Abstract - Digital photographs can be treated in a manner to preserve the photographic elements of an image, while using the capabilities of image manipulation software in a manner that does not fundamentally change the photograph. These techniques are similar to manipulations that could be performed in a traditional darkroom, as well as techniques particular to digital photographs that do not affect the image in an artificial manner.
Introduction
With digital photography comes the ability to produce photographic appearing images at will. People and objects can be juxtaposed in ways they never were in real life. These images have beneficial uses in both creative and especially commercial art. Photography, on the other hand, relies on recording a scene as it is, manipulating the image only as a whole, and not changing the juxtapositions or appearance of objects within a picture. While some measure of local control such as dodging and burning, lithographic masking and unsharp masking can be done in the darkroom, these techniques are difficult, expensive and time consuming.
By using photographic methods with software such as Photoshop, the difficulty of complex photographic manipulations can be greatly reduced and a photographer's productivity increased. This paper describes twenty techniques that I have developed or stolen for my own workflow. Using these techniques I am able to continue my development as a photographic artist. These techniques can, for the most part be performed with black and white and color images.
Noise
Digital cameras have an image sensor element that is really an array of tiny capacitors that change their charge holding ability according to the amount of light that falls on them. Because there is variability of the charge as compared between two capacitors, even in the dark, variations of the image intensity show up as digital noise. Such noise is somewhat larger than a single pixel because adjacent capacitors effect each other. Digital noise appears to be random across the sensor, however as conditions degrade to produce more and more noise, patterns tend to form.
The sensitivity of the image sensor, the ISO speed, has the largest effect on the amount of noise seen in a particular photograph. At low sensitivities, less than ISO 400, noise is hard to see in images. Moderate sensitivity, at speeds from ISO 400 to 1000 or so, noise is visible, but controllable. High speed photographs, above 1000 (I have shot photos as high as ISO 1,000,000, or 1 MISO) the noise is an integral part of the photograph, and must be dealt with in some way.
This is not new to digital photography, however. Film exhibits it's own type of noise called grain. The relationship of film grain to noise is comparable. However, in general digital photographs show less noise than do photographs on film, however, low noise films are more and more popular in todays dwindling film market. Even the stalwart Tri-X has been recently reformulated to produce less noise upon push processing.
Digital noise is usually coarser and more patterned than film grain. It is also poly chromatic, that is, made up of multicolored dots, where film grain is more uniform both in size and color. Because of this, digital noise is not as aesthetically pleasing as is film grain and methods and tools exist for reducing the amount of digital noise seen in digital photographs.
Most digital photographs contain some noise. The following techniques can be used to reduce, remove or modify the appearance of that noise.
Technique 1 - Gaussian Blur
The Gaussian Blur filter of Photoshop's Filters>Blur>Gaussian Blur menu can be used to smooth out out of focus areas of a photograph, such as a background.
- Select an area that requires noise reduction.
- Feather the section by 1 to 3 pixels (depending on the size of the image - bigger means use more pixels).
- Add Gaussian Blur with a radius of 0.4 or less.
- Optionally Fade (Edit>Fade) the Gaussian Blur to fine tune the effect.
A variant of Gaussian blurring is to blur only one channel in the same way. By examining each channel in the Channels palette, only the noisiest channel can be blurred, leaving the other channels untouched. Another variant is to first convert the color space to Lab Color (Image>Mode>Lab Color). Lab color is a special color-space that extracts the shades of gray into a separate channel called the L (for lightness or luminance) channel, leaving the color information in two other channels called the a and b channels.
Once the image is in Lab color, the L channel can be blurred in the same way as above and the noise in the selected area will be reduced, but delicate colors will not be affected.
Technique 2 - Lab Color, a/b Blurring
This technique does not remove noise, rather it makes noise monochromatic and more sharply focussed, and thus closer in appearance to film grain. This technique goes through the individual steps, however these can be applied automatically with a slider when a RAW format image is opened with the Adobe Camera Raw plug-in (the slider is labeled Color Noise Reduction.) To manually apply a/b blurring:
- Convert an image that shows mottled red/blue noise to Lab color mode (Image>Mode>Lab Color)
- Select the a channel in the Channels palette.
