Glossary

Additive color mixing is the process of combining light of two different colors. For example, adding red and green light makes yellow.

An anaglyph is a pair of black and white stereo images prepared for viewing with special glasses that have a red filter for one eye and a cyan filter for the other eye. One image is printed in red and can be seen only by the eye with the red filter; the other image is printed in cyan and can be seen only by the eye with the cyan filter.

Smoothing out the jagged appearance of edges in an image by blending background color into the boundary pixels.

The ratio of width to height of a rectangle; an aspect ratio of 1.0 corresponds to a square.

An image made up of pixels of only two colors (normally black and white). Each pixel in a binary image occupies a single bit of storage; thus 8 pixels can be stored in a single byte of memory.

A bit is the smallest piece of information that can be represented in a computer.

The BMP or bitmap file format is the standard method within Microsoft Windows for representing image data. BMP files are normally uncompressed.

A byte is a unit of computer memory consisting of eight bits. When interpreted as a number, a byte can represent a range of values from 0 and 255.

A cast is an excess of one color in an image. For example, an image that is too red is said to have a red cast. Color casts can occur in a digital image as a result of degradation of film over time, reciprocity failure, problems in film handling or processing, use of incorrect lighting for a given film type, improper scanning, or many other causes.

Charged Coupled Device. A specialized light-sensitive integrated circuit that can capture a digital image directly. CCDs are commonly used in camcorders, digital cameras, and scanners to convert images into digital form.

The part of a video signal that contains color information (as opposed to Luminance which is the black and white component of the video signal).

The Commission Internationale de l'Elcairage. An international standards organization.

Cloning is a technique for freehand retouching of digital images. Cloning involves copying small patches of one image over other parts of the same or a different image.

For example, to remove a dust speck from a photography of the sky, you could copy a small patch of the sky from right next to the dust speck over the speck to remove it. More ambitious cloning techniques can be used to repair damaged images, eliminate larger objects from the background, or to copy image elements from one place to another.

CMYK (Cyan, Magenta, Yellow, Black) is the color space used by four-color printing presses to reproduce color photographs in newspapers, magazines, and other printed materials. The fourth color, black, is required to create rich blacks since the printing inks are slightly impure and do not combine to create a good black.

Color depth (sometimes also called bit depth) refers to the number of bits used to represent each pixel in a digital image. The common values for color depth and the number of possible colors each pixel can represent are given in the following table:

bits 1	possible colors 2
4	16
8	256
15	32,768
16	65,536
24	16,777,216

The more bits are used by each pixel, the more closely a digital image can approximate a photograph (and the more memory the image occupies).

A color gamut is the range of colors that can be display, printed, or captured by a particular monitor, film, printer, scanner or other device that deals with color images. The larger the color gamut, the more closely a device can approximate the full range of colors visible to the human eye.

Since all colors are generated by combining primary colors, the color gamut of a device is mostly determined by the primary colors in use. For a monitor, the primary colors are determined by the physical properties of the red, green, and blue phosphors embedded in the face of its picture tube (CRT). For film, the color gamut depends on the dyes contained in the different layers of the emulsion. For printers, the color gamut depends on the colors of the inks that are combined and deposited on the paper.

Color management is the process of making colors look the same on two or more different color image input or output devices. For example, color management can be used to make the prints created by a digital color printer match the same images displayed on a computer screen.

Accurate color management requires that all devices in the system be calibrated (adjusted to standard settings) and characterized (have a systematically measured color response). Then sophisticated software is used to alter the colors so they will match as closely as possible. Perfect color matching is usually impossible due to differences in the characteristics of the devices.

Every visible color can be specified by a set of three numbers. There are many different schemes for classifying all the possible colors. The three schemes that Picture Window uses are RGB, HSV, and HSL.

The RGB color space breaks colors down according to the amount of each of the three primary colors they contain. Red, green, and blue are called additive primary colors because light of these colors can be mixed in different proportions to make any other color. Most computer equipment is based on the RGB color space and it is therefore the easiest for a computer to work with. Unfortunately, humans do not naturally interpret colors according to this system. The space of all colors in the RGB system is a cube.

The HSV color space categorizes colors by their hue, saturation, and brightness (sometimes called value).

Hue corresponds to the colors around the outside of the color wheel. Hues run from red to yellow to green to cyan to blue to magenta, and back to red.

Saturation refers to the intensity of a color. A fully saturated color is very pure and deep; as you reduce their saturation, colors become progressively more washed out until at a zero saturation they become shades of gray.

Value refers to the overall brightness of a color. All colors with a value of zero, regardless of their hue or saturation, are black. As you increase value, colors become progressively brighter. The brightest, most saturated colors are those with the maximum value.

