I shoot strictly RAW with my D100 and while I get great pictures of buildings or defined shapes like people, cars, etc. (things with definite contrast) I've encountered a weird issue when shooting nature images (and that's most of my work). For example, I went to New England last fall and shot over 1500 film and digital images under various lighting conditions (sunny, light cloud coverage, heavy cloud coverage, etc.) Invariably on the digital most images with leaves that were turning yellow or of a very light colors (in other word most of them) were "muddy" for lack of better word. It's hard to discribe but it looked almost as if they were from a painting. It's nothing major and I used many of these pictures in my show but it bugged me. Since I was relativelly new with this camera I though that it was me, except that I continue to have this problem with various settings, curves, etc.
I'm not sure where it comes from since it appears in the exact same way regardless of the lens I use. I'm not even sure if it is not me being paranoid or screwing up something since I am still relativelly new to digital.
At any rate I attached a couple of pictures for your curiosity and comments on this problem. They both represent a 100% crop of a small area of pictures (center of image to avoid any lens distortion) with various degrees of this "muddy" effect.
I'd appreciate any comments, ideas, recommendations.
What I mainly noticed in the second one is the white sky behind the leaves. It looks like the contrast when you were shooting was rather high, something that is always a problem when shooting outdoors. Slide film has about 5 stops of contrast with detail still in the shadows and hightlights. The D100 is a bit more forgiving with between 6 and 7 stops, but it is still easy to exceed this. This may be the source of the perceived over-exposure. On such a small sample though, it may not really be a problem.
As you were shooting hand-held, what was your focal length and shutter speed? The first two images also look a bit soft, but this may also be only evident at the size you have enlarged them.
Note on this one the fuzziness of the yellow leaves. The green undergrowth is clean in comparison and this makes no sense to me.
Note that, as I mentioned, this is not a grave problem and hardly anyone notices it (hence my feeling that I am becoming paranoid) and since I know that there is more leaway with the d100 (I always double up with my f100 and kodachrome 64 or more recently the E100G and GX series) it does not make sense.
My recommendation is to choose the following settings for the most flexible digital captures:
White Balance: Fixed (e.g., CDY-3, etc.) Contrast: Less Contrast Sharpening: Normal (or Low or None) Mode: RAW
Especially in a high-contrast situation, I would not want the Nikon auto tone curve to set my dynamic range. In the less contrast mode, the original image will be a little flat, but leave lots of room for curve manipulation in Photoshop, and will allow you to keep more of those highlights from blowing out. In RAW mode, you can try some minus exposure compensation to hold in the highlights a little better and hope to use Photoshop to pull up the shadows a bit, and/or use contrast masking to compress the dynamic range.
Based on what I can see, I agree with Bob that it may be an overexposure problem with the highlights. (Either that or the leaves are blowing in the breeze and the underbrush is relatively still. Photographing fall landscapes at small apertures with beautiful tree foliage can be maddening because of wind whipping the leaves.) If the image was still enough it would be an ideal candidate for dual exposures and recombination in Photoshop.
Basically, you can take two shots of the same scene, one exposed for the highlights and one for the shadows, then combine them in photoshop. If you are shooting NEF, you can actually make two images from the same RAW capture.
White balance helps to get close in camera if for no other reason that it's sure a pain to have to change every one later. Sharpening and Contrast, etc, are better done in Photoshop as you get *far* more control of the process that way including only adjusting a portion of the image.
The fuzziness really appears to be that you have blown out the highlights completely on the sky behind and in between the leaves. When viewed on the camera with the blinking highlights display, I'd expect that area to be blinking. Also, in Photoshop, are there areas that read 255 255 255? If so, this is your problem as it indicates you have lost all information in that area. Based on the jpeg, I can't tell for sure, but this looks likely.
Exposure subject brightness range on the D100 is better than film, but overexposing the highlights can kill an image.
The white balance in all instaces was set at a specific value - flash in this instances - and don't recall any significant wind. I also shot Raw with no sharpening but with normal tone.
I also understand and believe that Bob and you are right about the over-exposure and in many instances with these pictures you can see that the exposure difference between the underwood and the sky was tremendous. What surprised me was to find this problem on the picture of the house with a grey sky in the background and it leads me to conclude that the tone setting may also have been a factor and will try with less contrast.
...you should be able to change the tone curve settings of the existing image in Nikon Capture and see if you still have enough latitude left to retain some of the highlights. When you've got such a high-contrast image it's well worth the time to try a "virtual neutral grad" by exposing two identical images of the same scene about 2-3 stops apart in exposure. There are various ways of recombining these images to retain both the highlights and the shadows in a form that can be translated to print.