- Add Gaussian Blur (Filters>Blur>Gaussian Blur...) until the noise cannot be seen in the image. This can take as little as a 2 pixel radius for a relatively smooth images to over 20 for very noisy pictures.
Note: Do not add excess blurring to the a and b channels. That will dilute, that is flatten out the colors of the photo.
- Select the b channel and process it in the same way as the a channel was done. The amounts of Gaussian Blur to apply will not necessarily be the same for both of the channels.
- Optional - Add an appropriate amount of sharpening to the L channel.
- Convert the image back to the RGB color space.
I use this technique often to make monochrome photographs look more 'film-like'. With low noise photographs, however, the noise can be enhanced by overdoing this technique. Be sure to compare an affected photograph to one that has not been modified by this technique (easily done with the History palette.) If it seems o be over-done upon inspection - there may be a large scale undulations in the image - try it again using less blurring on each of the channels
Software for Noise Removal
Since noise removal is such an integral part of digital photograpic manipulation, there are several third party software packages that specialize in noise removal. I have used two such packages, Visual Infinity's Grain Surgery and PictureCode's Noise Ninja. Grain Surgery works as a plug-in for Photoshop, that is controlled by sliders acting on several parameters. With images with low to moderate noise, Grain surgery does a superb job of completely eliminating the noise without degrading the sharpness of the image. Images with more noise, particularly those shot at ISO 1600 or above, are not suitable for use with Grain Surgery. High ISO noise, especially noise found in shadow regions, takes on a rectangular shape that is emphasized by treatment with this program.
The newer program, Noise Ninja, does a much better job with high ISO noise, although it is not as customizable as is Grain Surgery. Noise Ninja runs much faster than does Grain Surgery and is easy to use. It is only available in beta form at the present, and sometimes places artifacts on an image. These show up as vertical or horizontal lines that can be cloned out with the rubber stamp brush in Photoshop, but hopefully this bug will be repaired in the final version, which should be released this summer.
I have not used Noise Ninja on any of my presented images, but have experimented with it. Once the final version is released I plan on adding it to my toolkit. Some other programs, Neat Image comes to mind, are available for the PC platform, but I have not used them.
Other Noise Reduction Techniques
Technique 3 - Curve Matching
This technique can be used to smooth out a fairly uniform region of a photograph, for example, the sky, where excess noise is very visible and objectionable. Since indiscriminate use of this technique can lead to loss of picture data, be sure to compare before and after samples carefully.
- Select an area to be smoothed, and feather (Select>Feather...) the selection 1-3 pixels to hide sharp edges.
- Call up the Curves dialog (Image>Adjustments>Curves...) and be sure to check te Preview box.
- Explore the selected area by dragging (moving the pointer with the mouse button down) the pointer (it will change into an eye-dropper) around. A small circular indicator will appear on the curve graph, indicating the brightness level of the pixels under the pointer. I this manner identify the range of brightness for the region to be smoothed.
- Bend the curve so that the portion of the curve that was identified inside the target region is flatter - more horizontal - than the unmanipulated curve. Use sharp bends, and try to keep the parts of the curve outside the target region at the same slope as before you made the changes, as shown here:
- Adjust the amount and width of the flattened area until the result is satisfactory then click Ok.
Technique 4 - Sharpen a Single Channel Only
If sharpening is applied, by sharpening only the smoothest channel, usually the green channel, the noise of the red and blue channels are not emphasized.
Technique 5 - Local Sharpening
Again, this technique is actually a modification of a subsequent sharpening step. By only sharpening the areas that require sharpened, the overall noise enhancement will be limited to those areas. Although the Sharpen Brush is a rather heavy handed tool, it can be moderated by turning down its strength in the options palette and by using a brush size that keeps the effect only on small areas that require sharpening.
If disconnected areas require sharpening, noise sharpening (and thus amplification) can be kept out of areas that don't need to be sharpened by using the history brush on those areas. Just be sure to set the history state to the step immediately prior to the sharpening step. This effect can be moderated by using the Fade (Edit>Fade History Brush ...) command.
