Georg N. Nyman PhD.

BACK TO HOMEPAGE

This page is optimized for a screen size of 1280x1024pixels - please adjust your viewing screen accordingly!
Many image files on this page are quite large - a broadband connection is recommended!

CANON EOS 5D
(or 36Hours of Pleasure)

It was only most recently that I got the opportunity to test the Canon EOS 5D from Saturday evening to Monday morning (Thanks to Bryce Cole from Englewood Camera in South Denver) and I must admit, it was a rather busy weekend for me. Of course, I was not able to test all functions in real life, to look at all details, to evaluate all dozens of customization options - what I was able to achieve was getting a pretty good idea of what this camera can do and how good it is.

This review is certainly not a repeat of a user manual or a marketing presentation - I do not get any financial incentive from Canon (and also no equipment, neither free nor discounted) - I try to stay as neutral and objective as possible but I am allowed to have an opinion and that opinion I do voice always.

My very first impression when I got the camera was - oh, that is a real camera, not a high-tech box made in China and out of plastic...nice for a change! The camera is very comfortable to hold and to carry, the weight is pleasant (not comparable with the old Kodak DCS 760) and the most commonly used controls are placed logically and ergonomically at the right place - at least for my hands. You switch the camera on and it is here, no waiting (0.2sec!), no lag, just like a conventional SLR - wonderful!

Before I start with the presentation of the results, I would like to share my impressions about some controls and features:

Viewfinder
What a nice and bright viewfinder - and of course, large as well because the 5D is a full frame camera. I know there are lots of pro's and con's for and against full frame sensors and APS sensors and this review is not the right place to add my two pennies worth, I just wanted to mention - the viewfinder is large, bright and clean - no unnecessary clutter spoils the clear view of the scene.

Viewing screen and Menu
As well - a large screen (2.5inch) and its brightness is really good - not only the image but also the menu listings can be read in most lighting conditions - only in sunlight, there is no way to read anything, but I have not seen any digital screen which is bright enough to be fine in sunshine. The menu is more or less logical, some points I would have placed differently, but one can get used to it quickly. One complaint to do have - the mirror-lock-up function is buried in sub-menus and not where it should be - a separate function which can be chosen quickly and easily. I just do not understand why Canon still did not get this - everyone using a long focus tele lens or does macro photography or runs into exposure times of a few fractions of a second to a few seconds needs that mirror-lock-up function....Canon, please do change this!!

Battery power
The battery power is OK, not outstanding, but fine and OK. After shooting a lot of images with viewing them on the screen, some cleaning up of unnecessary images and some longer exposures times, the battery told me that it was only 2/3 full anymore - that is OK I guess, but I would have expected to see a bit longer battery life (compare with the Nikon D2X...you can shoot and shoot and shoot and still have got an almost fully loaded battery).

Shutter
The shutter is very quiet and that is something I liked very much - that is a significant advantage of the 5D over the Nikon D2X with it's rather noisy shutter. Not that the 5D is as quiet as a Leica M, but quiet enough not to draw attention to the surrounding when you take a picture (Have you ever used a Kodak DCS460 - when you release the shutter, everyone within one mile radius thinks you fired a gun...)

Sensitivity setting, Pixels and Sensor
ISO100 to 1600 with standard setting - fine enough, I do not really need ISO50 and 3200, but it is nice to have and the 5D has got it. Anyway, I will show later some images taken at various ISO settings to compare the amount of noise from low ISO settings to the high one - not much in reality. Canon must have a very good algorithm to reduce noise that efficiently.
And of course the large pixel size (8.2micron compared to Nikon's 5.5micron) contributes in a positive manner as well. For illustration purposes, I have taken the graphic display of the various pixel sizes out of one of the DP-Reviews to show visually the difference between the various cameras again:

Both cameras use CMOS sensors and if one does the math, the D2x with its smaller sensor corresponds theoretically to a full frame 19MP camera but as their pixels are significantly smaller, the gain on theoretical resolution can not be realized due to the less efficient light collection and processing abilities of the smaller pixel geometry - which means that one can expect to see a slightly higher resolution with the D2X but definitely not as much as one could expect from the difference in pixel size - which confirms again that the size of the pixels is not the one and only determining factor for performance and resolution in digital images.

An additional remark - a reader of this review made the comment that the comparison of two cameras with different sensor sizes using lenses of different focal lengths to get the same field of view as result is not correct - I tend to agree but this is the way how a user of a DLSR would do it, I presume.
What is of interest for a photographer is to see what resolution does a camera deliver at a certain field of view, meaning with an APS sized DLSR, one needs to take a lens with a shorter focal length than with a full frame DLSR - the result is that both images with show the same angle of view/field of view. A purist will argue that one needs to take the same focal length and uses therefore a differently sized crop (which shows the same field of view) to compare the two cameras - with the full frame camera, the crop would be smaller than with the APS sized DLSR. In that case, the Nikon D2X would definitely show a better resolution (but more noise due to its smaller pixel size) than the EOS 5D, because more pixels will cover the same small area and therefore produce a more detailed image.
I guess, this question is not easy to answer - which of these two approaches are more appropriate. However again, in real life, I presume that a user is interested in the de facto resolution by comparing images which show the same field of view/viewing angle.

