5D Mk II vs. 5D jpgs - I'm seeing about a 1-stop noise improvement

<p>(Not on my own photos; my 5D II is supposed to arrive on Monday.)</p>

<p>I'm looking at the samples at imaging-resource.com, and it looks to me like the 5DII files are about as clean at ISO 3200 as the 5D files are at 1600.</p>

<p>Didn't do RAW conversions, because I don't care <em>that </em> much for now; I just wanted a quick idea. In a few days I'll have both cameras, so the outcome of this comparison won't affect my buying decision. Nor do I care how the 5DII compares to Nikon, because I'm not at a stage where I'm trying to choose between camera systems (and it sounds like the differences are pretty miniscule anyway, especially in prints up to a given size; one can go with Canon or Nikon for any number of reasons, but a clear edge in noise control isn't one of them anymore).</p>

<p>I simply went to the i-r.com comparometer (http://www.imaging-resource.com/IMCOMP/COMPS01.HTM) and downloaded relevant full-size jpgs (I prefer the still-life, with crayons and wine bottles, instead of the test charts) for the 5D and 5DII at various high ISOs. One can't really use the "Comparometer" feature to do side-by-side comparisons because the photos are sized so differently thanks to the pixel count, so I upsized the downloaded 5D images to 5616x3744 so they'd match the 5DII images (it's fairer to upsize the smaller files than downsize the larger ones when doing such comparisons). Then I went to Photo Mechanic, the program I personally use for editing, which lets me scroll back and forth through any section of the respective images at 100% (I'm guessing that similar programs like Lightroom allow the same?). I like darkish but not black areas when judging noise, such as the shadows on the back grey wall or the necks of some of the bottles (comparing in this case is an imperfect process because the lighting, alas, isn't identical in the photos; they were taken a couple of years apart).</p>

<p>Again, it looks to me like the ISO 3200 files from the 5DII are pretty close noise-wise to the ISO 1600 files from the 5D; some areas look a little less noisy, some a little more, but the 5DII's 3200 files are definitely not as clean as the 800 files from the 5D (no surprise there), nor is there any question when viewing both cameras' output at 3200 which one's from the 5D2; it has a clear edge (no surprise there, considering the hype).</p>

<p>The one thing that's really hard to tell from those samples is how much the noise control is adversely affecting preservation of detail; for that I'll just run my own tests once I have both cameras. (We all know that different companies take different approaches to the balance involved in suppressing noise vs. preserving detail and there is no "right" or "wrong" approach, so it's not worth arguing about.)</p>

<p>I know some people were hoping for more stunning results, but to me improving noise control by a full stop while increasing the pixel count from 12 million to 21 million IS very impressive. Brickbats welcome, but let's keep it civil.</p>

<p>There is a <a href="http://www.dslr-forum.de/showthread.php?t=414272">similar comparison</a> in a German forum.<br>

Those guys compared the 5D, 5D II and 1D III. They downsized the 5D and 5D II files to the 10 MP of the 1D III.<br>

The last two section are the original files of the 5D and 5D II not downsized.<br>

The 5D II delivers pretty impressive stuff, but to my eyes the good old 5D is not too bad in comparison.</p>

<p>I maintain that you can't really conduct an accurate comparison of camera with differing numbers of photosites (even if they have same sensor dimension) by downsizing/upsizing one or more of the samples to match the size of the other test cases. <br>

The best real tests would be:<br>

 

<ul>

<li>to make RAW photographs with each camera using, if possible, the same lens and the same settings, then take both photographs all the way to prints of the same (large) size.</li>

<li>to go through a typical workflow to create optimized jpg images for online distribution in typical posting sizes. (Obviously this would be a much less critical and less informative test.)</li>

</ul>

<div>There are so many variables otherwise that the results become quite confusing and, arguably, quite meaningless. Obviously, up/downsizing has at least some negative effect on images. On the other hand if you don't up/downsize and you instead compare 100% magnification crops you are looking "more closely" at the sample from the camera with higher photosite density. (For example, a 100x100 pixel 100% crop from a 12 MP camera is a larger portion of the image than a 100 x 100 pixel 100% crop from a 21 MP camera.) Regarding noise, the size number of "points" in a print/photo of a given size has some effect on the impact that a given level of noise has on the appearance of a photo, too.</div>

<div><br /></div>

<div>Dan</div>

</p>

<p>Up rezzing 12.8MP 5D files to match 21MP 5D II files doesn't sound very fair either. Certainly any noise artifacts in the up rezzed file will be nearly twice as apparent.<br>

Even if the 5D II files have exactly the same pixel level noise as the 5D, noise would be less apparent in a same sized print as far less up rezzing will be needed on a huge 21MP file than a 12.8MP file. When comparing 12x18 prints from my 40D and 50D a couple months back, the 50D prints were smoother with less noise, although at pixel view they were about the same noise wise.</p>

<p>Who cares about photosites.  All that matter is what's visible at a given print size.  The rest is irrelevant, except for tech talk.</p>

<p>I'm thoroughly impressed with the 5D II, far more so than I was the 50D. I wouldn't call it a full stop better than the original 5D, but the fact that it's better at all on a per-pixel level is to be applauded. There's essentially no difference between ISO 100 and 800. ISO 1600 introduces some oh-so-subtle softening, which becomes slightly more noticeable at ISO 3200. For small prints with careful noise reduction from RAW, I wouldn't hesitate to wind it all the way to ISO 12800. I'd say the same for Nikon's D3; for the 5D II to nearly match a body with only 60% of its resolution is quite a feat. </p>

<p>Hey Ralph, there are several sites now that have done production models tests between the two and they

all say the same thing. The 5D markII images are as good and better on higher ISO's and the overall IQ is impress

ive to say the least. None of them have said there was soft images because of the noise reduction, plus you

can now choose how much noise reduction you want to use, another nice feature. When you cons

ider going from 12-21MP the IQ they have acheived is impressive. I know for me it was a no brainer, have mine orde

red and can't wait until I get it. If you check out Photography Bay's website they have links to several of these review

s. Problem with 5D is was it was basically a 30D with FF sensor and 12MP, so feature wise the new markII is leap

s and bounds ahead of the 5D, focusing is better, live view with new high resolution LCD, mico lens adjusments and s

o on. If you are a landscape photogher like me and use a tripod for most of you shooting then start using the live vi

ew function, there are so many benefits with it I wouldn't consider a camera that didn't have it. I can get much mor

e precise focusing now, much easier to see what my compostion really is when checking the live view and turn on the

grid option and it makes composing images a lot easier, no more having to buy the grid screens, Whooo! If the wi

nd is blowing good and your using longer lens just turn on the live view, set it to 10x magnifaction and you can s

ee if your lens is shaking at all, saved me couple times now from getting soft images. I use mine like

a mini view camera, what an advantage. I'm surprised more people don't talk about this feature when shooting t

ripod mounted shots, such an advantage. Now with new high resolution LCD if wi

<p>"Who cares about photosites.  All that matter is what's visible at a given print size.  The rest is irrelevant, except for tech talk."</p>

<p>I totaly agree, however I have been using the Mk II for a little over a week now and I agree with the OP that there is at least a 1 stop improvement in iso on the Mk II, this is a fine camera.</p>

<p>Ross</p>