Technique 6 - Down Sampling
by reducing the overall resolution of the photograph, even slightly, the size and thus the effect of noise grains are also reduced in size. Once the size of the noise is one pixel or less, it is barely noticeable. When used in conjunction with a/b blurring, a very slight image size reduction results in a very film like grain appearance.
- Open the image size dialog (Image>Image Size ...)
- Be sure that the Resample Image checkbox is checked, and that the method is set to Bicubic.
- Set the Width parameter to Percent
- Type a percentage that is an even sixteenth into the Width field according to this table:
15/16 = 93.75%
14/16 = 87.5%
13/16 = 81.25
12/16 = 75%
|
11/16 = 68.75%
10/16 = 62.5%
9/16 = 56.25
8/16 = 50%
|
7/16 = 43.75%
6/16 = 37.5%
5/16 = 31.25
4/16 = 25%
|
3/16 = 18.75%
2/16 = 12.5%
1/16 = 6.25
|
Images resizing causes blurring by changing the structure of edges as sown in this simulation:
Imagine a six by one pixel picture that is made of two white pixels, two black pixels followed by two more white pixels. The black region in the center is has very sharp edges:
W W B B W W
If this is then resized to five pixels wide, the central black region becomes three pixels wide, with only the center pixel completely black, with gray pixels along the edge:
W G B G W
and the image is less sharp.
Resizing by even sixteenths results in less loss of detail than resizing by an arbitrary amount.
Technique 7 - Ink Jet Printing
Printing the image with a sprayed ink (or pigment) type printer, including normal ink jet, Iris printers, Glicleè process as well as inexpensive dye sublimation printers, will print with reduced noise because of overlap of the sprayed regions. Noise that is obvious on a computer screen will be diminished when printed in this manner. This effect is most effective when printing onto paper with a textured surface. Some paper, such as Smooth Pearl surface from Ilford, smooths the noise while retaining fine detail.
Embrace the Noise
Less a technique and more of a philosophy, accepting the presence of digital noise can lead to creative use of the noise. With the techniques above, noise can be made colorless, smaller, sharper or partially removed. Or it can be left alone and seen as part of the image. I usually follow the Embrace the Noise aesthetic to some extent, but try to reduce or modify it to my own preference.
Sharpening
Digital photographs are made by projecting light onto a sensor array instead of a piece of photo-reactive film. In order to produce a color image from the signal sent from a sensor, the sensor is covered with at least two filters. The first filter is a Bayer color filter, that alternates red, green, blue, green color filters over the individual sensing elements. In this way, the exact color of any group of pixels can be determined. The use of a Bayer filter, although widespread in digital imaging technology, can lead to color anomolies, particularly Moire pattern formation. In order to minimize this effect, a second filter, an anti-aliasing filter is placed immediately in front of the colored Bayer filter, so that light from one pixel is shared with the pixels surrounding it.
These intervening filters cause the image that reaches the sensing elements to be somewhat softened. Most digital cameras contain software that sorts out the softness and restores the sharpness of the lens. However, the amount of sharpening needed to do this is dependent on several factors, such as the brightness of the scene, the content of the scene and the particular optics of the lens. Since the manufacturer cannot know exactly how much sharpening to apply, the in camera sharpening algorithms are a compromise. Often the sharpening the camera applies is too much, resulting in halos and small 'burnt out' points, or the sharpening may be not enough, leaving the photograph soft in appearance.
Professional digital cameras (such as the Nikon D100 used for this paper) have the option to turn off in camera sharpening so that the correct amount of sharpening can be done using software after he photograph has been taken. This allows the use of different sharpening techniques, that can be tailored to the individual photograph, and the taste of the photographer.
In addition to restoring the optical qualities of the lens system used, after-the-fact sharpening has a limited use in increasing depth of field, and correcting out of focus images. Sharpening is not a panacea, however. Very little can be salvaged from an out of focus photograph by applying any sharpening technique.