AWB
(Automatic setting) works fine in most cases, better than what I have seen with most other cameras. Of course, it got a limited range of Kelvin to adjust (3000K-7000K) automatically but that is fine for 90% of all shots I guess. And one can adjust for a wider range in a non-automatic mode - if that newly released Expo-disk is used, adjustment to an individual lighting situation is for most situations easy and works well.

Contrast and Color
People who know me, think that I am more a film-photography person than a digital fan but they are not really right with that opinion. I am predominately interested in landscapes, scenic shots and what one maybe can call "quiet photography". I am using 4x5 inch, 5x7 inch and as a hand held camera most of the times either a Mamiya RZ67ProII or an Alpa12SWA.
But the digital technology has become within very short time a serious competitor even to medium format film photography. Still today, all digital systems, I have tested, do not reproduce contrast/shades of color/color "flavor" as accurate or pleasing as film can do, but it is getting very close to film and probably within one year or two, they will outperform medium format film photography as well.
The Canon 5D is a very attractive camera but still, the contrast rendition is not really close to perfect as the overall image contrast is pretty high. There are subprograms which allow for a slightly smoother contrast with less steep curves - and this helps a bit, but still, I do see in critically illuminated scenes too much contrast in the digital shots.
The color rendition on the other hand is very brilliant - the 5D is a color-friendly camera, the images are very pleasing but sometimes maybe a bit too colorful - this can be adjusted as well, but using the standard shooting modes, for my liking. the colors are a tiny bit too colorful.
Something else which I think I should mention is the fact that for my liking the images of the 5D are a bit too soft - they need considerably more sharpening to be crispy than the ones coming out of the Nikon D2X. At very first glance, one could think that the resolution is slightly limited, but no, all the details are existent, they are just very soft and this creates this initial feeling that the images are a tiny bit out of focus - but they are not. When you compare the raw images (in their initial factory default settings) of both cameras, of the 5D and the D2X, you immediately see the difference between them - Nikon's default raw image is crispier and a bit less colorful whereas Canon's default raw image is more colorful and a bit less crispy. But - please keep in mind, I am talking about minute differences, nothing which is degrading any of these cameras - it is like driving a Ferrari and a Maserati - two different approaches to one task with exciting features and performance. And of course with both cameras, all these settings can be changed to match exactly what you like to see - the amount of custom functions (21) and custom settings (57) is almost overpowering.

Maybe I am a bit (over?)sensitive to these differences - coming from several years in photofinishing industry - it was always the known "battle" between the European taste of color and the American taste of color - the photofinishing equipment for US customers needed to be adjusted such that the printed photographs showed a more colorful images than the ones for European customers - especially the Swiss, German and Austrian customers (I am Austrian) do not favor these colorful pictures too much as for their taste, all colors have to be a bit more subdued than you get them in the USA - just get a roll of film printed here and there and you can see the difference.

Image Quality
Now to some results. first here a comparison between the 5D and the D2X shot at the same location as I did some time ago the comparison of several digital cameras. It was again afternoon, a few clouds, but still very sunny. This is the entire shot at 24mm focal length and the crop which I have taken as usual is marked in the center with a gray frame:

EOS 5D, full frame at 24mm

Looking at the center crop, here one can see the details - it is the unaltered crop from a raw image, no additional sharpening and no additional color correction - the Canon shot is softer and gives the (wrong) impression that it is not as detailed as the one below taken with the D2X (default settings)

Canon 5D Detail

And for comparison here the exactly same shot with the Nikon D2X - same crop, no additional sharpening and no additional color correction - this shot appears to be more sharpened, crispier and showing more details with finer lines (look at the details in the front side of the white Jeep)

Nikon D2X Detail

Now the same two shots but this time with some additional individual sharpening, color optimization and a bit of contrast adjustment on the crops - now the Canon shot does not anymore show this softness but in my opinion is still a tiny bit less detailed than the Nikon shot.

