Should 5D Mark III images be sharpened differently from 5DC images?

<p>I just got a 5D Mark III. My images from it are noticeably softer than with the 5D classic, using the same lens, exposure and aperture. I first noticed it while comparing images I had taken at two different banquets, then verified with some controlled testing. I took the camera back to the camera store I bought it from, and they thought the problem might be (1) that I had been using MRaw for my tests (I was using it because that gives about the same pixel dimensions as native raw on the 5D Classic), or (2) that I wasn't sharpening the Mark III images properly. That seemed wrong; my research indicated that, if anything, MRaw should be sharper than full raw images, but in any event, I came home and tried shooting in full raw, but got the same results as before. I did notice that I could get fine detail in the Mark III images to look pretty good if I applied a lot of sharpening (i.e., moving the slider in Light Room over to around 100, as opposed to 25 for the 5D images), but that brought out a lot of noise in the shadows that simply wasn't present in the 5D Classic images. I suspect the problem is really with the camera, but just to be sure, I thought I'd ask here.<br>

Thanks for any insight.</p>

<p>This is a common complaint that has been raised many times on this forum: "I got a new higher resolution camera and the pictures are less sharp than my old camera." The short answer is: The 5D is a great camera that produced "sharp" images when viewed or printed up to a certain size with the proper amount of sharpening for that particular output size. The 5Dii (or 5Diii) will do the same at that same size (there is no reason it should be "sharper" for the same output size), and at even larger output size, again if sharpened for best results for that particular size. The sharpening parameters will be different, just sharpen until best results are obtained for each case (specific to output resolution). Caveat: Lens and other limitations (atmospheric, camera shake etc) will be more pronounced with higher sensor resolution if viewed at 100%. Rule of diminishing returns. Noise may be higher with a larger resolution sensor (smaller pixel pitch thus less light per pixel and lower signal to noise ratio).</p>

<p>"<em>Lens and other limitations (atmospheric, camera shake etc) will be more pronounced with higher sensor resolution if viewed at 100%. Rule of diminishing returns. Noise may be higher with a larger resolution sensor (smaller pixel pitch thus less light per pixel and lower signal to noise ratio</em>"<br>

Funny that did not happen to me when I upgraded from the 30D(8.2MP) to the 7D(18.0MP).</p>

 

Barry,

 

Did your controlled testing include tripod based live view 10x manual focus based images? If not and those images turn out to be

sharp then you may have a very slight front/back focus issue (which you can calibrate with the 5d3 itself).

<blockquote>

<p>This is a common complaint that has been raised many times on this forum: "I got a new higher resolution camera and the pictures are less sharp than my old camera."</p>

</blockquote>

<p>Yep, we've all dug that hole and fell in it after spending big bucks on an ultra MP dream cam. Happened to me when I went from a 5D to 5D2 and 40D to 7D (you won't notice much diff going from 5D2 to 5D3). The problem is we all start by using the same PP habits and pixel peep not realizing higher rez means much higher magnification at pixel level. Once I figured that stuff out I couldn't help but notice my larger prints--13x19 and above--were actually more detailed and defined from the higher MP cameras. Not much difference in smaller prints.</p>

<p>In my experience, MRAW is not sharper than full rez but has merely been down sampled. It may appear sharper on a huge monitor only because it is smaller and you can see more of the image. If you down sample the full sized file yourself it is the same story.</p>

<p>Of course, you may not have dug the same hole but, instead, are suffering from lens calibration issues or even a defective camera. </p>

<p>Well, but I am experiencing the softness at the <em>same</em> pixel dimensions as on the 5D. I'm well aware that there's supposed to be a difference on a "per pixel" basis, and that 100% at full resolution on the Mark III is supposed to look softer than 100% at full resolution on the Classic, but what's happening here is that when I either use Mraw, or when I shoot at full resolution then downsize to 5D size, I get noticeably softer images. I've never heard of that as a problem.<br>

Also, as I said in the original post, when I play with sharpening, I can improve the apparent sharpness, but at the expense of bringing out a lot of noise, so that doesn't seem to be an ideal solution.</p>

<p>I'd suggest you do full trial (MF, on tripod, live view 10x focusing) to see if you can get satisfactorily sharp images. If the camera is capable of making the shots sharp enough w/ the 'manual setup', then you have a concrete problem that you can address.</p>

<p>Unfortunately, the vagaries due to the difference in the camera tend to make something like this difficult to assess without a full trial.</p>

<p>Perhaps you could post some images that demonstrate the difficulty you are having.</p>

<p>I'm curious to know what ISO you are testing them at. I went from a 1Ds to a 5DII, so pretty much the same jump, and saw the resolution difference immediately!</p>

<p>In my experiece, with 5 upgrades in resolution, I have generally found that less sharpening and less saturation were required as I went up the ladder.</p>

Without more information (lens, shutter speed, aperture, AF or MF, IS on or off, flash used or not), I would guess that auto focus is to

blame. Either you're not using the right AF options for this type of shooting, or the camera needs to be fine tuned for this lens.