The space of all colors in the HSV system can be thought of as a hexagonal cone with black at the vertex. The following pictures are cross-sections through the hexcone at 50%, and 100% value. At 0%, it converges to a black point.

The HSL color space categorizes colors by their hue, saturation, and lightness. As in the HSV color space, Hue corresponds to the colors around the outside of the color wheel and Saturation refers to the intensity of a color.

The term Lightness refers to the brightness of a color. All colors with a lightness of zero, regardless of their hue or saturation, are black. As you increase lightness, colors become progressively brighter. The brightest, most saturated colors are those with a lightness of 50%. As lightness increases toward 100%, colors become increasingly whiter. All colors with a 100% lightness are white.

The space of all colors in the HSL system can be thought of as a double-ended hexagonal cone with black at one vertex and white at the other. The following pictures are coruscations through the double hexcone at 25%, 50%, and 75% lightness. At 0%, it converges to a black point and at 100% it converges to a white point.

Color temperature is a term generally used to describe the color of a light source or the maximum white that a monitor can produce. For accurate color matching between an image viewed on a monitor and a print or transparency, the color temperature (white point) of the monitor must be adjusted to match the color temperature of the light used to view the print or transparency.

Color temperatures are normally measured in degrees based on the black body radiation curve which describes the color change of objects when they are heated. Low temperature light sources light incandescent and quartz halogen lamps (around 3000 degrees) are somewhat orange. Standard daylight is around 5500 degrees. A typical uncorrected color monitor is rather blue (around 9300 degrees).

The complement of a color is the color you must add to it to get a neutral gray. You can find the complement of a color by locating the point on the color wheel an equal distance on the other side of the center. The following primary colors are complements of each other:

Red	Cyan
Green	Magenta
Blue	Yellow

There are two kinds of compression: lossless and lossy. Lossless compression techniques reduce the storage space required by image (or other) files so that when the files are decompressed, the original file is reproduced with no loss of information. Lossy compression reduces image size with some loss of information.

Lossless compression techniques are quite limited in their ability to reduce image sizes. For typical photographic images, compression factors of more than two are difficult to achieve.

Lossy compression techniques such as JPEG let you trade off image quality against amount of compression; JPEG compression can compress typical images by factors of five to ten while retaining very high image quality.

In a continuous tone image, each pixel can reproduce any visible shade of gray or color. This is in contrast to bilevel or halftone images where gray levels or colors are reproduced as clusters of pixels selected from a limited palette of colors.

Convolution is a mathematical process for modifying a signal or image by applying a so-called kernel which is really just a set of weights. Each pixel in the output image is computed by taking the weighted average of the neighboring pixels in the input image. By using different kernels, convolution can be used to perform many different operations such as blurring, sharpening, and edge detection.

A densitometer is a device for measuring the density of a piece of film or a print. Densitometers contain a light source and a light detector, similar to a light meter. Densitometers that measure film density are called transmission densitometers because they measure how much light passes through the film at a particular location. Densitometers that measure prints are called reflection densitometers because they measure how much light is reflected off the surface of the print at a particular location.

Density is a measure of how much light can pass through a piece of film (or how much light is reflected from a print). Density is measured using a logarithmic scale; a density of zero corresponds to clear film (or white paper) and the darker the image, the higher the density.

Density, D, is computed as D = -log(T), where T is the fraction of light transmitted by a piece of film (or reflected from a print). Thus, a piece of film that transmits 10% of the light that falls on it has a density of 1.0; 1% transmission gives a density of 2.0, and so on.

Color images are made up of color pixels, each of which is represented internally in the computer as three bytes (24 bits) defining its red, green, and blue values. This means a color image 1000 pixels wide and 1000 pixels high occupies 3MB of memory.

Black and white images are made up of gray pixels each of which is represented internally as a single byte (8 bits) whose value describes its brightness. Thus, a black and white image uses three times less memory than a color image of the same dimensions.

Binary images are made up of pixels that are either black or white; each binary pixel is represented as a single bit. Binary images use eight times less memory than black and white images. While Picture Window supports importing and using binary images, they are not very useful for representing photographic images.

The DNG (Digital Negative Format) file format was created by Adobe for storing RAW files in a uniform format.

The extent of the range of brightness levels in an image between the darkest to the lightest pixel.

Gamma is a technical term that refers to the curve that relates the RGB component values of a pixel to the actual intensity, as measured off the face of the monitor by a photometer. Since the human visual response is not linear, it is desirable to set the gamma curve to roughly match that of the eye. This is approximated by using a mathematical relationship called a power law where gamma is the exponent. The standard gamma value used in television sets, video recording, and most PC monitors is approximately 2.2.

