Part 15 - Camera Sharpening

Many consider "sharpening" to be the remedy for "out of focus" images. It's not. Out of focus pictures will ALWAYS be out of focus. Sharpening only enhances the edge contrast within an image to create the illusion of focus improvement. But the picture is still out of focus.

A complete article on sharpening is located on my sister site, Hub's iDarkroom - Part 13, Sharpening, Less is Better. Please read this article to get the complete "sharpening" picture (excuse the pun).

This brief post covers the basics of the sharpening function found on most of today's DSLRs.

Sharpening selections on a Nikon DSLR menu

Because of the way digital images are captured by the camera's sensor, a small degree of sharpening is required for EVERY DSLR picture. This sharpening can be accomplished in the camera as it is processed and stored on your data card.

One exception is in the case of RAW image files. RAW images are always recorded WITHOUT sharpening. RAW images still require sharpening but this is done either 1) as you open the image in a RAW converter program or 2) within most digital imaging programs like Photoshop.

From my experience, I believe sharpening is more flexible and precise in the computer using image editing software. For this reason, I routinely use LOW or NO sharpening within the camera. I recommend this as a starting point for any new DSLR owner who processes their own images with a program that provides sharpening options.

This discussion, including illustrations showing the impact of sharpening on your pictures, continues in Hub's iDarkroom.

Please let me know if you have questions or comments.

Part 14 - Yes You Can

In my last post, I mentioned that a picture I judged in a recent local photo competition was the trigger for writing Parts 13 and Part 14 of this blog. Shown below is my version of the image as it was presented as a print in the contest:

South Falls in Silver Falls State Park, Oregon

By the way, when you inevitably pass through your "waterfall picture-taking" period, you have to go to Oregon -- the waterfall capital of the world.

Although this waterfall picture is dramatic, to me the dark areas of the image are too dark and lacking in detail. The ledge from which the falls begin is totally black and difficult to recognize. The lush greenery below the falls is nearly lost.

At the same time, the sky and trees in the background are striking and well exposed. Was this a photographic exposure error? Since I went to the same location to recreate this photograph, I can say that the problem was created by the high dynamic range of the original scene. This is the correct exposure for the sky and background trees. But it's underexposed for the cliff and greenery. Looking at my camera's histogram, it becomes apparent that I've walked myself into a high dynamic range situation:

Camera's histogram for waterfall picture above.

The lightest part of this image is represented on the right side of the histogram, and those densities are all within the borders of the graph. The camera is saying, "No problem with the light areas of this picture (the sky and farthest background trees). They will reproduce just fine."

The histogram goes on to say, "However, forget about the darker areas of the picture (left side of graph). Those densities -- like the cliff and dark foliage under the falls -- are dead zones. They will reproduce as dark with minimal detail or totally black."

I know I can use the camera's exposure compensation controls to move the densities within the histogram. I can also do the same thing by switching to manual mode and selecting a larger aperture opening (smaller f/stop number). Making these adjustments will allow me to move the entire density distribution toward the right (in effect lightening the whole picture).

For the sake of demonstration, I did just that. I shot another picture at 1.5 f/stops lighter using my exposure compensation control (opened the aperture 1.5 stops):

Here's the resulting image and histogram at 1.5 f/stops increase in exposure. It's better, and it's worse. As the histogram indicates, the entire graph moved to the right and detail is now visible in the shadow areas. The bad news is the increased exposure has moved my highlights beyond the far right hand border of the graph. Now I've lost the sky and background tree detail. I've "blown out" the highlights.

This seems like a "no win" situation. I can't make an adjustment that will bring both the highlights and shadow areas of the picture within the borders of the histogram at the same time. The dynamic range of this scene is beyond the capability of my camera's sensor to record.

Ideally, the solution would be a combination of these two images that captures the highlight and shadow detail important to this picture.

Enter digital imaging technology. You can do just that using your camera, tripod and any imaging software that accommodates High Dynamic Range processing. Here's how I married these two pictures:

Camera mounted on tripod to create multiple exposures for HDR processing.
Notice the histogram display.

• Mount camera on tripod and compose the picture in viewfinder (make certain the camera is rigid and will not move throughout this process)
• Paying close attention to the camera's histogram shoot an exposure that captures the shadow densities (shadow portion of the graph is within the left hand borders of the graph)
• Now take a second picture that captures all the highlight detail within the right hand side of the histogram plot.

This example uses only two pictures to create the final HDR image. To be safe, I recommend you take three pictures (one over, one under and one at the suggested exposure) in most situations to ensure you capture all the entire range of densities your picture requires. In some extreme high dynamic range instances, I have seen some photographers take as many as 5 separate images at different exposure settings.

The separate image files are now downloaded to your computer. In programs like Photoshop, you next identify the individual image files that you want to combine. The program will automatically do the work for you, or allow you to "tweak" the image as it is being processed.

Below is the image that resulted from combining my two exposures:

Final combined HDR image

Final HDR histogram with highlights and shadows within reproducible limits

When compared to the original photograph and histogram, this image has good detail and color in both the shadows and highlights. In essence, this digital manipulation has allowed me to photograph a scene that is beyond the dynamic range capability of my camera's sensor to capture in a single exposure.

Some notes:

• a tripod must be used to ensure that each image aligns in every detail (otherwise the image will appear "soft" or out of focus)
  
• objects within the picture area should be stationary. (It appears I violated this rule in the waterfall picture. But, in this case, the motion of the water didn't affect the image's visual impact. Just be careful.)

