Table of Contents Introduction.......................................... 3 Acquisition Processing Exposure and the Histogram The Digital Workflow Definition of Terms ........................... 4 The Transfer Process......................... 29 The Histogram Defined .................... 6 IPTC Caption/Transfer Options....... 30 Evaluating Luminosity ..................... 7 Performing a Transfer ....................... 31 Understanding Color ........................ 10 Batch Processing ................................ 32 Working with Color .......................... 11 The Selection Process........................ 33 Image Processing Metering Nikon Capture Program Setup ......... 34 Which one do I choose and Why?..... 13 Great Egret ........................................ 37 Matrix .................................................15 Black Crowned Night Heron ............ 41 Spot .....................................................16 Jean-Claude ........................................ 46 Center................................................. 17 Chilean Jasmine ................................ 50 The Vietnam Memorial .................... 55 Composition Hard Rock Stratocaster ..................... 58 “the rule of thirds” ............................ 18 Moose at Sunrise ............................... 61 Guiding the eye ................................. 19 Tomb of the Unknown Soldier ......... 65 Angles, near/far technique ............... 20 Digital DEE: an overview ................. 68 Closeup work ..................................... 21 Jackson Lake, Grand Tetons ............. 71 Using the focused plane .................... 22 Breaks in convention & luck ............ 23 Photoshop Processing Architectural lines ............................ 24 CS RAW — BC Night Heron............. 74 Post-processing .................................. 78 Tips and Tricks ................................... 26 Images for your Enjoyment ............... 28 Lenses (from Digital-Images.net) ......... 86 Thanks are due to Dave Cheatham, Andrew Chan, Mike Worley, Kevin Gallagher, Vince Battaglia and Rick Wallace for their assistance in this project. All text, graphics material and images © 2004 — Ron Reznick All Rights Reserved RR Design 636 E. Harvard Rd unit B Burbank, California 91501 Telephone: (818) 843-8212 http://www.digital-images.net 2 3 The best photographs result from well- Special attention will be paid to the digital exposed, well-composed, properly- workflow, including organizational methods; processed images. This book is not intended batch-processing for pre-selection (to remove to be a camera or software manual which unwanted duplicates or mistakes from the expounds on features… instead I prefer to group to be processed manually); the selection go into detail regarding various techniques process itself; and a comprehensive treatise on which I use for the acquisition and process- image processing using Nikon Capture as well ing of images. While I use Nikon gear, the as some post-processing techniques in Pho- techniques which I will describe are appli- toshop. In my workflow, most of the image cable to other manufacturers’ products, and processing including the color work, White techniques described for processing images in Balance, EV, Levels, Curves and Gamma Nikon Capture are applicable to other RAW processing are done in Capture, with post- processing programs and image editors with a processing being primarily used for touch-up, few alterations. rotation, cropping, resizing, etc. The section on processing is documented with screen cap- Some acquisition techniques which will tures and presented in easily-understandable be discussed include the use of the histogram language. Zoom in to examine image details. to determine proper exposure; the use of the various metering modes along with exposure An article on Lens Selection and sample compensation to achieve that exposure in images are also included, along with lessons various situations; some reasons for the use of showing step-by-step processing of these im- aperture priority, shutter priority or manual ages, which were selected to illustrate com- mode; techniques for composition of various mon histograms. There will also be auxiliary sorts of subjects; and the use of depth of field files included to help you to understand lumi- for subject isolation or to assist in the creation nosity and color, and how they apply to both of a wide-angle composition. metering and evaluation of the histogram. 2 3 Exposure and the Histogram: A Definition of Terms Many people reading this book are probably aware of the terms and features I’m going to discuss. If you already know this stuff, bear with me or skip forward a bit. One of the most powerful things about digital cameras is the ability to see what we’re doing… not only can we see the image itself, but we can turn on a Histogram: a graphic distribution chart that shows the image in terms of light levels and the number of pixels that are present at each level. The histogram is a chart running left to right from 0 (Black) to 255 (White) with divisions at 64, 128 (middle grey), and 192 (see image to right). The histogram allows us to see the ex- posure. If you can evaluate the light levels (luminosity) of the scene, you can look at the histogram and determine whether you are The numbers below the vertical divisions underexposed or overexposed — for instance, correspond to luminosity level at each division I was with a student