A total beginner's perspective

I am a total beginner to DSLRs, and have been given my brother's 5D Mk III (Canon Camera Museum, Amazon, DPReview) to use indefinitely. I have been shooting film since the late seventies and SLRs since the mid-eighties. I am almost entirely a still shooter, not video, and I click for pleasure, it's not my job.

The 5D is also my first AF SLR. My last few SLRs were the T90, AE1P, and FTb. Therefore, when I was handed the 5D Mk III, I first had to figure out how this auto-focus animal worked. I'd seen DSLRs with friends, but never used one for more than half a dozen potshots. My review of the 5D will be very unusual -- it will only be of help to another beginner to AF and DSLR who jumps in with both feet with this model. I doubt whether there's another 5D owner whose first AF SLR is the 5D. This review may not be of value to experienced DSLR users. Nevertheless I will soldier on.

At first glance

The 5D (I'll use just "5D" to mean the "5D Mk III") is well built, comparable to the T90. The VF image is as good as the top-end manual-focus SLRs I've used. The VF information is (finally!) just right -- you get aperture, shutter speed, ISO, over-or-under-exposure indicator on an analog scale, focus confirmation and one more item which I don't find important. And I'm not used to the luxury of dioptre adjustment in the VF -- the T90 and AE1P didn't have it. I've grown older since I got those cameras, and love the feature, because I need it. (Most of you take the feature for granted, I guess.)

Getting started

It doesn't take time to get started with the 5D. Battery, SD card, switch it on, turn the mode dial to the green "A" and you can click your first photo. And those photos are not very satisfactory. You then move the mode dial out of the green "A", read the manual, and go to bed in a daze -- that manual can make a grown man's head spin. I'm a post-grad in engineering, I've written software for fun and profit, but the manual intimidated me.

Modes, dials, and states

Took me a few days just to distinguish between the manual's reference to the "main dial" and the "quick control dial". The camera has three rotating dials/knobs: mode, main, and quick control. They're all very visible, and the mode dial does not cause any confusion since it's used relatively rarely. The main dial and quick control dial caused me a lot of confusion. The manual refers to one of them here, the other one there, and these references are in the form of graphics. Initially I failed to notice the different graphics the manual used to denote the two dials. So the manual would tell me to turn the quick control dial to set exp-comp, and I would be turning the main dial.

The second problem I faced was with the state of the camera. When clicking and just before, the camera is in active state (my nomenclature). If the camera is idle for a few seconds, it falls into a semi-passive state, and if it's idle for much longer, it goes completely dormant. The difference? If you're in aperture-priority (Av) mode, then

• in the active state displays current shutter speed and aperture, both, in the LCD panel and VF.
• In semi-passive state, the shutter speed disappears from the LCD, and the VF display switches off.
• In completely dormant state, the LCD panel switches off too.

All of this may be very familiar to old EOS users (which most readers of this review will probably be). But I was confused. How does it impact the user? Well, certain controls will only work in the active state and not in semi-passive state. For instance, if you are in Av mode and want to set exp-comp, the quick control dial will only work if the camera is in the active state, not if it is semi-passive. But when you (or a new user like me) are using the camera, you will never guess this because you won't know there's a difference between the active and semi-passive states. I learned to set exp-comp only after using the 5D for about two months, purely because of this confusion.

The shutter button is supposed to have two levels: the half-pressing and the full pressing. The manual says that half-pressing the shutter activates auto-focus and full pressing takes the shot. In fact, there's a one-third-press of the shutter button too -- it changes the camera state to active, wakes up the meter, and shows you the VF information with metered exposure settings, without triggering the AF.

Focusing correctly

The first hurdle I had to cross was the less-than-perfect focus. My first few shots did not have the pin-sharp images I was hoping to see. I first thought that AF technology is capable of just this accuracy, no more -- I'd seen this same slightly imperfect focus in other shots friends had taken with their DSLRs too. The first few days went in experimenting with the auto-focus. I quickly realised that the completely auto-mode for AF area selection doesn't work well. In this mode, the camera keeps all 61 sensors active and selects a few which it thinks cover the subject area, then focuses the lens to make that part of the image sharp. I switched to manual control of AF area. After some experimentation, I have remained with manual selection of single AF area out of the set of 61. I now decide which AF sensor the camera will use, and I place it over the most important part of the image, and the AF works exceedingly well. I just had to discover how to use it correctly.

I have used the AF in places so dark that I cannot see clearly the image I am clicking. I push the ISO up to 6,400 and click, and I get a focus lock. Then I see details in the image on the computer screen which I could not see with the naked eye or through the VF. I don't know how good the AF is compared to other cameras, but I don't think I want anything better than what the 5D gives.

The 5D supports three AF modes. In "Oneshot" mode, it focuses on a (stationary) subject and stays there; you lose focus if the subject moves. In "AI Focus" mode, the camera locks on but then can re-focus if the subject moves. This is what I usually use. In "AI Servo" mode, the camera never locks on, and continuously tracks a moving subject. I have tried the "AI Servo" mode, letting the camera focus on my boy on a swing or a large pet dog running towards me. I get about 60-70% shots accurately focused. I am happy with this. I wonder how good this performance is for pro sports photographers. The "AI Servo" mode's AF tracking characteristics can be selected from some half dozen different options. I have retained the "Standard" setting which is the default. I find it difficult to even understand the descriptions of all the settings, and will experiment with them some other time, when my gizmo-love exceeds my desire to click good images.

