|
Resolution Roundup: A Comparison of the Canon 1Ds Mark II, Nikon D2x and Nikon D200 April 2006 Disclaimer: This is not a scientific test. I wanted to see for myself just how big and what the differences, specifically resolution, are between these cameras. Both Canon and Nikon make incredible digital cameras and even though I use Nikon for my work I tried to give an unbiased eye to the final results of this comparison. Also, please note that I am well aware of the difficulties in comparing these cameras. There is a laundry list of differences that make it hard to get a true comparison. Those include: different sensor dimensions, different anti-aliasing filters, different optics, differing RAW file formats and on and on. In this test I have tried to keep the variables to a minimum by shooting with the same settings in all cameras, using similar lenses from each manufacturer, and upsizing the lower MP cameras to match file sizes. With each example I will explain why I have chosen to test them in a certain manner and exactly how the cameras and files were managed. I realize that this ‘test’ might be disconcerting to many of my readers and there are highly likely many other ways of comparing these fine cameras. Feel free to email me if you want though I can’t promise I’ll have time to respond to every email. Otherwise, sit back and enjoy the ride. I don’t think there is any earth shattering info in this article but it is very interesting just how good all of these cameras are. |
|
 |
|
As with any test, we have to start with some reasoning for the test. I own a Nikon D2x and D200 (with which I am very happy) and I wanted to find out just how much of a difference there is between these two cameras compared to Canon’s top of the line 1Ds Mark II (here after 1DsII) not only for myself but also for my clients. This is not a comprehensive evaluation or comparison of these three cameras, I have chosen to focus my efforts on comparing the resolution of the three sensors in real world usage. Of course there are many other comparisons that could be made but I feel those have already been well established by better testers than myself. For example, high ISO noise has been tested very well by the folks at dpreview.com. Even without their exhaustive tests it is well known the Canon 1DsII outshines Nikon’s offerings in the High ISO department as to be expected. The fact that the Nikon D2x does so well with a DX sized sensor is an incredible engineering feat that Nikon should be very proud of. And maybe you are wondering why I left out the Canon 5D. The answer to that question is that I didn’t have access to one. Likewise a lot could be said about how each camera handles, the ergonomics and layout and design. As a Nikon user, I will admit I prefer the way the Canon pro bodies feel when I pick them up - but I have really big hands - and from an aesthetic point of view I prefer the way the Canon shutter button feels. On the other hand, I think Nikon did a much better job with the layout and controls of the camera, especially the manner in which you change autofocus points. The 1DsII and the D2x are both beasts of burden compared to the D200 but that’s par for the course. I could go on and on about subtle differences amongst these cameras but lets get on to the test. To compare the resolution of each camera I photographed a still life of a climbing rope, my Epson P2000 and a few other items. I have found my climbing rope is a great test subject for resolution as their are thousands of fibers that make up the sheath of the rope and it is a perfect subject to show differences in resolution. Each image was shot with the mirror locked up at 1/13th sec. and f/8 at ISO 100. I took great care to make sure each camera was set up on a tripod and the angle of view was exactly the same for each camera. Of course with digital cameras it is hard to make exact comparisons. How much sharpening do you apply in the RAW processing to account for the different anti-aliasing filters? I applied the default sharpening of 25 to all images in the RAW processing and I used Adobe Camera RAW for all images. How do you compare images from cameras of different resolutions and with different pixel dimensions? The Nikon DX sensor is just slightly more rectangular than the Canon - only by a few pixels but that becomes significant when enlarging the images. Since I don’t have weeks to compare these cameras I decided I would do three different comparisons - one of which would surely show me any difference their might be. First, I compared the three images at 100% with the Nikon files upsized (via Bicubic Smoother in Photoshop’s Image Size dialog box) to match the vertical pixel dimensions of the Canon. Why did I choose the vertical dimension? I chose the vertical pixel dimension because this gave a file size almost exactly the same as the 1DsMII and it gave the closest dimensions to the Canon sensor without cropping the Nikon image. I did this comparison on screen and on glossy prints made on my Epson R1800. I chose glossy prints because they show the most detail. You can see the screen shots of this comparison on page 8 - the middle row of images. And frankly on screen I was hard pressed to see much if any difference in resolution between the images - especially since no sharpening was added at this point save for that done in the RAW processing. Now at this point I can hear everyone saying that the Canon has a much stronger anti-aliasing filter than the Nikons and it does - so that might negate this test. I would counter that when you upsize an image with Bicubic Smoother in Photoshop it adds a fair bit of smoothing to the image as well - hence this comparison was a draw and if the Canon resolved more detail I’d need a microscope to tell the difference. The fact that I couldn’t see a difference is why I chose to do two more rounds of testing. For the second comparison, I upsized all images (again via Bicubic Smoother) to 30x45 inches figuring this would definitely show a difference. I compared the images without any added sharpening and with sharpening added to each image. The images were compared in both print form and on screen, again on glossy photo paper. The on-screen version of this comparison without sharpening added can be seen in the bottom row of images in Figure 1 below. I must add at this point that this newsletter at screen resolution is a horrible medium for judging resolution. But the end result isn’t much different than what you see here - I could not see any noticeable difference in the on-screen comparison with or without sharpening. |