Left shot - Canon 5D                                                                                                           Right shot - Nikon D2X

On the late afternoon the day before, I took pretty much the same photographs - this time the sun was covered with clouds - so less contrast and less colors - here the outcome, again the left shot from the Canon 5D and the right shot from the Nikon D2X

And a further enlargement of part of this photograph shows the small differences between these shots a bit clearer - again, minute differences only are visible - for my liking, the Nikon shot slightly more detailed, but this is barely noticeable and not as much as one would expect from the theoretical difference in the amount of pixels per square millimeter - with other words, Canon have done a fabulous job with the new 5D

As I was not able to get a Nikon D2X for the whole time (I would very much like to get the D2X for a few days for an in-depth review, but...), I was out shooting with the 5D various scenes which I think are representing different challenges for a camera - the first one is downtown Denver and you might remember this shot from the comparison of the new Fuji color negative films, which I recently made. Here the entire shot, taken with an RZ67 Pro II on the new Fuji Pro160C film:

As usual, I took a small crop out of this shot and compared it with the same shot, taken a week later from the same spot with the Canon 5D - here the crop out of the Fuji negative film shot:

Mamiya RZ67 Pro II and Fuji Pro160C

And here the same crop taken from a shot with the EOS 5D - same field of view, same crop - and the quality is superb - of course, there are more details in the medium format shot, but see how close the 5D comes to the film shot! As you might notice, the sun was slightly different, but this does not influence the comparison much I think.

Canon EOS 5D

One of the shortcomings of film based photography can be the grain of the film - sometimes it is an interesting and desirable component, but often it is disturbing more than helping. In my opinion, grain is one of the negative aspects of film photography - let me show you an example which shows how good digital photographs are in scenes where the film grain is really a nuisance. This shot, again one, I have taken before on the new Fuji film, was enlarged and shows a detail of another one from the previously mentioned comparison (Fuji Pro160 film).

Mamiya RZ67 Pro II with Fuji Pro160C

The same shot, taken with the 5D, reveals an important advantage of digital photography - the absence of grain in the image - here the same crop in digital:

Canon 5D

It is necessary to apply a "grain killer" program to make the film shot as smooth as digital, or even still better - here the same film shot as before, but treated with the program "NeatImage". As with all such manipulations, there is the danger to make the final image appear too artificial....

RZ 67 Pro II, Fuji Pro160C and NeatImage

Oct 22/2005:
I was asked it the grainy structure could also come from the texture of the outside of the building - I went back and took a photograph with a tele lens to show that yes, part of it is the structure of the outside which is granite - the digital camera is not able to resolve this granite structure at all and at the film shot, the structure is buried within the grain of the film itself. Here the detail shot with a tele lens (taken with my own DCS 460)

I also wanted to find out how good the resolution of the 5D is in terms of lines per millimeter. The old fashioned Patterson resolution table and chart helps and provides very good and objective results. The distance between "film plane" and lens needs to be 40 times the focal length - so with a 50mm lens about 200cm. The entire chart consists of four quadrants and here is the result, taken with the 5D, 50mm and 200cm distance.

Patterson resolution chart

To make an evaluation easier, I have taken the B&W part and scanned it into a scanner to show the fine details - that is how an ideally perfect camera would reproduce the chart:

And now the camera shots - I have put the B&W outcome of the Canon 5D in the upper half of the image and the original chart in the lower half. In this test, the 5D resolves about 36lpm. which is impressive. For the purists - this shot is taken with the chart in the center of the image field of view to eliminate as much as possible the influence of lens aberrations.

Oct 22/2005: Several readers of this review mentioned in emails that it would be good to see if the camera lens I used, had an influence on these results - to show that the results are basically identical with other Canon lenses as well, I have done the same test again with the Canon 17-40mmL and the Canon 100-400mmL lens - here the results:

Canon 17-40mmL

Canon 100-400mmL

Another interesting task was to see how the camera reacts to different settings of sensitivity - from ISO 100 to 1600 - how does the noise increase, how is the overall impression of the image at various ISO settings. I did the first test at night - here the results:

First, the entire test shot - taken at ISO 100 at an exposure time of 5 seconds

Now the two series of crops the first on taken from the left roof area at speed settings of ISO 100-200-400 (first row) and 800 and 1600 (second row)

The second set of shots taken from the middle part - bright areas with some very bright highlights - again in the first row the ISO settings 100, 200 and 400 and in the second row the ISO settings 800 and 1600

At ISO 1600, the noise in the dark areas becomes quite visible, but this is not unexpected and all other crops show how good the camera digests highlights and shadows and how little noise is visible - again, very impressive.

Another option which is very interesting, is the Picture Style Setting - when the camera is switched on, the default setting is "normal", but there are other settings with pre-set values which should give different results - according the names of these settings: Portrait, Landscape, Neutral and Faithful (plus a B&W setting). I have again taken one and the same shot - one with high contrast - full sun on a bright building and a lot of dark details in the shadows and used the before mentioned various settings for shooting. Here the results:

Standard mode

Portrait mode

Landscape

Neutral

Faithful

A bit difficult to decide which one is the best as the best exposure was done in the Portrait and Landscape mode, the best color rendition in Neutral and Faithful. The standard mode, as I expected, resulted in overexposure of the brightest parts of the image. This is nothing bad, it is clear as this scene was very challenging - the exposure difference between the darkest parts and the brightest ones was about 8 f-stops...just too much for digital and to be fair, also too much for film. One can take the best shot (exposure) - Portrait mode - and set the colors to where they should be and the result is again pretty impressive - here it is:

It is of course again a matter of taste - some people prefer the exposure set more for the highlights, resulting in rather too dark shadows, others like the highlights to be on the very bright side but therefore preserving more details in the shadows - as the Canon 5D's Style Modes can be manually set according to preference, no problem at all - you can adjust the output according to the scene's requirement and your preferences.

