images below are 100% crops from the respective test pictures. I could post full-size images or our service provider would shut us down, so you’ll have to go on these crops and my own observations of the full size pictures. In each picture I’ve indicated which lens, focal length and aperture was used. The frame on the left is the AF exposure, and on the right my manual focus exposure.
(Note: I assume most of you won’t be running off to take pictures of wall-mounted newspapers with your expensive cameras and lenses. This test uses a controlled setup to minimise any inaccuracies that could crop up in random testing in the field. The results obtained will therefore even themselves out in practice, so that minor focusing errors in all likelihood will go unnoticed in real world situations. That’s why many reviewers don’t include them in their write-ups).
Besides all the apparent conclusions we can draw about each lens’s relative sharpness, colour balance, and so on, what’s immediately obvious is that I found no significant front- or back-focus problems with either lens at these settings. You could argue some of the frames (like the AF frame of the Sigma at 200mm) are “visually” soft, but that’s more a condition of the lens wide open at its longest point than a focus inaccuracy.
You could also argue my manual focusing technique is way off par, something that surprised me looking at the results, and something I am working hard at to rectify. I did find after some of these shots were taken that my dioptre was slightly miscalibrated, which would result in my not being able to accurately judge critical focus through the viewfinder. Subsequent frames were much closer, and almost identical, to the AF exposures.
What it does show quite blatantly is that poor focus and soft pictures go hand in hand. Next time you look at a picture taken straight from the camera and consider it soft, think carefully whether the softness is a result of poor focus, poor technique, or camera or subject movement, before flagging a fault with the camera.
I have stopped short of including the pictures from the f/5.6 and f/8 exposures because in each case they are either visually identical or sharper than these examples. That’s because as each lens is stopped down it not only gets closer to optimal sharpness, but depth of field also increases, which could hide any front- or back-focus inaccuracies I would have found wide open. As it happens, there were none.
Also remember that what you are looking at are 100% crops of unsharpened images shot at each lens’s largest aperture. I explained in the review that almost every picture taken with a digital SLR will have to be sharpened to some degree, because of the filtering it is subjected to in-camera. Compare the crop to the full-frame image (much reduced, below) from the Sigma. Minor focus inaccuracies detected at 100% won’t be visible when the picture is printed full size, especially when printed smaller than 8”x10”, something to keep in mind if your final output is going to be print.
Another test
Just to be sure that the results of the wall tests aren’t
skewed, I set another newspaper flat on my desk and close-focused on a small
section of text with the lens at 45-degrees to the page. Again I opened up both
lenses to their largest apertures; with the added magnification, depth of field
was extremely narrow so any focus inaccuracies would become immediately
obvious.
The only other difference, equipment-wise, to the first test was that I used a Canon 500D close-up lens on the Sigma zoom, because without it the minimum focusing distance was about six feet, hardly useful for a macro-type test.
In each of the frames below focus was locked using the centre focus point between the “O” and the “R” of “STORY”.
I’ve seen some results posted on the Internet that looked vastly different to these, but all I can offer you are the results from my camera and lenses. In this case, it’s clear that focus was dead-on accurate. I tried these shots about a dozen times over with each lens, using different combinations of aperture, shutter speed and focal length. In each case the pictures appeared properly focused, with the zone of focus well defined and just where it should be. Even the narrowest of zones (using the Sigma at f/2.8, but not shown here) was right over the “OR”, with just about everything north and south of this point rapidly blowing out of focus.
Sharp enough?
Seeing is believing, so I won’t bore you with a scholarly
analysis of what you’re seeing. Still I have no doubt that some of you are
experiencing very real problems with getting sharp images out of your 10D’s, so
here are some thoughts that may help you around the dilemma.
As with any SLR system, many different factors come into play before a final picture is realised. Consider for instance the sheer number of variations available to you when building your system; while your 10D may be new, you could be using lenses several years old, some of which have seen their best days. Then there is the matter of variations in lens quality. Let’s face it, not all lenses are equal, and some are far more likely to produce a sharp, focused picture than others.
There is a reason why a professional grade Canon L zoom, for example, costs several multiples more than a consumer cheapo: Quality. Using better glass, more advanced technology and more stringent testing, some lenses are produced with much more precision, and at greater expense, than others, and the quality shows in the pictures they help produce. A picture made with an inexpensive, poorly calibrated lens will always produce soft-looking pictures, especially at high magnification. Poorly made lenses also do the camera no favours in helping it lock focus.
Lenses aside, often we need look no further than the
photographer behind the camera to find the weak link in the quality chain.
Please don’t take this as a personal insult, but most cameras today are better
than many photographers’ ability to use them. Expensive cameras with expensive
lenses will still not be able to produce sharp, detailed pictures if the photographer
using them disregards the basic principles of photography and employs poor
technique in his picture taking.
For example, how many times have you seen photographers complain about a soft, out-of-focus picture, only to concede that it was taken hand-held using a slow shutter speed in less-than-adequate light? Cameras and lenses will only record what you show them using a narrow band of settings and options; they won’t pull a rabbit out of the hat if it’s not there to be pulled, no matter how much they cost you.
The 10D is a new camera, and as such will take some time to master, even in the hands of experienced professionals. Consider this: racing drivers spend months testing minor modifications to their cars, on tracks they’ve raced hundreds of times before, because every change affects the way they respond to the car and the way the car responds to them. A new tool brings with it new challenges, even though you may be familiar with similar tools that came before it.
You may think the best investment you can make is sharp glass. I disagree; invest some time getting to know the camera, how it responds in different light, and how it balances in your hand. After that, it’s just a matter of using the tool the way it was designed to be used.
Of course, if you’ve tried everything else, your technique is refined, you’ve used the heaviest tripod in brilliant light, then it’s not unreasonable to believe you bought a rotten camera. It’s simply not unheard of to get a handful of bad apples from an otherwise exemplary tree, and while I have faith in Canon’s manufacturing process for the 10D, I can see how a small number of exceptions will make it past the finish line.
I can’t argue for or against the camera you have in your hand; I can only speak for mine. The results of the tests I’ve made speak for themselves, as do the hundreds of pictures I’ve taken in day-to-day use in a variety of conditions on a variety of subjects.
You already know what I think about the 10D. I hope these tests put to rest some of the fears you still have about the camera, and help get the focus back where it should be: your photographs.
Copyright © 2003