in Yellowstone National in RGB value (RGB is Red, Green, Blue). There Park recently, and we were shooting some are also greyscale (K) equivalents to each of pronghorns in strong light. I pointed out to these values, from left to right on the histogram him that the white part of the pronghorn was they are 100%K (Black), 75%K (64RGB), 240, and to expose accordingly. He took a shot, 50%K (128RGB), 25%K (192RGB) and checked his histogram, and saw the highlights 0%K (White). The greyscale values correspond dropping off just past the 192 division on the to the percentages seen in the greyscale below. histogram. He applied +0.7 EV (Exposure Value Compensation) and took another shot, nailing the exposure right on the money. What does the EV control do? The EV control alters the exposure by raising or low- ering the aperture or shutter speed. If you If you overexpose an image, the contrast is have your camera set to Aperture Priority, lowered and it is possible to severely overexpose the aperture (lens opening) is fixed. When highlights. This will result in a large spike on the in Shutter Priority the shutter speed is fixed. right side of the histogram (255) and a loss of The EV control alters the one that isn’t fixed, highlight detail. opening up the aperture on the lens or lower- ing the shutter speed for positive EV, and clos- If you underexpose an image, the shadows ing down the aperture or raising the shutter and midtones start to get too dark (block up), speed for negative EV. and to recover this underexposure yields a noisy result. 4 5 Another powerful feature of the digital camera is the ability to alter the ISO (film speed) without changing film (since we don’t If you set your aperture to get the depth of use film). This works the same as the volume field you want, and the exposure is correct but the control on your stereo -- it cranks up the gain. shutter speed is too low, raising ISO increases the As the signal-to-noise level is fixed (the noise shutter speed at the expense of additional noise. is increased along with the signal level when You’ll have a sharper shot, but with more noise. you “raise the volume”), raising the ISO speed increases the shutter speed for a given aper- ture and exposure at the cost of a higher level of noise. On the following pages, I am going to get into some rather exhaustive detail regarding the histogram, and will show you how to use a greyscale and color charts to train your eye to see luminosity (light levels) in the subjects around you. This will give you a reference to light levels that will assist you when looking at your histogram, so you can judge when an image is overexposed or underexposed. I am providing you with several files that will allow you to calibrate your eyes (and incidentally adjust the brightness and con- trast of your monitor in case you aren’t using calibration tools), teach you about color and greyscale luminosities, and RGB color levels. If you spend some time studying these as out- lined in the pages to come, you will soon turn into a human light meter and will be able to shadows midtones highlights closely estimate the level of highlights and midtones, and therefore will be able to look at the histogram and you will know whether or not you have the exposure correct. The freedom this gives you is tremendous. I wish The left side of the histogram shows the you could see the looks on the faces of my stu- shadows, the right side shows the highlights. dents when they realize they are consistently Midtones are in the middle. You want to acquire hitting exposures within one-third stop, and the ability to estimate luminosity values of the often within one-tenth. BIG smiles. midtones or highlights so you can adjust the EV. The histogram above shows the RGB values Plan on reading the following informa- associated with each division of the histogram as tion several times to understand it. This is not well as a greyscale, and a very important piece of rocket science, of course, but the information information: the distance in EV between each of requires some digestion before it becomes the divisions from the mid-shadows to white. If second nature. Open the greyscale and color you practice, you will learn to evaluate the sub- charts in your image editor several times per ject and highlight luminosities, and can look for week. Put your mouse cursor over the color them on the histogram. With known values, you patches and get to know the RGB levels (look can make adjustments based on a knowledge of in your Info palette). Soon, you’ll have a good the distance between the divisions and adjust the feel for luminosity. exposure as close as your camera will allow. 