Gaussian blurring is a technique for blurring an image that closely models what happens in an optical system when an image is defocused. Gaussian blur, as normally implemented in image editing programs, can create very strong blurring by choosing a large radius setting.

The GIF (Graphics Interchange Format) file format was created by CompuServe for distributing images on-line. The GIF file format supports a maximum of 256 colors per image and is therefore not well suited to representing color photographic images accurately.

The process of representing continuous tone images using a small palette of colors, usually performed in connection with printing. Black and white images are typically halftoned to a pattern of black and white dots; Color images are usually halftoned using cyan, magenta, yellow, and black dots. Different shades of gray or color are created by placing clusters of dots together. There are many different styles of halftoning which have different visual appearances.

The HEIF (High Efficiency Image Format) file format was created by Apple for storing compressed iPhone images.

The histogram of an image is a graphical representation of how many times each possible brightness level in the image occurs. It is computed by looking at each pixel of the image, computing its brightness, and then counting how many times that brightness value occurs.

The shape and smoothness of a histogram can be used to gauge the overall quality of an image as well as its brightness, contrast, and dynamic range. A predominantly dark image will have a histogram that peaks at lower brightness values; a predominantly light image will peak at higher values. An image with a full spread of tonal values will have a histogram that extends all the way from one end of the brightness scale to the other.

ICC stands for International Color Consortium, a group of companies active in the digital imaging industry. They have created the ICC Profile standard for characterizing different color output devices. These profiles facilitate the reproduction of color images in a consistent manner regardless of the device being used.

Interpolation is a technique for computing data between measured values. When applied to digital images, interpolation can be used to resample an image from one resolution to another. While interpolation does not add any new information to the image, it can be used to change the size of the image to match a specific printer or other output device.

JPEG stands for Joint Photographic Experts Group, an association that has defined a standard image compression technique and file format that is supported by many different image editing programs. JPEG allows for adjustable levels of compression -- image quality goes down as you increase compression. Amazingly, the JPEG compression technique can reduce the size of an image by a factor or 5 to 10 with little or no visible image quality degradation.

JPEG compression works by breaking up your image into 8x8 pixel squares, analyzing each one using a mathematical technique called a Discrete Cosine Transform, and discarding frequency components that fall below a certain threshold based on the compression level you select. The remaining data are compressed using Huffman coding and stored in the output file. JPEG files are decompressed using a similar technique.

Landscape orientation or landscape mode refers to an orientation of an image where the width is greater than the height.

The black and white component of a video signal or the perceived brightness of a pixel. The luminance of a pixel is computed as a weighted average of its red, green, and blue components, based on the relative sensitivity of the eye to each color:

Color	Weight
Red	30%
Green	59%
Blue	11%

A mask is a black and white image used to isolate part of another image of the same dimensions. Masks can be used to protect part of an image from the effects of an operation.

Minimizing a window means reducing its to an icon that can be expanded later by double clicking on it.

An interference pattern caused by the interaction of two different patterns of dots or lines. Moiré patterns commonly occur when you attempt to photograph images using a digital camera. In this case, the moiré pattern results from interference between the regularly spaced patterns in the scene and the sensor matrix.

Monochrome simply means black and white. For example, a monochrome monitor is one that can only display different shades of gray. Similarly, a monochrome image is black and white only.

The term pixel is derived from Picture Element. A pixel is a tiny dot of solid color; photographic images are made up of thousands or even millions of these dots arranged in evenly spaced rows and columns. Whether being displayed on your computer screen or printed on film or paper, digital images are made up the same rows and columns of pixels. By zooming an image in to a high magnification factor, you can see its individual pixels.

The PNG (Portable Network Graphics) file format was introduced as a replacement for the GIF file format when it turned out that GIF incorporates a form of file compression patented by Sperry Univac. The PNG format is actually a considerable improvement on GIF but its popularity remains limited.

Portrait orientation or landscape mode refers to an orientation of an image where the height is greater than the width.

Posterization is the process of representing a continuous tone image using a small number of different grays or colors. For example, a very high contrast image is a 2-color posterized representation of an original. Less radical forms of posterization use intermediate grays or colors. Posterization refers to the process of creating images for posters that are printed using a small number of different inks.

PostScript is a language developed by Adobe Systems for using in printing documents. It supports the use of mixed fonts, images, rotated text, drawings, and many other advanced features. Viewing or printing PostScript requires software called a RIP (raster image processor) that can convert the PostScript data to a raster image.

