Jim followed up on my reply to his initial question by supplying a couple of sample photos that he was not totally happy with. I won’t reproduce them here since they aren’t my photos, among other reasons, but I will post a few comments and observations.

The two photos were both taken in Yosemite, this summer I think. I took a more careful look at a photo that I think I recognize as being taken along the Tuolumne River in Tuolumne Meadows, with Mt. Gibb in the distance. I have several observations about this photo:

1. The exposure seems pretty good. Judging from the jpg I saw, it may be slightly underexposed – the histogram stops a bit before the right edge and detail is missing in the shadows of the trees on the right side. (This detail might still be recoverable in the RAW file.)
2. It looks like the photo was made at a f/22… and this likely is a problem as far as sharpness is concerned. The best aperture for optimum sharpness on crop sensor cameras is generally right around f/8, plus/minus a stop or so. While closing down (e.g. – f/22) would seem to provide greater depth of field, it also emphasizes diffraction effects that decrease the maximum sharpness of the lens. Imagine that you were perfectly focused on an object at f/8; you would likely be obtaining the sharpest possible image of that object. Now decrease the aperture to f/22; this object in the focus plane is now less sharp than it was at f/8. In Jim’s photo, most of the scene is a pretty good distance away… and out of focus due to diffraction. There is little to gain by increasing depth of field, and I suspect that the small aperture is responsible for the fuzzy focus.
3. In terms of effective framing of the scene, this might be more effective with a tighter crop. There is not enough to hold my interest on the right and left margins or in the foreground, so cropping some of these areas out of the frame might produce a more effective image I think. (It might be possible to go for a squarer format and do most of the cropping from the sides.)
4. There is a strong blue cast on parts of the water’s surface, probably picked up from the sky. I would adjust this by selecting this area (with a feathered edge) and reducing overall saturation and especially saturation in the blues.
5. Other parts of the image might be improved by a bit more saturation – perhaps about a 10 or 12 setting in PS.
6. I would perhaps make a feathered selection of the distant ridge and by levels or curves adjustments give it a bit more contrast.

The other photo is of an impressive scene that almost all Yosemite visitors have gawked at – including me. It was shot along the road descending into the valley from the north, looking east into the valley.

The difficulty with this shot is mostly the lighting. This is a very difficult place to photograph unless the light is just right, typically very early or late in the day during the summer. It looks to me like Jim was there during the middle of the day when the light is almost directly overhead and is quite harsh – tough conditions!

I didn’t take a close look at the technical aspects of this image, though it does appear to be fairly well exposed. It may be a bit overexposed, judging from the histogram of the blue channel which picks up the intense blue of the sky – but again there is a good chance that this could be handled in the RAW conversion.

Dan

—–

Jim wrote to ask about exposure, color, and sharpness results on photos taken on his Canon 350D/XT. Here is what I wrote in reply to his original email:

First of all, in regards to exposure the histogram display is your friend. I have my XT set to display the very small version of the image after exposure along with the histogram display. The histogram can tell you quite quickly whether your exposure captures the full dynamic range. In fact, I often use the first shot as my light meter – I let the camera make the first exposure in aperture priority mode and then look at the histogram. Then I frequently switch to manual mode and set aperture and shutter speed based on what I saw in the histogram.

If the full dynamic range will not fit into the histogram display – in other words there is a sharp cutoff at the bright and dark ends of the scale and the highlight warning flashes the display – I will give priority to avoiding blowing out the highlights. In other words I close down or increase shutter speed so that I avoid losing detail in the bright areas, even at the expense of seeing the curve shift toward underexposure. When this happens there are still a couple of things you can do to salvage the shadow areas, especially if you shoot RAW mode: you can bring up the dark areas in your software on your computer if they aren’t _way_ too dark, or you can shoot multiple exposures of the scene (obviously on a tripod) with some optimized for highlights and others for shadows and then combine them in Photoshop. (I did that in a major way to create the Pacific Sunset from Windy Hill shots I recently posted.)

Are you shooting in RAW mode or shooting .jpg files? I always shoot in RAW mode because it does not compress the image data at all – it basically dumps the photosite data into a file. (Hence, “raw” I suppose.) RAW files contain greater dynamic range and give you more room to make post-camera corrections in software. In fact, you really _must_ make a number of post-processing adjustments in order to get the most out of RAW files. Plain unprocessed RAW files are not very impressive and, in particular, they are not very sharp looking. (jpg files may appear sharper straight from the camera due to in-camera sharpening and other automatic adjustments, but they cannot be improved much beyond this point.)

In general, when I convert my RAW files (use the Adobe Camera Raw component of Photoshop) I increase the contrast and saturation a bit (the exact amount varies), decrease the exposure (and sometimes compensate by adjusting brightness) so that highlights are not burned out, adjust the shadows setting to the lowest black level, and adjust for chromatic aberration if necessary. Yes, a lot of steps! But that’s not the end.