4 5 Exposure and the Histogram: The Histogram Defined Properly exposing an image makes a major difference in the amount of im- pact an image has, the character of highlights and shadow detail, and noise levels among other things. It also reduces the amount of processing effort considerably. Our goal is to consistently expose images within one-third of a stop (underexposure is preferred to overexposure to avoid blowing highlights and losing detail), although it is pos- sible to recover images that are underexposed by as much as a whole stop or overexposed by as much as 1⁄2 stop depending on your stan- dards for noise or loss of highlight detail. Underexposed images generally leave a gap on the right side of the histogram and are bunched on the left side, overexposed images generally leave a gap on the left side and are bunched on the right. It is possible to have an overexposed or underexposed image that does not look this way, for instance if the brightest highlight was 200 and you show 250, you’re 2/3 stop overexposed. To the right are two histograms of the same subject from Nikon Capture (one under- exposed and one overexposed). Note in the upper histogram how the highlight data, shown at the bottom of the graph, drops from the midtone region just after the centerline (128 division), and the brightest highlight peters out just after the first division left of the right-side margin (192 division). The proper exposure for this image would have shown the drop down from the upper midtones at 200 and the brightest high- lights just barely hitting 255. The upper image is underexposed by 1 stop. The lower image is overexposed by just under a stop. Note the spike at the 255 mar- gin (right side) that goes all the way to the top, and especially the height of the data above the baseline before it rises to the spike. Also, no- tice the significant gap on the left side. When you see these three characteristics together, it indicates overexposure. 6 7 Evaluating Luminosity This image to the right is what you I have provided you with a number of files on will see when you open the file the CD. Throughout this section, you will want “Greyscale.tif”. You will notice that 128 Grey to have these files open in your image editor to be (the strip above the greyscale) is equal to 50% able to follow along with the explanations. grey. This corresponds to the center dividing line on your histogram. You can verify the Open the file “Greyscale.tif” on the CD. RGB luminosity of each of the K-percentages (listed in the white strip below the greyscale) by using the mouse cursor in your editor (e.g. Photoshop) and looking at the RGB values in the Info palette. This is an easy way for you to get familiar with the RGB levels at various A 5% greyscale. The percentages listed be- luminosities, it will make it much easier to es- low the greyscale are in K-values (derived from timate where highlights and midtones should CMYK: Cyan, Magenta, Yellow and Black). fall on the histogram, and will greatly improve Black is 100% K (100% Black), and white is your ability to get the correct exposure. 0% K (0% Black). The vertical lines on the histogram cor- respond to the greyscale as shown in the table to the lower right. The divisions from the left to the right margin are listed in both K-values (as seen on the Greyscale) and as RGB-values (which can be seen on the greyscale by hold- ing your mouse cursor over each patch). The RGB values are also shown on the histogram shown above the table, which you’ll recognize as the one seen on page 5. By memorizing the various luminosity levels in the greyscale above, and comparing your knowledge of these luminosity levels with what you see around you, you will have an extremely good feel for how much EV compensation to apply to any given scene de- pending on the metering mode you are using. Remember that the meters will average the scene to ANSI grey, which is roughly 102 grey or 60 on the greyscale above. It is most useful for you to know 128 grey, however, as it cor- Division on Histogram % K RGB responds to the mid-line on your histogram. If you meter on a subject that is 128 grey, Left Margin 100 0 you’ll see a spike in the center of the histo- gram. If the spike doesn’t appear where it’s between left & center 75 64 expected, you need to apply EV compensation Center division 50 128 — if it is left of the expected point you apply between center and right 25 192 positive EV, and if it is right of the expected Right Margin 0 255 point you apply negative EV. Clear as mud, right? I’ll give more detail on the next page… 6 7 To the right is the familiar histogram, and below it is an image that shows greyscale luminosities that correspond to the divisions in the histogram. You have a large version on the CD called “PrintGrey.tif”. You can print this image to create a test-target (it will print on a letter page at 10” x 7.75”. The test target can be used to help you learn how to adjust EV (exposure compensation). Notice in your shot of the print where each spike shows up on the histogram, and apply EV as necessary to adjust the result so that each spike lands on each division. This will be the correct ex- posure. There is one EV between 192 and 255, and 1.3 EV between 128 and 192. (There is 1.7 EV between 64 and 128, but hopefully you will rarely underexpose by this much). This information is very useful to help you determine how much to alter the exposure to correct for what you see on the histogram. If you are metering something that you know is 192, for instance, and it just hits the right border of the histogram, you are 1 stop overexposed. If you apply -1.0EV, you will either be right on the money or just a bit over- or underexposed, and can make the final adjustment in Capture, knowing you’ve pre- served the highlights. You can adjust exposure in 1/3 stop increments by setting your camera to 1/3 stop in the menus (this is the default). One thing that I should emphasize here: many people try to preserve all highlights, even those specular highlights that are blown in real life (e.g. sparkles on water, etc.). This is a mistake that will results in underexposed images, and should be avoided. If it’s white in real life, you should expose accordingly and al- low a spike in the histogram on the right side There is a larger version of this image — otherwise you are going to severely under- above on your CD called “GreyPrint.tif” which expose the rest of the image, and recovering will print on an 8.5” x 11” letter page. You can this using gamma and EV compensation after use this print to practice making EV adjust- the fact will result in lowered shadow detail ments, as you will see spikes at the correspond- and noise in darker areas of the image. I will ing divisions on the histogram when the image discuss how to use gamma and EV to adjust is properly exposed. an image in the processing section. Let it go if it’s white. Don’t try to hold the highlight. 