IMATEST Results

The program IMATEST, which has in the past year become a de-facto standard for quantitative digital camera performance evaluation reveals among other valuable information the color error - the amount of deviation of a color shown in a file from the original color. There is no digital camera existing, which has zero color error, but the larger the error is, the more correction is necessary and the smaller, the better the color rendition of the original image file. I took the 5D and shot the Gretag Macbeth color chart at a time of the day, where the color temperature was well within the range of the automatic setting capabilities of the 5D - around 10am. The results confirm what I have seen on the image files - the color rendition in standard automatic mode is already very good, one can class it as excellent.
Here the resulting data charts, diagrams and information - I am sure you are already familiar with this kind of presentation so I do not need to explain it in detail - besides, most of the data presented are more or less self-explaining.

Color Error

The little squares indicate the location of the color on the Color Checker and the little round marks indicate the location of the color as reproduced by the camera, in this case, the 5D. The smaller the error, the closer is the color in the image to the color of the Standard. The numbers are identifying the colors, not the amount of deviation from ideal!

White Balance Error

An ideal camera response would reproduce not only all color on the exact locations but also all white-gray-black shades without any white balance error - some cameras produce gray tones which are tending towards brownish tones, others, like the 5D, obviously gray tones which tend towards blue. But again, keep in mind, these errors are minute and the display shows the errors grossly exaggerated to make it possible to see the error tendency better.
In this context it was interesting to check if that EXPO disc, which is a filter sized white plate with a lens array in front of it and which can be used to get to a perfect white balance with manual setting, if this device is able to make the results even better - well, I tried it and yes, there is a slight improvement visible, not much, but the White Balance Error graph shows it - the numerical values for the deviation from real white are smaller then without it.

White Balance error - compensated with EXPO disk

I think that this little device, called EXPO Disc is a nice one, not suitable for all adjustments, but if used with care, very helpful - just to round this up, here a comparison of two shots, taken in the Botanic Garden in Denver - the left one with the 5D on AWB setting and the right one after manual adjustment of the white balance with that disk - the greenish tint is gone....

Finally now some more shots taken with the EOS 5D on the same day, the one, I was able to have that camera for myself. If I would not have a lot of Nikon lenses from my 35mm system (F100, F90s), I would definitely go shopping and get this new EOS 5D. It is very well designed, works very well and has also a very attractive price and the quality of the images is extremely good. It is a camera which I can highly recommend to anyone.

How will Nikon now react to this new addition to the complete line of digital SLR's from Canon? Nikon has got the D50 and D70(s), the much older D100, which I do not rate as up to date anymore and the D2H and D2X - what is missing is quite a bit to be really competitive, I think.
The D2X, the only DLSR which I class as a competitor, is not a full frame camera - From a marketing standpoint, a full frame camera has got a lot of advantages - especially when it comes to the viewfinder, the wide-angle lenses and applications like scenic photography, landscapes and also others like portrait photography.
For portrait photography, a full frame sensor allows to separate the area of focus better from the background and surrounding as the depth of focus is smaller at a comparable field of view.
For landscape and scenic photography, the advantages of full frame are obvious - more field of view, especially since the state-of-the-art digital cameras with their high resolution sensors are getting now to the point where this type of photography is not anymore excluded.
And when you have compared the viewfinders of a full frame DSLR camera with an APS DSLR camera, then I think that the preferences are clear - full frame viewfinders are larger, brighter and clearer, simply better. The D2X has got a very good viewfinder, but I think the 5D has got a better one.

Nikon needs to act and come out with a very attractive, full frame DLSR with pixel size preferably around 7-8micron (the current 5.5 micron would be OK if the algorithms for image acquisition and noise reduction were superb) at a price which is significantly below the Canon 1DsMkII and definitely not much, if at all, higher than the current price tag of the D2X. If not, I guess, Nikon is not able to compete with Canon anymore, what a shame this would be.

This review and article is solely based on my own experiences, findings and opinion after being able to use the Canon EOS 5D for 36hours - certainly incomplete, not touching all necessary topics but maybe suitable for potential buyers to make a decision....

BACK TO HOMEPAGE

Send mail to Webmaster with questions or comments about this web site.
Last modified: 19-May-2007