 

Do the tests that have been suggested above. I would add that you might try testing in bright sunshine for maximum contrast and a high

shutter speed. Test at f/5.6. This should be a sweet spot between sharpness and diffraction for most lenses. Turn IS off when using a

tripod, of course. Good luck.

<p>I have a 5D, 7D and 5DIII. My initial thoughts on the 5DIII were similar - the images just don't seem as sharp as the 5D. After goofing with the sharpness on the 5DIII and shooting with it for awhile the problem somehow disappeared. :) I use Digital Photo Professional a lot. It comes with your camera. Download the latest version from Canon. Shoot in raw and do some tripod tests as described above and process a few images in DPP to see what happens.</p><div></div>

<p>The 5dIII has a lot more pixels than the 5d. Therefore, when sharpening you'll need to raise the value of the "radius" in unsharp mask to get the same effect. Pretty simple.</p>

<p>To answer some of the questions above, I've been doing my tests with both cameras at ISO 200, shutter speed of 1/200, and f/6.3. I haven't been using a tripod or live-view focusing, but instead have been shooting a series of images at each setting and picking the sharpest of any group (and, honestly, there has been almost no variation; the images have been very consistent). I've minimized camera movement by bracing my elbows on a table, and I've had my test subject (my daughter) do the same on the other side of the table. Once I've got the best image out of each group, I then compare those best images with each other.<br>

I like the idea of doing a tripod + live-view focusing test just to see if the lens on that camera is capable of sharp focus, or if instead there is an autofocus issue. I note, however, that I've done some informal testing shooting at a detailed subject tilted away from the camera and have found no indication of front or back focusing; the sharpest point of focus is always right where I expected.</p><div></div>

<p>And here's an example of the softer results. Each image was processed identically in LR 4.1 -- which is to say, all that's been done here is cropping. Otherwise, the images are just as they came out of the respective cameras. I've used the default sharpening in Lightroom, which is Amount = 25, Radius = 1.0, Detail = 25, and Masking = 0. The softer image was taken using Mraw -- again, with the idea of having a file with the same pixel dimensions as the 5D Classic produces. It hasn't been resized in post.</p><div></div>

<p>It is not clear to me why you are testing the 5Diii camera using the MRAW format rather than the full RAW format. I would take the picture in RAW, process in LR incl. sharpening, then downsample in PS and sharpen again. The 5Diii has a smaller pixel pitch, so if you compare the same number of pixels, I would not be surprised if the 5D outperforms the 5Diii.</p>

<p>I think Peter E. and some others may have hit on the solution. I just did some studio testing with MRAW and FRAW, and found that the MRAW images were uniformly softer than both the FRAW and the 5D Classic images. The FRAW images might be ever-so-slightly softer than the 5DC images, but my understanding is that that is to be expected due to the finer pixel pitch, and is something that can be dealt with by adjusting the sharpening. In any event, the difference is so tiny that I doubt it could be seen as a practical matter in any print.<br>

FWIW, I found in my particular camera that the autofocus was bang on; there was no difference in result between autofocus and manual live-view focus.<br>

Thanks to all for the suggestions.</p>

<blockquote>

<p>My images from it are noticeably softer than with the 5D classic, using the same lens, exposure and aperture.</p>