8 9 The histograms to the right show the same subject which has a balanced character and midtones that occur exactly at 128. The top histogram is properly exposed, the middle one is overexposed by 0.33 EV, and the bottom one is underexposed by 0.33 EV. Of course, this can be corrected after the fact in Capture, but it’s good practice to correct it in-camera. By memorizing the luminosity greyscale and the regions of the histogram, and taking test shots for a while, your eye will become sensitive to subject luminosities and you will soon be able to estimate the luminosity closely without taking test shots… if you do have time you can check yourself and make required EV ad- justments as necessary. One other thing while we’re on this sub- ject… many people seem to think that the histogram has to show an even distribution, or that it has to show an exposure that goes all the way to the right margin, or some other preconception of what the histogram should show. What this means is that they really don’t know what a histogram is: a distribu- tion curve that shows how many pixels are at a given luminosity value. There simply is no specific required look to a histogram… if there is no white in a scene, and the bright- est highlight is 200, your histogram should show the data dropping out just past the 192 division mark. If there is white in the scene, and there is a fairly strong specular highlight, you will have a spike at 255 (right margin). If there is an odd color balance, you could conceivably achieve full saturation at a lower value than 255 (e.g. when shooting a red rose on a black background — a subject that is of- ten overexposed to the point where all detail in the rose is lost). In the next few pages, I will show you some files that I have provided which will help you learn about color and greyscale, and discuss how you can use these files. These will assist you in understanding how the camera sees color, and how various colors will be ex- posed so you will know what to do with that pesky histogram to make sure you are getting the proper exposure in situations where the color balance of the scene is unusual. 8 9 Understanding Color: Defining the Situation The differences between exposing for Some situations where people — even grey and for color are somewhat professionals who are used to working with complex. I am going to tie the two principles their light meters but who don’t have an in- together with the aid of two tools that I have stinctive feel for color — make this mistake provided for you on the CD: are when shooting red rock (e.g. at Bryce Can- yon (Introduction page and the image below) “RGB135_Grey.tif” and or Sedona, AZ), or when shooting primary “RGB135_Converted.tif”. red subjects like roses, etc. Now that you know what the histogram An example of how this will help you is is showing you — the number of pixels at the case of the red rock I mentioned… at Bryce each luminosity level from 0 (black) to 255 Canyon’s Visitor’s Center, there are a number (white) — and you know that the histogram of printed photographs taken professionally and shows you a greyscale average because there printed in everything from books to NPS’s bro- are no separate Red, Green and Blue (RGB) chures. Every single shot I looked at showed over- histograms displayed, you need to be able to exposure of the reddish-orange rock to yellow. If understand how the camera evaluates color so you set exposure properly, you can maintain the you can bias your evaluation of the histogram red channel without blowing it to yellow, or los- depending on the color balance in the scene. ing highlight detail. If the scene has a good balance of Red, Green and Blue (RGB), there will not be a problem, but quite often you will find yourself in situ- ations where one color predominates, or the color balance is skewed in a way that favors one color over another. The image sensor, which reads grey, has a filter over it that limits the light that en- ters each well. Most of the sensors we use have what’s known as a Beyer Pattern filter (which is shown on page 12), that alternates RGRGRG or GBGBGB per each row, limiting the color that enters each well to wavelengths corresponding to these colors. This means that the sensor, which has twice as many Green-sensing wells as Red or Blue, has a color bias towards green and yields a higher resolution when there is green blended with red or blue. The issue is that you can achieve full satu- ration in the red channel quite easily, and in situations that have a predominance of red, you will oversaturate the red channel long before the histogram shows the highlights at 255. You need to be able to evaluate color balance to know when this will be a problem, and the only way to do that is to know how colors are put together and their greyscale equivalents. 10 11