Digital Media The material in this handout is excerpted from Digital Media Curriculum Primer a work written by Dr. Yue-Ling Wong (ylwong@wfu.edu), Department of Computer Science and Department of Art, Wake Forest University and supported by the National Science Foundation under Grant No. DUE-0127280. http://digitalmedia.wfu.edu/project/digital-media-curriculum-development/ This handout was prepared for students in MMP320 Multimedia Networks at Borough of Manhattan Community College, City University of New York as part of a curriculum redesign project supported by National Science Foundation Grant No. DUE NSF-0511209, Co PI s Christopher Stein (cstein@bmcc.cuny.edu) and Jody Culkin (jculkin@bmcc.cuny.edu) http://teachingmultimedia.net
Capturing and Editing Digital Images * Methods of capturing digital images- scanning and digital photography Scanners: common types Flatbed- most common and versatile. Consists of glass plate, motorized scan head, consisting of light source and array of sensors. Sheet-fed- smaller portable scanner, fixed scan head, cannot scan books or thicker documents. Hand-held- also portable. Often rely on user moving scan head, making it difficult to get consistent results. Most commonly used for text. Drum- very high resolution. Used for print publishing. Scanners can handle film negatives and slides as well as flat documents. There are specialized scanners for handling these types of media, and some flatbed scanners have adapters to scan film. TWAIN: a standard or specification for the interface between image capture devices such as scanners and digital cameras and image software that handles these types of files. Capturing Images by Scanning: Optical resolution versus interpolated resolution- optical resolution is dependent on the hardware, the actual numbers of light sensors that the scanner has to capture the image information. Interpolation increases resolution by adding pixels with software to those captured by the hardware. Generally scanner resolution is referred to in dpi (dots per inch). We re returning to the concept of sampling. The number of sensors available in this single row corresponds to the sampling rate in the x-direction. The discrete stepwise movement of the scan head is related to the sampling rate in the y-direction. The sensor corresponds to the dot in the unit dpi. Each sample results in a pixel of the scanned image. For example, many flatbed scanners now have an optical resolution of at least 2400 dpi on each direction. If you scan a 1-inch-by-1-inch picture at this scanning resolution, you will get an image of 2400 x 2400 pixels. Determining Scanning Resolution Before making a scan, you must consider the requirements of the final output. What will the scanned image be used for? Is the scanned image intended for the Web or printing? If it is intended for printing, then what will the physical dimensions of the print be and what are the requirements of the printing device? If the final image is for the Web, then what are the required pixel dimensions of the image? You should scan at a high enough resolution (in dots per inch (dpi)) to generate enough pixels in each dimension in order to produce an image at the expected print size. If you are uncertain, it is better to capture a larger amount of pixels. Your image quality will
suffer if you have to add pixels by interpolation afterwards. It is possible to reduce the number of pixels in imaging editing software without sacrificing image quality. To determine the correct number of pixels, remember you must consider the pixel dimensions as well as the ppi (pixels per inch). The inch in the pixel per inch (ppi) and dots per inch (dpi) is in linear inches. Remember, just like any digital images, scanned images do not possess any inherent physical dimensions. The physical dimensions materialize when the image is printed out, defined by both the image's pixel dimensions and the printing ppi (pixels per inch). Math for determining the print size in scanning: Step 1: Determine the total pixels Scanning a photograph of 4x6 inches at 600 dpi (dots per inch) will give a scan of 2400x3600 pixels. Step 2: Determine the physical print size If this image is printed out on a printer at 600 ppi (the same resolution as scanned), then the print will be the same as the original, 4x6 inches. 2400pixels/600ppi = 4 inches 3600pixels/600ppi = 5 inches If this same scan is printed out at 300 ppi, then the print will become 8x12 inches. 2400pixels/300ppi = 8 inches 3600pixels/300ppi = 12 inches If this same scan is printed out at 200 ppi, then the print will become 12x18 inches. 2400pixels/200ppi = 12 inches 3600pixels/200ppi = 18 inches If this same scan is printed out at 1200 ppi, then the print will become 2x3 inches. 2400pixels/1200ppi = 2 inches 3600pixels/1200ppi = 3 inches As seen in the example above, you can print the same captured image at many different print sizes. Generally, you have an idea of what the final output will be. Math for determining the print resolution. Determine the total pixels, or the pixel dimension of final image Pixel Dimensions = Print Dimensions (in inches) x Print Resolution (in ppi) 150 to 300 ppi is a decent range for resolution when printing to an inkjet printer. Say you have an image that you want to print at 5 in. x 7 in., and you have decided to print it at 300 ppi. The original is a 35 mm slide, approximately 1 in. x 1.5 in.
The desired pixel dimensions of the final print are: 5 inches x 300 ppi or 1500 pixels 7 inches x 300 ppi or 2100 pixels To calculate the scan resolution, you use the formula above in reverse because now pixel dimensions and the source's physical dimensions are known in the equation. Scan Resolution = Pixel Dimensions/Scan Source Dimensions (in inches) 1500 pixels / 1 inch = 1500 ppi or 2100 pixels/ 1.5 inch = 1400 ppi There is a discrepancy here, because the aspect ratio of the final output is not exactly the same as the original. The image will have to be cropped Tonal Adjustments during the Scanning Process It is best to optimize the tonal range and correct any significant color problems during the scanning process. While it is possible to adjust the image after the scan, you will only have the data you captured- trying to get the fullest range of data at the time of scanning will give you a wider range to work with. (a) (b) (c) (d) Scanned images with their histograms. (a) Scanned using a narrow tonal range. (b) The tonal range of the scanned image from (a) is re-adjusted by stretching the histogram. (c) Scanned with highlights cropped off. The lightest colors of the image all become white. The lost highlight information during the scan cannot be recovered by stretching the histogram or image editing. (d) Scanned with a maximized tonal range. Do not sharpen your image during the scanning process! It is best to do this afterwards, as part of the image adjustment phase. Capturing Digital Images by Digital Photography In traditional photography, images are made up of tiny grains of silver on the negative. Digital photography uses light sensors instead of silver grains.