</blockquote>

<p>They are probably at least as sharp and perhaps (once you are finished) even sharper. Keep in mind that if you compare 100% magnification crops from the two cameras, you are looking at a significantly smaller area of the photograph, which is equivalent to viewing it under greater magnification. (I know that might seem like an odd thing to say, but 100% crop does not mean "the same size" - it means the same number of pixels in a a given area.)<br /> <br /> As to the separate question of whether the sharpening settings should be different, if you are picky about this, standard sharpening settings are merely starting points and often must be modified for different subjects and different lenses, even when shot with the came camera. For example, if you are applying a large radius USM process (often done to increase what might be termed "local area contrast") a given radius is a smaller percentage of the frame width on a higher MP camera, so you might find that you need a larger radius to get roughly the effect you go with your old lower MP camera. You may also find that if you do "output sharpening" (to compensate for dot gain, or ink spread during inkjet printing) that you need to "over-sharpen" a bit more than you would on the 5D, but that the higher MP camera can also tolerate a bit more such sharpening. <br /> <br /> A fun thing about sharpening is that once you get used to how it works and play with it a bit, you start to see ways in which you can use it in other than the default, boilerplate ways.<br /> <br /> However, do watch out for some common misconceptions about the resolution performance of high MP cameras. You'll hear things like "higher MP cameras are more prone to diffraction blur" or "higher MP cameras are more prone to motion blur" and so on. All nonsense. If you make two photographs with the same lens at the same aperture, one with a 12MP camera and the other with a 21MP camera, and then print them at the same size, they will exhibit exactly the same amount of diffraction or motion blur. What is true is that if you <em>try to print even larger as a consequence of using a higher MP camera</em> you need to be more careful about such things. However, there still is no more blur in the images from the higher MP camera.<br /><br />I, too, wonder why you got a camera with the potential of increasing resolution at 21MP and then shoot in a mode other than plain old normal RAW.<br /> <br /> Dan</p>

<p>G. Dan: Thanks for your comments. I believe they'll be helpful as I play around with sharpening. You ask, ultimately, why I would want to use MRAW. The answer is that ordinarily I would not. I happened to be using MRAW because I was taking pictures for a function where the images themselves were not important to me personally, and I knew from experience that no one was going to be ordering more than about a 4x6 print (and more likely, would just want something at screen resolution), so I figured that there was no point in using up all that storage space. The only reason I discovered the softness issue was that the camera seemed to be underexposing with my flash, and while I was testing that out, I noticed that I was getting noticeably sharper images with my 5D. That then got me curious about the softness issue, and in fact one of my first hypotheses was that the problem lay in the MRAW process. However, some Web research suggested to me that, if anything, MRAW ought to appear sharper even than full RAW, which, if you'll look at my comment just before yours, seems not to be the case. (So much for the Web.) WhenI took the camera in to the shop, they agreed that the images were definitely softer, but were baffled at the cause. It was only after a lot of further testing, as detailed above, that I think I've gotten to root of the problem.</p>

<p>And, actually, there is a further reason I would want to use MRAW, which is that I didn't really buy the camera for its pixel count. Other features, such as improved focusing, highlight handling, exposure, and frame rate actually were more interesting to me. I ordinarily print only at 8.5 x 11, and even my old 20D was overkill for that. I was more than happy with the resolution of the 5D.</p>

<p>If you ordinarily only print at 8.5 x 11, you really don't have to worry about camera sharpness issues at all. With no interpolation, you'll have resolutions in the middle to upper 400 range from that camera. Since that is your target output, instead of looking as the screen, just whip up a couple prints in that size and see if you have any sharpness issues.</p>

<p>The fact that mraw seems sharper is also a manifest of the first issue I described, namely the "larger MP at 100% will always look softer even when it isn't" phenomenon.</p>

<p>Dan</p>

<p>BTW, and to the forum not you, the "Sexy Asian Women Love American Men" ads appearing in the sidebar seem quite wrong for this site. I'm not a prude, but there are a number of issues there, and ya'll might want to think about putting some boundaries around the ads that appear here.</p>

<p>Yes, this is what I have also noticed in comparing between 5D and 7D even at the same shutter speed and lens.</p>

<blockquote>

<p>BTW, and to the forum not you, the "Sexy Asian Women Love American Men" ads appearing in the sidebar seem quite wrong for this site. I'm not a prude, but there are a number of issues there, and ya'll might want to think about putting some boundaries around the ads that appear here.</p>

</blockquote>

<p>Haha, my ads are all for photo gear! However, I think ad materials are selected by reading your cookies and web browsing habits. Once my wife used my computer just before me and all sorts of women's product ads appeared...</p>

<blockquote>

<p>I think ad materials are selected by reading your cookies and web browsing habits.</p>

</blockquote>

<p>I don't browse anything that would label me as interested in that subject, and these ad don't show up on the many other sites I visit that where I know that ads are selected based on browsing history. </p>