Category Archives: Canon 5D

Canon’s landmark full frame camera

Canon 5D sensor dust

Could this be the reason?

I was Googling the subject of sensor dust, which so seems to bedevil the 5D, and came across a very funny criticism by one user who may well come from eastern Europe, given the grammar. What it misses in terms of the Queen’s English it gains in clarity.

“Is vacuum pump, not camera”.

So it got me to thinking. Could the lens design have something to do with it? After all, per square inch of sensor, the 5D should gather no more or less dust than its siblings, cropped or full frame, yet the 5D seems to be a problem more often than the other models.

Part of it may be that users of the full frame sensor in the 5D are either consistently making big enlargements or pixel peeping in Photoshop just to appreciate the gorgeous definition of which this sensor is capable. Bottom line, they enlarge more, because they can, and more magnification means more dust becomes visible. On the other hand, a like-sized print from a cropped sensor camera requires a 60% greater enlargement ratio, so maybe this is not a good explanation.

Yet why is it that I can put away my 5D with 24-105mm in place – my ‘standard’ lens – only to find that sensor dust has reappeared even though the lens has not been removed?

So I took the 24-105mm off the camera and, holding the rear to my cheek, worked the manual zoom ring. Sure enough, a ‘whoosh’ of air could be felt when the ring was activated vigorously. You can gauge the stroke here with the lens set at 24mm and 105mm, respectively:


Focusing seems to generate no air rush, probably because the lens uses internal focusing. Likewise for my 85mm f/1.8 and 200mm f/2.8 lenses. By contrast, the 15mm Fisheye and the 50mm f/1.4 use traditional focusing, but the throw is so short that no detectable rush of air could be felt using the ‘cheek test’.

So let’s assume that cropped sensor Canon users for the most part avoid the 24-105 (the effective range of 38-168mm being far less useful to them than the 24-105mm on a full frame sensor). So 5D users, many of whom favor this lens, do indeed have a ‘vacuum cleaner’, or more correctly, an air pump, attached to their cameras. The only thing I cannot figure out is why 1D and 1Ds/Mark II full frame users rarely complain of this malady. Maybe they opt for the 17-40mm, 16-35mm and 70-200mm L lenses, all of which seem to have a superior reputation for dust sealing? In this regard, the 24-105mm is anything but ‘pro’ quality, though there’s no arguing with the superb optical performance.

For me the cure is probably to avoid using the 24-105mm in dusty, dry conditions, opting for fixed focal length lenses in lieu. Not very satisfactory.

Maybe someone out there has done some comparative testing of the issues and causes?

All this said, the net throughput even when sensor dust intrudes, is still exceptional, especially if a ‘roll’ needs the use of the stamp & clone feature in Aperture, which allows simultaneous removal of dust motes in the image from multiple pictures at the same time.

Canon 5D – can you say speed?

The fastest camera I have ever used

For over thirty years my camera of choice for street snaps was a Leica M2 or M3, typically with a 35mm or 50mm lens. This was the street snapper’s outfit par excellence.

Leica, of course is clueless about how to make a digital camera – they either badge engineer Panasonic’s and increase the price or make an overpriced, dated version of the fabulous M2 in the crippled-sensor M8.

In my experience, after using the 5D for some 15 months now, the 5D with the 24-105mm L zoom is faster in every respect. True, it’s a good deal bulkier than an M2 and not a bit noisier, but the automatic exposure and mind-numbingly fast and superbly accurate autofocus blows away the excellent Leica rangefinder in every way. We are talking operating speed here. Not high speed motor drives which are of little use in these situations. No I mean the speed of execution from seeing the picture to its recording.

Case in point. In continuing pursuit of pure color, and the hell with form and traditional composition, here’s the sort of thing which is second nature to this powerful lens/camera team (not to mention the operator):


Tattoo You. 5D, 24-105mm at 93mm, ISO 125, 1/350, f/4.5

It is hard to describe just how fast I had to be to snap this; the whole raising the camera/zooming/composing/snap cycle is just a blur in my memory. Exposure? Automatic. Focus? Automatic. Zoom? Manual – forget motorized zooms. Result? You be the judge.

And, while we’re at it, here’s another pair:


Twin Carb. 5D, 24-105mm at 105mm, ISO 250, 1/1500, f/5.6

Beating the burn

In overexposed highlights, that is

I have written before of the tendency of digital sensors to burn out highlight details. While highlights can be recovered using the Highlights slider in the Aperture Adjustments HUD, this is limited to one stop using RAW, in my experience. Thereafter, not all highlight details can be recovered.

Accordingly, in high contrast situations like outdoor sun, it’s far better to underexpose and use the Shadows slider to bring up shadow detail, rather than trying to recover highlights. The technique is illustrated here. A low noise sensor, like in the 5D, can sustain a lot of shadow enhancement before noise rears its ugly head.

Strangely, I find the large sensor in the Canon 5D more susceptible to highlight burn than the miniscule one in the Lumix LX1. Tfhe 5D’s sensor is some 1/2 stop more sensitive than the indicated ISO, compared with the Lumix. Given that HDR cannot be used with dynamic subjects (the three or five images required dictate the use of a tripod on a stationary subject) I simply underexpose by 1-2 stops in high contrast situations. Single image tone mapping can help, but it adds maybe half a stop at best; any more and the effect is garish. Canon provides exposure compensation on the 5D but is is horribly documented in the miserable book with a miniscule typeface that passes for instructions that comes with this camera. For $3,000 for a body only, this has to be the height of cynicism. Canon, please exclude accountants from the design of your machines and instruction books.