RAM stands for Random Access Memory; it refers to the main memory installed in your computer. This memory is used to run the operating system, run programs, and hold the data you are currently working on. The more RAM your computer has, the more programs and information it can access quickly. RAM is normally measured in megabytes (MB) a megabyte is roughly one million bytes.

RAW file format is a generic term for files saved by various cameras. The format specification varies from manufacturer to manufacturer and from camera model to camera model. The advantage of RAW files is that they capture all the information from the camera. The disadvantage is that they need to be converted to one of the standard image formats before they can be edited. RAW converter software interpolates the image to remove the effects of the Bayer color filter in front of the sensor, applies the necessary gamma correction, performs noise reduction, removes camera and lens distortion, chromatic aberration and light falloff, and lets you adjust the exposure to compensate for some level of overexposure or underexposure.

RGB stands for Red, Green, and Blue, the additive primary colors. Any color can be created by adding red, green, and blue light in the correct proportions. Computer monitors use this technique to display images. The red, green, and blue component values define a color in the RGB color space.

Red eye is a red disk that commonly appears in the center of the eyes in certain flash photographs:

This phenomenon is caused by using an on-camera flash to photograph a subject looking directly into the camera. The red disk in the pupil of the eye is actually a reflection of the flash from the back of the eye. Some cameras attempt to reduce this reflection by pre-firing the flash just strongly enough to get the subject's pupils to contract before taking the photograph.

Resampling is the process of increasing or decreasing the number of pixels in an image.

The resolution of a digital image refers to the amount of detail it can hold. This in turn depends on the number of rows and columns of pixels in the image. The more pixels an image contains, the higher its resolution and the more detail it can reproduced. This concept is similar to that of film grain where finer grained films can capture more image detail.

Sometimes resolution is also used to indicate how densely pixels are packed within an image of a fixed size. In this case, resolution is normally given in dots per inch (dpi). This indicates how closely spaced the pixels are in the image. The more dots per inch, the higher the resolution of an image. Images displayed on a computer screen typically have a resolution of 60 to 100 dpi. An image printed on paper using a continuous tone printer at 200 to 300 dpi approaches photographic quality. When printing the same image using a conventional printing press, a resolution of 1200 to 2400 dpi is required to obtain roughly equivalent results.

Scrolling is the process of moving a viewing window across a document or image that is too large to fit in the window.

Subtractive color mixing is the process of combining inks or pigments of two different colors. For example, adding magenta and yellow makes red.

Cyan (a blue-green), Magenta (a red-blue), and Yellow are the subtractive primary colors you were taught about in grade school. You can mix any other color by combining paints or inks of these colors. The subtractive primary colors are just the complements of the additive primary colors, red, green, and blue.

An area on your hard disk that is used by a virtual memory system to store data or programs that are not currently being used.

TIFF stands for Tagged Image File Format -- the most widely supported image file format. If you want to create image files that can be used on Macintosh or UNIX computers as well as PCs, TIFF is usually the best choice. TIFF files may or may not be compressed.

The Undo command in most programs is used to reverse the effect of the most recent command. Some programs permit the user to undo multiple actions.

Unsharp masking is a versatile sharpening technique that is related to Gaussian Blur. Unsharp masking works by first blurring an image and then comparing the original version to the blurred version. Where the difference exceeds a specified threshold value, the image is sharpened by exaggerating the difference between the original and blurred versions.

Virtual memory is a technique for extending the memory capacity of a computer. It involves writing unused portions of memory to a swap file on your hard disk and reading it back later when it is needed. Your computer first uses up all the normal memory (RAM) in your system; when that runs out, it starts to swap inactive data or programs to the swap file. If the data that is swapped out is needed at a later time, it is swapped back in.

The larger the swap file, the more data you can handle without getting Out of Memory errors. However, if you do not have enough normal memory to hold the active information you are working on, something called thrashing may occur which is a condition where memory is continually being swapped in and out and the computer grinds to a crawl. The more RAM you have, the larger the amount of information you can work on without thrashing.

The white point of a monitor refers to the color of the brightest white it can generate. If uncorrected, most computer monitors have a white point that is somewhat blue compared to normal daylight or artificial illumination. To achieve an accurate match between images on a monitor and prints, the color temperature of the monitor must be matched to that of the illumination used to view the prints.

Zooming is the process of altering the magnification factor for a viewing window for a document or image. Zooming in increases the magnification; zooming out reduces the magnification.

The degree of magnification (or reduction) is generally given as a zoom factor. When an image is displayed at a zoom factor of 1:1 each pixel in the image is displayed as a single pixel on the screen. At 2:1, each pixel in the image is enlarged to a 2x2 square of pixels, thus magnifying it by a factor of two; at 1:2, every other pixel is displayed, causing the image to shrink in size by a factor of 2.