I'll play with the tone and see what comes out of it but unfortunatelly I don't think it will yield much as I tried many approaches with curve correction to no avail. However, I will definitelly try the two or tree images approach and see what I get. With the braketing option I should be able to see quite quickly what effect it has on the result.
...at the Nikon Digital School. The instructors pretty much always have their cameras set to less constrast to prevent the tone curves from adding too much contrast to the image before processing in Photoshop. Obviously, it doesn't matter in RAW mode since you can always change it later, but can be very important if you are shooting JPEGs.
These picture were taken handheld with a 24-85mm as well as an 18-35. I though that it maybe related to vibration and I tested with the anti-vib on and tripod and noticed a slight but not definitive improvement.
AS I mentioned before, I used an F100 as may main camera sharing the same lenses and same handheld approach (D100 was my B roll if I can put it this way since I was getting acquainted to this camera) and the problems encountered and discussed above are non-existant.
Well, I definitely like your work. I guess my comment/question is, and maybe I am a little more critical of technology since what is possible and what is actualy is driven by a slow consumer market... IS this just possibly that we are running into a limitation of our 6 megapixel camera? The same picture in slide film would have some 70+ Megapixels and film would be 50+ Megapixels.
Don;t get me wrong I LOVE my D100 and feel like a traitor saying this, but I had to pose the question.
The Nikon Super CoolScan 4000 ED scans at 4000 spi. For a slide or negative that is roughly an inch by an inch and a half then, this works out to 24 megapixels.
A 50+ megapixel image would be roughly 6000 spi with a 70+ megapixel image (from 35mm film) requiring about 7000 spi. What scanner are you thinking of?
Also, simply comparing pixel counts is very misleading since film scans contain a lot of information regarding film grain, something digital captures lack so they can be smaller and still have the same amount of real informaton.
As to exposure differences between film and digital, my D100 has the ability to capture about 6 to 7 stops of subject brightness range while still retaining detail, versus about 5 stops for slide film. Digital is more unforgiving of overexposure than slide film, but can yeield tremendous amounts of detail from otherwise completely underexposed images.
Is this a limitation of our 6 megapixel cameras? If you want to call it that. Film has limits and so does digital. If we are critical of one, we need to be equally critical of the other.
Point well taken Bob but this is something that I've also been wondering. One thing you have to take into consideration is the complex logarythm that goes into any digital camera when it comes to "constructing" the image. I will not go into the detail as I am sure you know this issue better than I. However, Foveon is addressing this complex issue and seem to have put their finger on an obvious advantage of film and that is color sensitive layers.
I love digital and it allows me a level of control I only could dream about with film, but it may have some limitation when compared to the "output" of film. The good thing is that Kodak has come up with their 14 Megapixel full frame CMOS (I understand they have some problem but I am looking forward to test drive this camera) and Foveon is only started with their SX3 (I think that's what theyc call it) so the near future look grand.
Nikon has stuck with the smaller sensor size and even come out with the 12-24 DX lens (although I wish they would ship it ). They didn't ship a new digital camera at PMA which disapointed a lot of folks but must mean they are working on *something* new. A 25mm diagonal 10 megapixel Foveon chip would look pretty darned nice in a new D2X
>No, Foveon has announced a second chip at 10 megapixels. >Both are listed on their >products page.
I see 2 chips, the 10M and the 5M. The 10m is a 3.4mp (3.4 * 3 channels = 10 million) chip with a larger size. The 5m is a 1/1.8" sensor designed for consumer cams with 1.5mp output. Am i missing something?
Believing marketing hype is an act of faith for Sigma devotees and the Foveon faithful. They would have to address the more significant shortcomings of the chip to interest me even if the next chip did have 10 million pixel sites three layers deep. The low light and long exposure performance of the SD-9 is miserable, color rendition is often poor.
And considering how many folks in this forum love to work with out-of-the-camera jpegs, not many would be pleased at having the SD-9's raw-only mode capability with mandatory PC processing of image files. I'm not sure how much CPU horsepower it takes to deal with the Foveon raw data, perhaps this is just Sigma's choice to cut the cost of the camera. Or perhaps a real 10 megapixel Foveon camera will need Pentium 4 inside.
Here's my three cents worth, On the last two pics they look over sharpened? I have had some problems with the same type of photos. I would.. 1.turn camera sharpening OFF. 2.Manual focus, I hate auto focus for details. 3.Try to stay away from jpeg for more detail. I actually returned an 80-200 2.8 because of the focus/detail problems, actually turned out to be the autofocus/jpeg combination syndrom..Hope this helps. Harry