Metering the light

The 5D gives you four metering modes: evaluative, centre-weighted full-frame, partial area, and spot. All are commonly found in all pro SLR bodies these days. The evaluative metering, which Nikon calls their Matrix Metering, is what auto-everything snapshotters are expected to use. My T90 gave me centre-weighted, partial and spot. When you put the 5D into the green-"A" mode, you can only use evaluative metering.

The evaluative auto-everything metering mode seems to work surprisingly well even for back-lit subjects upto a point. I wanted to try the spot metering, and I saw that it works well. Having used the T90 for a long time (they used to call it "a spot meter with a camera body around it"), I feel a bit naked working with a body which does not have it, though I use it only rarely.

The 5D's metering has an unusual feature. If you operate in evaluative metering mode and manually select an AF area for auto-focusing, the metering system concentrates its light reading overwhelmingly around the selected AF area. This really helps most of the time. If you switch out of evaluative metering mode to spot metering, then the central 2% of the frame is used for metering, like the older T90-style spot metering I am used to. Took me a few months to begin to understand this intelligence in the 5D's evaluative metering, and it explains why this meter mode works so well -- it almost seems to read your mind sometimes.

On closer testing, I now realise that the bona fide spot meter gives different results from the implicit spot metering coupled with the AF area. The bona fide spot meter is the one I get when I set metering mode to "spot" and point the central 2.5% of the frame on the area I want to meter. I use a single AF sensor as my AF area, and the evaluative meter is supposed to use just this bit to meter -- that's the implicit spot meter. But when I tried pointing this area on a dark patch, with brightly lit patches in other parts of the frame, I found that the bona fide spot meter gave me readings about one stop brighter than the implicit metering from the evaluative meter. And after clicking some test shots, I found the bona fide spot meter's reading more accurate. I presume the evaluative meter continues to look at other parts of the frame even if it emphasises the AF area's little patch, while the bona fide spot meter is more purist, and completely ignores the rest of the frame. See samples here. In these samples, the area of interest was brightly lit and the surroundings were dark. So the bona fide spot meter got the balloons correctly exposed and the dark surroundings went pretty under-exposed. With evaluative metering, I placed the AF area on the balloons but it took a more "averaged" view of the light level differences. I will keep this discrepancy in mind.

High ISO and low light

I tried high ISO photography. I pushed the ISO to 6,400, even to 10,000, to shoot hand-held in artificial light. Everyone says images at high ISO settings add noise. I wanted to see at what ISO setting the noise became unusable. My benchmark: 4000dpi scans of mass-market film frames. I found that I could operate at 6,400 ISO, even 10,000 ISO, in low light and still get images cleaner than scanned images from film. I clicked photos at night in the open during Diwali when my five-year-old was burning fireworks, lit only by the light from fireworks. I've always wanted to click such shots, and I never got good results with film and manual focus. The 5D not only makes it possible, it makes it easy. This is in part due to the cleanness of the image, in part due to the 5D's amazing low-light AF ability. I kept the camera on evaluative metering the whole time, and clicked a hundred frames over two hours. Some frames were over-exposed because the light level changed sharply between time I half-pressed and full-pressed the shutter button -- that's how it is if fireworks are your only light source. Of course, I had to learn how to work the row of buttons on the top plate beside the LCD panel first before I could even change ISO. Each of those buttons has two functions. You press the button you want, and then use the main dial to set one attribute, the quick control dial for the other attribute. I learned all this only after I learned to distinguish between the two dials. (Yes, I feel silly, but there it is.)

Balancing the white

I also messed around with white balance a bit. I was tired of having to correct the blue colour cast in all photos clicked in open shade -- it's been an irritating little part of my photography since I started processing my images on the computer. So I tried pushing the WB setting off AWB and tried "Open Shade". I was expecting the camera would now do the equivalent of an 81A or 81B on the lens. I discovered a few interesting things.

When a frame has a lot of red-orange-brown, it applies a milder correction, say an 81A. When there is less of the earth colours in the frame, it corrects more, and sometimes over-corrects. You can see the degree of correction changing simply by moving closer or farther away from the face of the subject in a portrait, when there is a lot of browns and maroons in the background. I was expecting that setting "Open Shade" would give me a fixed amount of colour correction, like fixing a specific filter on the lens. I think that's not how the 5D works. The same thing works both ways. If you set the camera to AWB and click in open shade, the camera gives you more or less blue tint depending on the mix of colours in the frame. I haven't done objective tests, but this is the impression I'm getting.

Another thing I noticed was that the "Open Shade" setting seemed to be working like a proper 81B on a film camera -- it was cutting the exposure a bit. Those brown-tinted frames appear very slightly less exposed than similar ones clicked with AWB. Of course, a physical 81B on a good SLR would not impact exposure -- the TTL meter would compensate for the filter. Here, in the all-software-cybernetic DSLR, I seemed to see a slight darkening in the image. All this was valuable, at least for me.