Once the file comes into Photoshop there is still a lot of work to do. At a minimum (in all but the rarest cases) I adjust levels so that the dynamic range of the image fills out the available dynamic range of my printer. Depending upon the image I may also make adjustments to curves, saturation, and so forth. And then, in classic style, I dodge and burn as necessary. On some images I employ other more exotic techniques (local adjustments to contrast, levels, saturation) to get what I’m looking for.

So a short summary:

No, I don’t use separate light meter, but I do use the camera

as a meter by means of the histogram display.

I try to get the histogram curve to fit into the middle of the

display and avoid blowing out the high end especially. (A good

curve goes _to_ the high end but not beyond.)

If the dynamic range is too great I take multiple frames at

different exposures and combine them later in software, or

if the problem isn’t to severe I can rescue the shadows from

a single exposure in the RAW converter.

I always shoot RAW.

Post-camera adjustments to levels, exposure, saturation,

contrast, and (especially) sharpening are necessary when

you work with RAW files.

—–

Jim wrote to ask about exposure, color, and sharpness results on photos taken on his Canon 350D/XT. Here is what I wrote in reply to his original email:

First of all, in regards to exposure the histogram display is your friend. I have my XT set to display the very small version of the image after exposure along with the histogram display. The histogram can tell you quite quickly whether your exposure captures the full dynamic range. In fact, I often use the first shot as my light meter – I let the camera make the first exposure in aperture priority mode and then look at the histogram. Then I frequently switch to manual mode and set aperture and shutter speed based on what I saw in the histogram.

If the full dynamic range will not fit into the histogram display – in other words there is a sharp cutoff at the bright and dark ends of the scale and the highlight warning flashes the display – I will give priority to avoiding blowing out the highlights. In other words I close down or increase shutter speed so that I avoid losing detail in the bright areas, even at the expense of seeing the curve shift toward underexposure. When this happens there are still a couple of things you can do to salvage the shadow areas, especially if you shoot RAW mode: you can bring up the dark areas in your software on your computer if they aren’t _way_ too dark, or you can shoot multiple exposures of the scene (obviously on a tripod) with some optimized for highlights and others for shadows and then combine them in Photoshop. (I did that in a major way to create the Pacific Sunset from Windy Hill shots I recently posted.)

Are you shooting in RAW mode or shooting .jpg files? I always shoot in RAW mode because it does not compress the image data at all – it basically dumps the photosite data into a file. (Hence, “raw” I suppose.) RAW files contain greater dynamic range and give you more room to make post-camera corrections in software. In fact, you really _must_ make a number of post-processing adjustments in order to get the most out of RAW files. Plain unprocessed RAW files are not very impressive and, in particular, they are not very sharp looking. (jpg files may appear sharper straight from the camera due to in-camera sharpening and other automatic adjustments, but they cannot be improved much beyond this point.)

In general, when I convert my RAW files (use the Adobe Camera Raw component of Photoshop) I increase the contrast and saturation a bit (the exact amount varies), decrease the exposure (and sometimes compensate by adjusting brightness) so that highlights are not burned out, adjust the shadows setting to the lowest black level, and adjust for chromatic aberration if necessary. Yes, a lot of steps! But that’s not the end.

Once the file comes into Photoshop there is still a lot of work to do. At a minimum (in all but the rarest cases) I adjust levels so that the dynamic range of the image fills out the available dynamic range of my printer. Depending upon the image I may also make adjustments to curves, saturation, and so forth. And then, in classic style, I dodge and burn as necessary. On some images I employ other more exotic techniques (local adjustments to contrast, levels, saturation) to get what I’m looking for.

So a short summary:

No, I don’t use separate light meter, but I do use the camera

as a meter by means of the histogram display.

I try to get the histogram curve to fit into the middle of the

display and avoid blowing out the high end especially. (A good

curve goes _to_ the high end but not beyond.)

If the dynamic range is too great I take multiple frames at

different exposures and combine them later in software, or

if the problem isn’t to severe I can rescue the shadows from

a single exposure in the RAW converter.

I always shoot RAW.

Post-camera adjustments to levels, exposure, saturation,

contrast, and (especially) sharpening are necessary when

you work with RAW files.

—–

(Note: I wrote this piece a back in 2006. While much of this information is still likely to be useful, I would do a few things differently today. Rather than edit the article – and rebuild the photo! – I’ll just point out one important difference here at the outset.

Today, I rarely use the Photoshop “levels” adjustments. Instead I much prefer to use “curves” layers. Curves can do anything you can do with levels, and much more. While you can use curves and levels to adjust the black and white points in your photo, curves also allow you to control mid-tone contrast, to make important color balance changes using the gray and white “eyedropper” tools and much more.)

---

Someone just emailed and asked a question about how to get a photograph that recreates the feeling evoked by the original subject of the photograph. He was frustrated because the images coming from his camera are not doing that. There are probably a million ways to approach this, but I thought it would be fun to explain the basics of the process that I used to create a recent photo, Pacific Sunset, Windy Hill. (I’m not finished with this photo yet, so the interpretation may continue to change.)