|
 |
|
Figure 1: Still lifes of my climbing rope, Epson P2000 and a Gretag Macbeth mini color chart shot on the same tripod with a Canon 1Ds Mark II, Nikon D2x and Nikon D200. Each camera had the mirror locked up at 1/13th sec. and f/8 at ISO 100. Canon image shot with a USM 28-70mm f/2.8 zoom at approximately 70mm while the Nikons used a Nikkor 28-70mm f/2.8 zoom set at 47mm, thus keeping the angle of view consistent between sensors. Top image is the full image, middle images are a crop of the rope at 100% with the Nikon files upsized to match the Canon file size. When I compared these middle images at 300 ppi all at the same resolution and pixel dimensions (or as close as was possible since the Nikon files have different pixel dimensions than the Canon) I could not see any difference in image quality - especially between the Nikon D2x and the Canon 1DMkII. Hence I enlarged the three images to 30x45 inches to see if the differences became more apparent. On screen it was still hard to see a difference and honestly this newsletter and screen res samples don’t show any difference that I can see. To see the real difference I had to print 30x40 samples. Now I don’t know about you but I was surprised I didn’t see a big difference at this point. Upsizing the Nikon files should have given them a huge disadvantage compared to Canon files in the first comparison and again in the second comparison since the Nikon files were upsized by 3.16x (D2x) and 3.47x (D200) compared to the Canon file being upsized only by 2.7x in the second comparison. In the print comparisons I had three different comparisons going: the first comparison, the second (both sharpened and unsharpened) and a third comparison in which I prepped each file individually with sharpening - to make them each look their best - and all the images were sized to the native file size of the Canon which in theory should give the Canon an advantage. In the second print comparison (sorry if this is getting confusing) of 30x45 inch prints on glossy paper I could only see a hair’s breath of difference between the images when looking at the prints with my loupe! And that difference was so small it is hard to say there was a difference at all. That hair’s breath difference was a little more noticeable comparing the D200 (10.2 MP) to the 1DsMII (16.7 MP) as I would expect but it was not anywhere near as big a difference as I would have expected. In my third print comparison it was the same story but even closer. Looking at the images with the loupe I was hard pressed to see any differences at all even between the 10 MP and 16 MP cameras! So what does this mean? I tested the images just about every way I could think of. I am sure some could pick my testing methods apart but the reality is a difference of two or even four megapixels is a very slight difference! About the only advantage I can still give to the Canon is that it produces a larger file size without having to upsize the image. But as my testing has shown even that isn’t a big advantage as I couldn’t see a difference in resolution even when I upsized the 10 MP and 12 MP images to match the 16 MP image of the Canon! To see an appreicable increase in resolution you would have to double the MP. Now I can hear all of the Canon aficionados going ballistic already - believe me, I was prepared for Nikon to get it’s butt kicked. I’m frankly astonished there isn’t a bigger difference myself between these cameras. I am not trying to say one camera is better than another with this review. I have no agenda. No one is paying me to do this comparison. I bought my Nikons at full price and a good friend of mine, Marc Romanelli, graciously lent me his Canon 1Ds Mark II to shoot with for this test. Under different conditions there might be a much different result but I would bet that isn’t due to the sensors being different sizes. In another test (see Figure 2 below), I shot some portraits of Marc Romanelli hand held with both the D2x and the 1DsMII as he held up a Mini Gretag MacBeth color chart. I used the same lenses as before and both images were shot at f/8 and a shutter speed of 1/500th a second. Now in this test - where again the Nikon D2x image was upsized to match the Canon file I can see a difference in sharpness but I am not sure it is due to the resolution of the sensors. Why can I see a difference in the sharpness here but not in the other images? I’d be willing to bet it is because I was hand holding the cameras - especially the D2x. In my experience with the D2x I have to be very conscious of the shutter speed if I want critically sharp images. Any lapse in my concentration or camera technique and the image quality suffers - the D2x is merciless when it comes to camera technique, much more so than the 1DsMII or any other digital camera on the market. Why? It all comes down to how big the pixels are on the sensor and the D2x’s pixels are packed in so tightly, to the tune of 5.8 microns per pixel in diameter, that in lines per millimeter it’s sensor actually out resolves anything else on the market and hence it is much more sensitive to camera shake. I’ve found that I have to use a shutter speed of 1/800th a second to get sharp images from my AF-S 80-200mm f/2.8 Nikkor zoom lens - that is if I want critically sharp images. Just as with film, there is a range of what is considered “sharp”. I can shoot at slower shutter speeds and still get usable images but if I want razor sharp images where I can count the pores on someone’s face then I need to pay great attention to my shutter speed. So in this comparison with the images shot handheld the Canon wins not because it sensor is out resolving the Nikon’s but because of the Nikon’s sensitivity to camera vibration even at 1/500th a second shutter speed. This being a real world test everything counts and the Canon has a noticeable advantage. If I put both cameras on a tripod then that advantage goes away. |