I tried to see how well the "Tungsten" WB would correct images clicked indoors, with warm-daylight CFL lighting as the only light source. Warm-daylight CFL only approximates tungsten filament lamps only very roughly. I was wondering what the 5D would do. It seems the 5D in "Tungsten" WB does not deal with warm-daylight CFL -- the images were as greenish-yellow in "Tungsten" setting as in AWB setting. Strong colour correction with "Colour --> Curves" in GIMP was called for.

One of the interesting aspects of my Diwali firecracker clicking session was that I set WB to "Daylight". I wanted the camera to not try to correct for colour casts -- those colours are the reason fireworks are so attractive. I was happy with the results -- I got all sorts of crazy colours. Just what I wanted.

Clicking in the RAW

I click JPEG; I refuse to click RAW. (Am I the only 5D user who does this?) I click for pleasure, and I hate the thought of poring over 40 MB files for five hours after clicking for an hour. I do not have colour-corrected computer screens either. But then I click for my pleasure, and looking at large prints from the JPEG files the 5D generates gives me giant gobs of that pleasure.

Inspired by Luminous Landscape's review, I downloaded a trial version of the Capture One software from Phase One, and tried a few RAW images. I set the 5D to record both RAW and JPEG, and compared the JPEG which Capture One generated (all settings at default) with the one the 5D had originally produced. It was a highly casual test, if you know what I mean. I could see no difference, so I switched back to pure JPEG. I use the "Fine Large JPEG" image format, where each file is between 5 and 10 MB. Maybe someday I too will begin to look down at JPEG capture and will start messing with RAW file processing.

Since I use the JPEG the camera provides, I have to set various controls in various menus to ensure the camera does not do too much processing. So I set a few settings

• I set the Picture Style to "Standard". I tried "Neutral" for a bit but found that I am touching up all my images in GIMP to make them look like "Standard" images, so I switched back to "Standard".
• I switched off Auto Lighting Optimiser. Keeping it on is too risky, one might as well use the green "A" mode.
• I also set High ISO Speed Noise Reduction to "Low". (All my Diwali firecracker photos were with this setting.)

These give my camera-generated JPEGs a certain degree of predictability.

Six months of living together

I got the 5D in April 2012, when it had not yet reached all shops. And I've clicked a couple of thousand frames by now. Today, I am as comfortable with the 5D as I've ever been with any SLR. I regularly use

• the row of double-function buttons beside the LCD
• the two dials, the main and the quick control
• the menus -- quite easy to navigate once you know how
• the image playback and adjustable magnification features
• the LCD, the VF information
• the AE lock button
• I am beginning to like the "Silent" drive mode where the mirror moves slower than usual, but the camera makes much less sound

The controls for selecting the AF area are fiddly. I can't claim I can do it easily or effortlessly, and I have to move the camera away from my eye each time, but I cannot imagine how an improved interface would operate either. Some may feel that a touchscreen interface, where one touches the point which one wants to use for AF sensing, will be easy to use, but I am not sure I want a touchscreen interface anywhere on my SLR.

And I do not use

• the "Q" button and "Rate" button
• any of the image post-processing functions built into the camera
• the "Lock" switch below the quick control dial
• the M-Fn button
• the green "A" mode
• Live View mode (using the LCD instead of the optical VF for viewing and shooting)

This is as much a comment on me and my clicking habits as on the 5D.

I say I am as comfortable with this body as I've ever been, but I can't help noticing that a DSLR seems to be a more deliberate engagement. There's more of left-brain thinking for every frame than I had experienced with my manual-focus film SLRs. There were no focus points to worry about, no ISO to change. You just kept focusing, shooting, and winding. Focus, shoot, wind. It used to work faster and more instinctively, somehow. Maybe this too will become instinctive with the years.

Currently, I use the 50mm f/1.8 II lens and the 100mm f/2.0 USM lens. I am perfectly happy with these. When I find the money, I'll buy a 24mm f/2.8 lens, and maybe I'll be rich enough to buy the 24mm perspective-control lens someday. I refuse to use the popular kit zooms for my clicking. I hate the dim VF image from those small-aperture lenses, and I do not want to suffer reduced AF accuracy with them. Even between the 50mm and 100mm lenses, I seem to find that the 100mm focuses better in difficult low-light situations than the 50mm does. Both are more than good enough for my needs, though.

I have printed images at 20"x30" and even 30"x40" sizes, from shots clicked hand-held. The details I can see in the prints thrills me -- I have seen such sharpness and fine detail with film only when everything turned out right. Extremely sharp images are easier to get with the 5D than with my old film equipment and darkroom chain.

There is an idiomatic expression: "Is he man enough for the job?" I want to say that the 5D is "camera enough" for any job I can possibly throw at it, or have thrown at any camera in my three decades of clicking. I will continue to use film SLRs, for black-and-white clicking. I just love the look of enlarger-projected and chemical-washed B&W prints very differently; not a comment on the 5D. For colour and for low-light shooting, I'm totally satisfied with the 5D.