Digital noise and sharpening are related. If a noisy picture is sharpened, the noise will also be sharpened, and will become more obvious. Conversely, if the noise is aggressively removed from a photo, the sharpness will suffer. At the time a photograph is taken, if a low ISO speed is used to minimize noise, motion blur may result, while high ISO speeds that yield short exposure times create noise. The diagram below shows the interconnection between noise and sharpness. The vertical axis represents the noise and the horizontal axis is the sharpness. Unfortunately cameras do not come with controls that can affect these parameters independently. The blue line represents the controls available at the time of taking the picture, ISO speed, the white balance and the exposure. The red line diagrams the range of post-processing control the photographer has such as noise reduction and sharpening:
Sharpening Tools - USM
The primary sharpening tool used by digital photographers is the Unsharp Mask (Filters>Sharpen>Unsharp Mask...), often referred to as USM. This basic tool can be applied in many ways, from a simple overall shot, to complex layer and channel based manipulations. Typically sharpening techniques are applied after all the other processing steps are taken. This way, subsequent changes to the photograph will not adversely change the sharpening.
USM is applied by adjusting three parameters, the amount, the radius and the threshold. The amount determines the strength of the effect, more amount means more sharpening is seen. Adjusting the amount changes the contrast difference given to adjacent pixels, adjusting the radius changes how far the contrast adjustment will go from each pixel. Threshold will limit the low end, so that sharpening of noise can be limited. Unless the contrast between adjacent pixels is greater than the threshold, the pixel will not be sharpened. This sounds more complicated than it really is, and the process becomes intuitive with a bit of practice.
Technique 8 - Standard USM
Once the picture has been prepared for web or print publication, it is customary to apply standard USM. The actual amounts are determined by trial and error. When sharpening for print, typical parameters are amount:100, radius:0.8 and threshold: 4. When sharpening for web or computer monitor viewing, the parameters are different. Amount: 400, radius:0.2 and threshold: 2 are typical.
When a high contrast edge is sharpened, a lighter region is created along the light edge and a dark region on the other side of the edge. This is called a halo, because it appears to be a glow coming from the object. Because of the halo, there is a limit to the amount and radius that can be applied.
- Choose Unsharp Mask from the Filter menu (Filters>Sharpen>Unsharp Mask...) to obtain the USM dialog.
- Be sure the Preview box is checked. Even with the Preview off, a section of the image is shown (initially) at full size inside the dialog.
- Adjust the sliders to the appropriate amounts depending on the application.
- When you are happy with the result, click OK.
Note that USM sharpening data is written in shorthand as (amount:radius:threshold).
Technique 9 - Darken Layer USM
To sharpen a photograph, without creating a bright halo, this technique may be used instead. In fact, since developing this technique, I use it in all cases where I would have used standard USM. This has become my standard way to sharpen photographs.
- Duplicate the image into a second layer (Layer>New>Layer via Copy)
- From the style menu at the top of the Layers palette (Window>Layers), set the style of the new layer to darken.
- Apply USM to the layer. Larger values may be used for the amount and radius, and often, the threshold can be set to zero. Typical values for images to be printed are (200:1:2), but values as high as (500:1.2:0) have been used with no detectable halo formation and a very natural appearance. A typical web dark layer sharpening would be 500:0.3:0.
- Adjsut the opacity of the layer from the slider in the Layers palette to moderate the effect.
- Flatten the image (Layer>Flatten Image).
- Optionally, apply a small amount of standard sharpening (80:0.2:0) to the flattened image.
The main drawback of this method is, when sharpening images with noise or dust spots (dust on the sensor is a notable problem for interchangeable lens digital cameras), black dots can result. If the image is very noisy, this sharpening method is not advised.
Other USM recipes
Technique 10 - Red Channel Sharpening
This method can be used with either standard sharpening or darken layer sharpening. since skin tones contain a large amount of red, sharpening only the red channel (which will be the lightest channel) will keep the skin smooth in appearance and will not sharpen wrinkles and blemishes. This technique works equally well with light and dark skin.
- For darken layer red channel sharpening, create a darken layer as above. For standard sharpening, omit this step.