The 5D has a two position power switch, illustrated above. (The peeling on the screen is my stick on protector, not delamination of the LCD!). Normally, the switch is clicked up one notch to ‘On’ when using aperture priority – Av on the top left dial. Click it up one more notch to the line and the rear wheel activates exposure compensation, visible on the bottom of the viewfinder. It’s also displayed on the top LCD screen which is much easier to see than the viewfinder readout. By the way, a Manfrotto QR tripod plate is visible in this snap, permanently attached to the 5D’s base. Highly recommended – this is the ‘Architectural’ version with an alignment lip to preserve its position on the relatively heavy 5D body.

Click the on-off switch to the line, take a first pressure on the shutter release button, and rotate the wheel while looking though the viewfinder or at the top LCD screen. You can elect 1/2 or 1/3 stop intervals using the custom functions in the camera’s software. 1/3 is confusing precision with accuracy in my book, so I have it set to 1/2 stop intervals. I dial in the camera to, say, -1.5 stops, then immediately move the switch back to the regular On position to preclude accidental adjustment – the compensation setting wheel is disabled in this way, although the setting you dialed in is preserved. Better still (or not, if you forget), the setting survives switching the camera off. Power up and there is your setting, preserved.

I prefer to use average exposure metering in fast paced outdoor settings (Canon’s matrix metering leave me unimpressed) as there’s rarely time to take a proper exposure reading in the interest of capturing the moment. With this approach, you gain a stop of highlight adjustment while preserving some three stops of shadow recovery. Now that’s what I call dynamic range.

Here’s an example taken in bright sun yesterday:


5D, 24-105mm at 70mm, ISO 250, 1/3000, f/5.6, one stop underexposed

Without the underexposure, the white sheet would have been comprehensively burned out. Here, detail is preserved.

Mirror slap in the 5D revisited

The Custom setting on the mode dial fixes things

In my little experiment to determine the seriousness of mirror slap-induced blur, I concluded that locking up the mirror before taking long exposures on a tripod made sense. Sharpness was improved – not something that would make any diffference for regular prints or web images, but clearly an improvement with big prints.

At the same time I grumbled about the difficulty of finding the right setting on the LCD panel on the rear of the 5D when it came to actually locking up the mirror, a problem compounded by the poor visibility of the LCD screen outdoors. Well, until someone comes up with a neat software fix to reprogram the little used ‘Print’ button on the back of the camera to lock up the mirror, here’s the next best thing.

Set up your camera for your preferred mode of use, go the the LCD screen and set the mirror to lock-up on the first pressure on the shutter release, using Custom Function 12. Now set the mode dial to ‘C’ (the dial to the left of the prism atop the body), go back to the LCD screen and click on ‘Register Camera Settings’. In this way, anytime you set the mode dial to ‘C’ you will have mirror lock-up available.

In my case, the preferred settings are Aperture Priority (meaning I set the camera up with the Mode Dial at ‘Av’), ISO at 250, center area average metering and the RAW file format.

So next time the camera goes on a tripod I will simply move the mode dial to ‘C’ and off we go.


The 5D’s Mode Dial set to ‘C’

Mirror slap in the Canon 5D

Does it matter

Hot on the tail of yesterday’s gripping piece extolling the virtues of camera supports, whether monopods or tripods, it ocurred to me to make one more experiment to see if the slapping rise of the mirror in the 5D has any effects at longer shutter speeds. To magnify things further, I placed the 200mm f/2.8 Canon L lens on the camera and mounted the latter on my super sturdy Linhop S168 tripod. You can see the cantilever braces between the legs and center column here – note the latter is retracted for maximum stability.

My mechanics professor would be pleased with those cantilevers and I, as a photographer, think it’s the only way to make a tripod sturdy without excessive weight. He used to lecture extensively on Euler struts and how light and strong could live together with proper design – someone at Linhof was listening. Leonhard Euler, who died in 1783, got it right first time around and his math is taught to all budding mechanical and civil engineers to this day.

I chose a different picture in the book, just for fun, as no benefit accrues from using the same one. To take human shake out of the equation I released the shutter using Canon’s plug-in and wildly overpriced RS-80N3 wired remote control. Cable Release to you and me.

All pictures of the target were processed in Aperture.

I once more used 1/8th at f/5.6; 1/8th is very slow (like 1/2 second with a 50mm lens) but not so slow that any vibration from mirror slap becomes too low a percentage of the exposure time, if you get my drift. In other words, any camera vibration will be a significant component of the total exposure time.

Here’s the full frame snap – once focus was set I switched to manual focus to prevent any changes:

The enlargement ratio in the following snaps would yield a 40″ x 27″ print. No sharpening has been added.

Here’s the result with normal mirror operation:

While the difference is not great, the picture taken with mirror lock-up is a tad sharper – look at the contrast of the lettering.

And here’s the one with the mirror locked up using Custom Function #12 in the Canon’s menu system:

By the way, the anti-aliasing of the image in the 5D dictates that sharpening be used for best results. In the following rendition of the ‘no mirror’ snap, above, I have used Aperture’s Edge Sharpen with the variables for Intensity/Edges/Falloff at 0.70/0.22/0.57 – slightly more modest than the defaults of 0.81/0.22/0.69 suggested by Apple’s settings. As you can see, the absence of grain (at ISO 160, as before) and high resolution from the superb Canon 200mm lens makes 40″ prints from the 5D a piece of cake. You are looking at a JPG conversion from Aperture. In reality, a print from the RAW file will be sharper with no artifacts.

Maybe a ‘Medium Format is Dead’ piece is required to go along with ‘Black and White is Dead’ and ‘Film is Dead’?

Now Canon, when the 5D Mark II is released, can you please assign the mirror lock-up feature to a mechanical button, the sort humans like, rather than to a blasted choice in the myriad selection of Custom Functions on that silly little LCD screen? The fact that the latter is unreadable by landscape photographers outdoors does little to help, in addition to the insane complexity of actually finding the function.