|
 |
|
Figure 2: Another comparison between the Canon 1Ds Mark II and the Nikon D2x - this time a handheld portrait of my good friend Marc Romanelli. I used the same lenses as in the previous test and just as before the D2x image was upsized to match the Canon file. Both images were shot at 1/500th sec. and f/8. It is nearly impossible to see any difference in these images at right - especially in this newsletter - but the Canon was a hint sharper and had slightly more detail than the Nikon. Knowing my D2x and how sensitive it is to camera shake this test does not isolate the sensors resolution as the previous test did - but still in the real world where cameras are handheld this shows the Canon can resolve slightly more detail which is as expected from a 16 MP vs. a 12 MP camera. So what does all this mean? Well for one, the megapixel war, at least for 35mm cameras as I predicted a year ago is over. The Nikon D2x and the Canon 1Ds Mark II already out resolve many of the lenses both companies manufacture. Hence, an increase in megapixels will only push the current crop of lenses that much harder. At this point I’d say that more megapixels aren’t going to necessarily increase image quality until we have better lenses to take advantage of more megapixels. I have heard rumors that Canon is introducing a 22 MP 1DsMark III this fall and while I am sure it will be a great camera the actual image quality will not be much different than their current 1DsMII. Of course they may have added some other features that make it a must have like better color, real 16 bit images, better autofocus, etc. Nikon on the other hand could if they so desired introduce a 28 MP full-frame DSLR based on the D2x with the same chip technology just expanded to full-frame. If you do the math that is what a 5.8 micron pixel size comes out as when you create a full-frame 35mm chip. I don’t think we’d ever see that camera because if their 12 MP camera is so demanding on Nikkor lenses what would be the point of a 28 MP camera unless they build it with a medium format lens mount. Not likely. All fantasies aside, the final verdict on this comparison is that all of these cameras are fantastic and I’d say all of these and most likely the Canon 5D as well are giving medium format resolution in a 35mm body. And that is great news for every photographer. Better image quality in a smaller package. If you think my comparison has to be way off base then I would encourage you to do your own testing. I’m not the only one to say these cameras have remarkably similar image quality. There are quite a few very well done tests and comparisons online. Check out www.dpreview.com’s review of the Canon 1Ds Mark II where they also compare it to the Nikon D2x and find that you’d need to blow images up HUGE to see a real difference. I am guessing by HUGE they meant bigger than 30x45 inches as I saw very little difference at that size. Bjorn Roslett also published a very thorough and detailed comparison of the Nikon D2x and the Canon 1DsII on his website at http://www.naturfotograf.com. And finally for you Nikon photographers out there wondering if the D200 is good enough compared to it’s much more expensive bigger brother the Nikon D2x you can take image quality out of that buying quandary. The bigger question is do you need superfast autofocus, a bullet proof camera body, and any of the other features that the D2x has above and beyond the D200. The D200 does have a black and white mode and has great autofocusing as well, though not as responsive as the D2x. And the D200 is quite a bit lighter and a LOT cheaper! For the money, I don’t think anyone would argue with me that it is the best deal going right now for a high end digital SLR. And it is built fairly tough itself and should be able to deal with many years of abuse. As an aside, while on assignment for Adobe the Lightroom Product Manager, George Jardine, who was shooting with a Canon 5D remarked that he couldn’t tell if an image was shot with a Canon or Nikon in terms of color. I know from my experience with Nikon that I have to custom white balance every 5 minutes or so near sunrise or sunset because the color temperature is changing so fast. This has become a major headache with digital and I hope Nikon and Canon improve how they handle color during the golden hour. Just as with film light can make or break a photo. With digital we have to get used to a whole new color palette and some of us old film dogs are taking a little longer than others to get use to the change. I miss the colors Fuji Velvia produced but the quality from digital is so much better on several other fronts that there is no going back - save for shooting medium format film which I still enjoy quite a bit. So whether you are shooting Nikon, Canon or a Holga, get out there and make some nice images and forget about megapixels and all the marketing voodoo!
|
|
all images copyright © michael clark |
contact info: (505) 438-0828 | mjcphoto@comcast.net | www.michaelclarkphoto.com |