Pacific Sunset, Windy Hill. San Francisco Bay Area, California. December 30, 2006. © Copyright G Dan Mitchell. (Sales)

This photograph was made along Skyline Road on the ridge above the San Francisco Peninsula. On December 30 it looked as if high thin clouds might produce an interesting sunset, so I headed for Skyline and drove about a half hour along the ridge looking for the right spot. I thought that Windy Hill might be the spot, but I continued driving since I knew I had another 20-25 minutes before the moment I was looking for, just immediately after the sun slipped below the horizon over the Pacific Ocean. Soon I gave up on finding anything better and headed back to Windy Hill, arriving just in time for the light I wanted to photograph.

I pulled off the road and quickly got out and put my camera on the tripod with the 24-105mm lens that was already attached. I framed and shot a couple pictures and then realized that 24mm was not wide enough; so I went back to the car and switched to the 17-40.

The elements of the scene were spectacular. The sky ranged from intense orange-yellow where the sun had just dropped below the horizon, through reds and pinks on the clouds, to yellow-greenish hues a bit higher in the sky, and finally to deep blues above the clouds. New grass (yes, that happens in winter in California) was intensely green, and some of the sky colors were reflecting on a patch of last years grass that had blown over.

The problem was that the dynamic range was absolutely huge – there was no way to capture it in a single exposure. (I don’t use ND grad filters.) So I had to bracket, starting with exposures for the extremely bright portion of the sky still lit intensely by the sun and, 7 stops later (!), opening up to capture detail in the shadows in the foreground. So, I ended up with 6 or 7 different exposures of each framing of the scene.

Back home I opened the RAW files in Adobe Bridge. As I looked them over it seemed like I could get sufficient details from the scene by combining three different exposures: one for the sky, one for the middle distance hazy hills, and one for the foreground grassy area. Here are the three original images (somewhat modified during the ACR conversion):

These three images were brought into Photoshop as “Smart Layers” (thus retaining their connections to the original editable RAW files) and then combined using masks. The initial masks were created using either the Magic Wand or Polygon Lasso tools and then fine tuned by painting on the masks. As I worked I returned to the RAW files a few times to do additional fine tuning of the conversions. The resulting composite image follows:

There were two large power transmission line towersin the distance just below the horizon on the right side of the scene – if you look closely you may be able to spot them. Yes, the next step was to use the Stamp tool to remove them from the photo. I don’t think that is worth adding another example here, so onward.

Next I made a basic Levels adjustment to the entire image. The bright end was already close to maxed out so I did not adjust this much at all. However, there was some room at the dark end to create slightly darker blacks. Here is the result:

There was not enough punch to the middle of the photo encompassing the reflected light on the foreground green grass and dried grass on the hill and the hazy canyons and ridges below the horizon. I made an elipse shaped selection in the middle of the image, feathered the edge, and created another Levels adjustment level. However, I positioned it above the two layers holding the foreground and middle portions of the image, but below the layer containing the sky. In this way I could alter the levels of the lower portions of the image without affecting the sky, which already was looking about the way I wanted it. Here is the result:

I wanted to see a bit more of the reflected light on the shallow U-shaped contour of the hillside, so I painted another selection along the green ridge in the left foreground and the lighter ridge on the right and made another levels adjustment. (I’m sill considering how much to emphasize this effect – I’m thinking the current version is a bit too bright here.) Result:

Next I fine tuned more or less the same area (but a bit larger) using a Curves adjustment:

I wanted to draw the eye toward the center of the image including the curve of the two hills meeting and the dark hills beyond. But I did not want to increase brightness or contrast here at this point since things were already beginning to go a bit too far in that direction. So instead of adjusting this area of the photograph I selected the area outside of this section and darkened it a bit using a Levels adjustment:

Since the grass reflected not only the brightness of the sky but also some of its brilliant colors, I created a saturation adjustment layer in which I greatly increased saturation. Then I created a mask for this layer set to “Hide All” – in other words, the mask eliminated the increase in satuation. Then I selectively painted on a few small areas of the mask to allow bits of the saturated colors to show though a bit in a very few selected areas. The effect is subtle, but you may be able to notice it a bit on the foreground ridge on the right:

As I worked on the earlier portions of this process, I had also been doing some dodging and burning. The final dodge/burn layer is added in this example:

Next was some sharpening, done in two stages. The first used “Smart Sharpen” to show fine details captured in the file that are not visible in the original RAW image. (Sharpening is a necessity when working with RAW files.) Here is a 100% crop (enlarged way beyond print size!) of a section of the unsharpened image followed by an example of the sharpening:

Next I applied Unsharp Mask to create “local contrast enhancement” that delineates larger areas of a given color/brightness a bit more:

At this point I felt that a slightly taller and narrower crop would be better than the original 3:2 ratio, so I cropped a bit off of both sides (more from the left) to provide the 4:3 ratio in the (current) final version:

Pacific Sunset, Windy Hill. San Francisco Bay Area, California. December 30, 2006. ©