Your autumn landscape pictures will generally fall within these parameters. The possible exception will be leaves moving in the breeze. But light motion in distant leaves will usually not be noticeable. Avoid hurricanes and tornadoes.

A future post in my digital darkroom blog, Hub's iDarkroom, will cover the details of using the HDR feature within Photoshop.

How will you know when HDR should be considered as an option? Keep your eye on the camera's histogram EACH time you take a picture. It will graphically tell you.

Hope this helps during your fall shooting excursions. If you have any questions or comments, please let me know.

Part 13 - "I Can't Take That Picture."

With Mother Nature Productions hard at work directing her annual fall epic of color, millions of photographers will be spending much of their time outdoors documenting the splendor of this year's display. While judging a photo competition recently, I was reminded just how challenging, and often frustrating, landscape photography can be to beginning photographers. So, as we gather our gear and head out to capture that perfect fall image, I thought this would be a good time for a photographic reality check.

Because the fact of the matter is: Even after carefully calculating your exposure, what-you-see may not be what-you-get in some lighting conditions. You might even walk away from a scene saying, "I can't take that picture."

This post lays the ground work in basic terms for Part 14, that will show you how to capture those dramatic landscape pictures containing an extreme range of densities -- just what you're likely to encounter this fall.

In the earlier two-part post on "Exposure is Everything", you learned much about the limitations of capturing and reproducing an image. Let's take some time here to expand on this information.

It all starts with the human eye...

Although our eye does not physically capture the picture that's recorded on a camera's data card, it is the image we "store" in our brains. This leaves open the possibility that the image our brain remembers is not the same picture that's stored as ones and zeros in the camera. As a matter of fact, this discrepancy occurs in nearly every picture we take. The difference between our eye's perception of a scene and the camera's digital version of that same image may be minor and of little consequence, or it may be extreme and of vital significance to our final printed picture.

Our eyes don't take snapshots. The activity of the eye is continuous, and our mind assembles multiple, adapted views to construct the image we remember. One of the most important analyses performed by our eye/brain combination is that of being able to see detail in the brightest and darkest parts of a scene. As the eye scans a scene, it constantly adapts to the available reflected light in every portion of the scene in an attempt to find and recognize the details in all areas of the scene.

This includes detail in the brightest snow as well as individual pieces of coal in a coal bin. (From an evolutionary perspective, this makes perfect sense. Who knows in nature where danger lurks -- in the shadows?-- or where the next meal may be found -- high in a brightly lit tree?) Having found all the perceptual detail, the eye assembles the image we see and remember -- moment by moment. Speaking in photographic terms, the range of brightness to darkness within which we can distinguish detail is called the image's dynamic range. Biologists call it survival.

When on location, the human eye can distinguish details in the snow, as well as in the shadows of the blue spruce tree, in this high dynamic range scene.

The camera is not nearly as fortunate. The lower dynamic range of today's DSLR sensors "see" far less detail in the brightest (highlights) and darkest (shadows) areas of a scene. Falling back on the previous posts on exposure, this difference in dynamic range can be illustrated in terms of a camera's histogram display.

An "Eye" histogram

Here's the histogram for a scene as the eye/brain might "record" it. Notice that all densities in this particular scene fall well within the boundaries of the graph. And we "see" the details in all areas.

The camera's histogram of the same scene

But, the same scene captured by a DSLR might appear as the histogram above. It's obvious that the lower end of the graph (shadows) and the upper end of the graph (highlights) extend beyond the borders of the histogram. These extremes will contain no detail in the final print.

The result in this example is that the eye "sees" more detail in the shadows and highlights than the camera's sensor is capable of recording. The dynamic range of the normal human eye exceeds the dynamic range of a camera's sensor.

Here are the same results in picture form:

Although this may be the picture your eye "saw" and remembers,

The camera "saw" and recorded it quite differently. Notice the clouds and trees are light and without detail, whereas the lake lacks detail in the darker shadow areas.

The difference between the dynamic range of the human eye and the sensor on a modern DSLR can also be thought of in terms of f/stops. Although f/stops are nearly impossible to accurately specify for the eye because of the way it functions, some approximations do provide a clearer understanding of the photographic problem.

• the dynamic range (darkest to lightest part of a scene) found in nature on a bright sunny day is approximately 12 f/stops or more
• the dynamic range of a normal human eye is approximately 10 to 15 f/stops
• the dynamic range of a typical DSLR sensor is approximately 5 to 9 f/stops.

Even if these numbers are not "dead on", they do point out the fact that what you see may not be what you are able to capture with a camera. There is also one piece of the puzzle missing in these numbers:

• the dynamic range found in nature on a cloudy day is 3 to 5 stops.

Aha. Dynamic range changes in nature depending on the type or quality of light striking the scene. That's true. Consequently, many pictures we take may fall within our camera's ability to record accurately. It's the exceptionally high dynamic range landscape-type images found on brighter days that concerns this post.

Jargon Alert: High Dynamic Range is often seen abbreviated to HDR.

So, was the camera's exposure incorrect in the mountain and lake scene above? NO. The exposure was as correct as the camera would allow. The dynamic range of this particular scene exceeded the camera sensor's capability to record. And, YES, the camera's histogram would have revealed that this image was going to be problematic.

But take heart DSLR owners, the same dynamic range issues have faced traditional photography since the day film became available.

Does this mean I can't take pictures when the dynamic range is greater than 5 to 9 f/stops? Not necessarily. Stay tuned for Part 14.