- From the Channels Palette (Window>Channels), select the red channel.
- Apply sharpening to the red channel with the USM dialog. About 25% more sharpening can be done this way than would be used if the entire RGB image were sharpened.
- Use the fade command (Edit>Fade Unsharp Mask ...) to reduce the effect if you went to far.
- Re-select the RGB channel from the Channels Palette.
Technique 11 - L channel sharpening
The lightness channel of the Lab color space can be sharpened independently. Bu doing so, color fringing, which can be seen if highly colored adjacent areas are sharpened by standard or darken layer sharpening.
- Convert the image to Lab color (Image>Mode>Lab Color).
- From the Channels Palette (Window>Channels), select the L channel.
- Apply standard sharpening to the L channel with the USM dialog.
- Use the fade command (Edit>Fade Unsharp Mask ...) to reduce the effect if you went to far.
- Convert the image back to RGB color (Image>Mode>RGB).
Technique 12 - Double Sharpening
For images to be shown on a computer screen, instead of printed, standard or darken layer sharpening can be applied with a radius of 0.2 or less twice. If this causes too much sharpening withthe second application, the effect can be faded (Edit>Fade ...) to get the correct amount of sharpening.
Technique 13 - High Radius Sharpening
This technique, not so much sharpens an image, but it increases the contrast is what may make repeating patterns standout as if they were sharpened. If standard or dark layer USM were applied, aliasing or moire may result in these cases. By using a low amount and high radius, the picture will appear sharper than if nothing is done, however if correctly applied, the effect will be subtle.
To apply high radius sharpening, use a standard USM with an amount of 40 to 60 and a radius of 20. The threshold does not matter, so leave it at zero. Special effects can be done by increasing either the amount (up to 200) or the radius, up to the maximum of the slider, 250. This will almost always result in extremely high contrast.
Non-USM Sharpening Techniques
Technique 14 - Contrast Increase
An effect similar to high radius sharpening can be obtained by increasing the contrast of the image with either the levels (Image>Adjustment>Levels ...) or curves (Image>Adjustment>Curves ...) dialogs.
Technique 15 - High Pass Sharpening
This is a popular technique that I never use.
- Create a duplicate layer with the New Layer via Copy command as described above.
- Run the High Pass filter (Filter>Other>High Pass ...) with a radius of 10.
- Zoom in to 100% view so the next step can be applied to suit the individual image.
- Set the Layer Style to Hard Light from the style menu in the Layers palette.
- Change the opacity of the Layer to 10-50% with the Opacity slider in the Layers palette to get the effect you want.
- Flatten the layer (Layer>Flatten Image).
When NOT to use sharpening
Sometimes it is not advisable to sharpen digital photographs. High noise photos, typically available light shots, degrade when sharpened, as do highly compressed JPEG files. And for very small images (less than 25 pixels wide), sharpening can obscure details. Also, with some documentary photos such as forensic evidence, sharpening may change the information contained in the photo and should not be used. Finally, if a soft focus effect is sought, sharpening can ruin the effect and should be left out to the work flow.
Monochrome Conversion
Much of my work is done in monochrome, that is black and white or a toned version of black and white. Since the image from digital cameras is inherently color, further processing must be done to convert the color picture to a monochrome one. Not all photographs lend themselves to a monochrome treatment and monochrome images must be previsualized to determine what the final image should look like. Fortunately the tools in Photoshop make image selection and re-visualization a trial and error process, where on screen feedback can be used to arrive at an optimal monochrome rendition.
There are four tools. Well, there are even more, for example Calculations (Image>Calculations ...) but even my math minor leaves this one counterintuitive, so I never use it. If one of the simpler tools gets you the picture you need, great! Use it and you will save a lot of time. However, I'm going to cut right to it, the Channel Mixer (Image>Adjustments>Channel Mixer ...) is the most versatile tool for rendering black and white images, ever. But let's start with the simple tools.
Technique 16 - Greyscale Mode
The simplest way to convert an image to black and white is to convert it to grayscale color mode (Image>Mode>Grayscale). This usually results in a flat image that can be 'punched up' with Levels or Curves adjustment. This method is very quick, consisting of only a single menu command.
Technique 17 - L Channel Extraction
Occasionally the L channel of Lab color space contains an interesting tonal range and is usually the sharpest channel within all of Photoshop's range.
- Convert the image to Lab color.
- Choose the L channel from the Channels palette.
- Select the entire image (Select>All), or a crop you wish to work with, and copy (Edit>Copy) the selected image.
- Create a new document the size of the copied image (File>New ...). The size already inserted in the New Document dialog will be the size of the copied data.
- Paste (Edit>Paste) the copied data into the new document, and optionally flatten (Layer>Flatten Image) the image.
This image is once again in RGB mode, but all three channels are the same. Changes to tones may be made with the Levels and Curves tools. This is also a very fast way to a monochrome image. Often batch processes use this method because the resulting image often is more pleasing then straight conversion to grayscale mode.
Technique 18 - Color Channel Extraction
Some subjects lend themselves well to primary color filtering. For example, a deep red filter gives a good look to a clouded sky, and swarthy male skin tones respond well to a green filter. In order to quickly perform a primary color filtration effect, a single color channel can be isolated in the same way the L channel was gotten to make a black and white image.
- If it is not already, convert the image to RGB color.
- Open the Channels palette. Select in turn, the R(ed), G(reen), and B(lue) channels. Look at how the different color tones are rendered in each of the channels.
- If you see a channel that has the qualities you want, choose that channel.
- Select the entire image (Select>All), or a crop you wish to work with, and copy (Edit>Copy) the selected image.
- Create a new document the size of the copied image (File>New ...). The size already inserted in the New Document dialog will be the size of the copied data.
- Paste (Edit>Paste) the copied data into the new document, and optionally flatten (Layer>Flatten Image) the image.
This image can now be adjusted with the Levels or Curves dialogs. If none of the color channels look just like you want, it is time to try mixing the channels together.
Technique 19 - Channel Mixing
In order to get the fullest control over the tones, as well as be able to simulate the effect of any color filter, Channel Mixer is your tool. Channel mixing can also be used to help to reduce digital noise or even enhance noise. This tool works by blending the three color channels into a single black and white channel. The amounts of each color to be blended are set by slider that ranges from -200% to 200%. A good starting point is with equal amounts of each channel: 34% Red, 34% Green and 34% Blue. Since this adds up to 102% take off the two percent by setting the Constant slider to -2. Setting the constant to a negative number increases the overall contrast, while positive constants lower the image's contrast.
The actual values that will make the final image can be widely varied. Since you examined each color channel in technique 18, above, you will have a feel for which channels will lighten which parts of the image. Using this knowledge as well as the full size live preview, adjust each of the sliders until the perfect blend is reached.
To do Channel Mixing
- If it is not already, convert the image to RGB color.
- Select the Channel Mixer tool. (Image>Adjustments>Channel Mixer ...)
- Check the Monochrome checkbox at the lower left corner of the Channel Mixer dialog.
- Slide each of the color channel sliders to various positions and evaluate the image.
- Once you are happy with the image, click Ok and your image will be converted to the chosen mix of channels. It is still an RGB image, but each of the color channels is the same grayscale image as is the RGB image.
It is interesting to discover what the effect of negative influences can be on images. The blue channel can be subtracted in moderate amounts (up to -20) to dramatically darken the sky and water. When the blue channel is given a very low value, say -180 and the green channel is given a correspondingly high value, a pseudo-infrared rendition is achieved, with black skys and very light foliage. A negative red and increased blue mixture can radically darken skin tones. Because the image is updated as the sliders are moved, the proper mix can quickly be arrived at.
Because some channels are noisier than others, typically the blue channel is the worst and the green channel is the cleanest, the noise present in the final image can be controlled by channel mixing. Beware, however that taking any channel to either positive or negative extremes will result in noise enhancement, so the best channel mixes from the noise point of view, are those done with moderate differences between channels. Channel mixing data is written in shorthand similar to the way USM data is. It is given as slider values in this order: (red:green:blue:constant).
Colorizing Monochrome Images
With film prints, the job was often not completed until the print was chemically toned. These processes either replaced or compounded the silver of the emulsion to make a substance with a different reflective color. One very popular toning process is sepia toning, in fact some cameras come with a built in sepia mode.
Of course, when I sepia toned prints, I did it under light and monitored he color, and could moderate the process by pulling a print from the toning bath prematurely, to obtain richer coloration of the print. The next two techniques are different ways to expand the colors in a monochrome image without going all the way and making it a full color image.
Technique 20 - Duotone
The Duotone option in photoshop allows you to pick other base colors then pure black and pure white. Variations of the Duotone process are Tritone and Quadratone. To simulate pulling a print from a toner bath early, I use the Tritone process to add a 'gray of a different color'
- Convert a black and white image to 8 bit (Image>Mode>8 Bits/Channel) if it isn't already, then to Grayscale mode (Image>Mode>Grayscale). If a dialog asks if yu want to disgard the color information, tell it yes.
- Then convert the image to Duotone mode (Image>Mode>Duotone ...). Check the Preview box.
- From the Type menu at the top of the dialog, select Tritone and the first three color descriptor lines will activate.
- Create names for each of the three colors. I call then offblack, offgray and offwhite.
- I use the following values for these colors, but you should experiment with the colors and make your own descriptions of the three shades.
- Offblack - R:G:B = 3:3:2
- Offgray - R:G:B = 232:227:95
- Offwhite - R:G:B - 255:255:244
- Click Ok. The image may be re-converted to RGB mode if desired, or it may be printed directly from the Tritone rendition.
Technique 21 - Layer Mixing
This last technique combines channel mixing or extraction with a full color rendition of the photograph to yeild softened colors, or with appropriate masking, selective coloring. It is especially effective with high key images, where most of the tones are brighter than middle gray.
- Create a duplicate layer of a color image (Layer>New>Layer via Copy).
- Convert the duplicate image to a grayscale image by channel mixing or other method.
- Use the opacity slider of the Layers palette to allow a portion of the color image to show through. Adjust the opacity until you like the image.
- When you are satisfied with the image, optionally flatten (Layer>Flatten Image) the image.
If you want to let part of the underlying color image to come through full strength, use the eraser tool to remove the monochromatic layer where you want to see the color image.
There are many, many more techniques, such as burning and dodging, layers and curves, masking and more complex combinations of the presented techniques. They are often one time processes and can be found by trial and error. I don't consider an image to be finished until I have completely exhausted my skills at trying different aspects of post processing. However, all the post processing I do is within the realm of what can be done to the image as a whole, or by localized masking or painting effects. Once a portion of an image is displaced or external images are added or elements removed, the image ceases to be a true photographic image.
I consider the techniques presented in this paper to be a good set of photographic manipulations. These manipulations, while sometimes radically changing the image, keep the image a photograph, where the images were originally formed by an optical system. I believe that it is important to preserve the art of photography from being combined with digital art. As I stated above, digital art is very useful and in many cases (such as those exhibited in this class) very beautiful, but they are more akin to painting than photography. Or perhaps to photomontage.
Reading List
- Reznick, Ron - Digital Photography, Acquisition and Processing Techniques, RR Design, 2004
- Kelby, Scott - The Photoshop CS Book for Digital Photographers, New Riders Publishing, 2003
- Adobe - Adobe Photoshop Users Guide, Adobe Systems, Inc., 2003
- McClelland, Deke and Katrin Eismann - Real World Digital Photogrphy, Industrial-Strength Techniques, Peachpit Press, 1999
- Shekter, Louise and David Fono - Grain Surgery 2, Intelligent Noise Manipulation, Visual Infinity, Inc., 2002
- Clarke, Graham - The Photograph, Oxford University Press, 1997
- National Association of Photoshop Professionals - Photoshop User (magazine), NAPP, 2004
- Element K Journals - Inside Photoshop (magazine), Element K Journals, 2004
- Askey, Phil - Digital Photography Review (online discussion forum), DP Review, 2004
|