Light hits the sensor and triggers electrical signals based on the light intensity. Such electrical signals are then converted into digital data and stored on the camera's storing device. Nowadays the sensor that a digital camera uses is either a CCD (charge coupled device) or a CMOS (complementary metal-oxide semiconductor). The size of the sensor and the number of light sensing sites determine the maximum resolution of the digital camera. In traditional film photography, generally, the smaller the grain size, the higher the definition and detail of the image you can obtain. In digital images, however, a pixel is the basic unit that makes up the image. Therefore, when you look at the specifications of digital cameras, you often see them specified in number of pixels. Digital cameras are often advertised using the number of megapixels. Megapixels The total number of pixels in a digital image are the pixel dimension width multiplied by the pixel dimension height. For example in an image 1600 x 1200: 1600 pixel dimension width 1200 pixel dimension height total pixels 1,920,000 This is 1.92 megapixels, and might be rounded up to 2 megapixels Does a digital camera with more megapixels give better image quality? Image quality depends in a digital camera depend on the optics, the electronics and the camera s image processing software. Remember when we are determining print size, a pixel has no physical dimensions, it is simply a point sample. To repeat the formula for print size: Print Dimensions (in inches) = Pixel Dimensions / Print Resolution (in ppi) Using a Digital Camera There are a number of types of digital cameras, like film cameras, most of them are digital point and shoot or digital single lens reflex (reflects the image optically onto the focusing screen or viewfinder by using the light coming through the lens. This means SLR cameras allow you to see the image area accurately from the viewfinder.) There are also digital camera backs for medium format cameras, so you do not need to replace the entire camera body. Digital Image Editing The common tools available in most digital image editing programs allow you to perform image retouching such as tonal adjustment, color correction and enhancement, and sharpening of the image. Many programs also have layers that allow you to composite and blend images in more creative ways beyond basic retouching. Common Steps in Image Editing 1. Cropping and straightening the image 2. Removing dirt and dust with a clone tool like the rubber stamp tool or healing brush in Photoshop
3. Adjusting overall contrast or tonal range 4. Removing color casts 5. Sharpening the image About sharpening: The sharpening algorithm finds edges which are defined by significant color changes. Sharpening creates the illusion of sharpness by creating a lighter line and a darker line on each side of the edge. Thus, applying sharpening will alter the colors in the image, especially where there are color changes, that is, at the edges. Because the edges are detected by the extent of color changes, all the color and tonal corrections or adjustments have to be made prior to the sharpening. Image resizing has to be done before the sharpening because resizing often softens the image focus. Sharpening should be the very last step of your image editing project. In Photoshop, you can exert the most control over sharpening using Unsharp Mask. Settings in Unsharp Mask: Amount: the amount of the contrast increase around the edges Radius: the radius of the surrounding area where the color comparison sharpening take place Threshold: how much the pixels differ from surrounding pixels in order to be identified as "edges" to be sharpened Color and Tonal Adjustments There are a number of tools for adjusting color and tonal information. These tools work by mapping the existing color or tonal values of the pixels to new ones. The common tools include adjusting the histogram, color balance, color curves, and hue/saturation. Adjusting the Histogram A histogram shows the relative number of pixels at each color intensity level. It is a bar chart. There is a histogram for each channel in an image, and each one has a slider representing the shadow tones, midtones, and highlights (in a grayscale image, this would correspond to black, grey and white.) By defining by the new positions of the black and the white sliders, you re-map the whole spectrum of color values of the pixels. In repositioning the black slider to the leftmost peak and the white slider to the rightmost peak, you stretch the histogram. That means you make the darkest color in the image to be black, and the brightest color in the image to be white. All other colors will be re-mapped in between in the same relationship of their original gray tones. By stretching the histogram, you maximize the use of the full spectrum of the gray tones. As you see, you increase the overall contrast of the image by increasing the difference of the color values between colors. Many digital cameras show you the histogram right after you have taken (captured) the image. Types of Selection Tools in Image Editing Software Predefined shapes such as rectangle and ellipse
Lasso and its variant s create a freehand outline Selection by color: Magic Wand, Eyedropper tool with Color Range tool Selection by drawing: Pen tool- creates vector paths Selections can be added to, subtracted from, intersected, feathered, moved. If you make a complex selection that you might need later on, you should save it. Layers Layers in Photoshop are like a stack of transparencies- the layer at the top can block out the layers beneath it. You can move them around, rename them, delete them, copy them. The background layer is specialized, you cannot make it transparent or move it, but you can turn it into a regular layer by double clicking it. Adjustment Layers allow you to apply tonal adjustments etc to a layer or layers that are not permanently changing the image. A layer mask associated with a layer lets you obscure just obscure, not delete part of the image on that layer. This non-destructive method of editing offers the advantage of preserving the original image in that layer. Printing To produce high-quality images, it is best not to scale the pixel dimensions of the image. Scaling the pixel dimensions is not the key to controlling the print size. But what is the best method to control the print size? It is important to understand that both the pixel dimensions of an image and the image resolution (ppi) affect the image's print size. Print Dimensions (in inches) = Pixel Dimensions / Print Resolution (in ppi) To adjust the output resolution or print size of an image, look in the image size setting. In Photoshop, for example, it is under Image > Image Size... When you want to maintain the pixel dimensions of the image, you will need to uncheck Resample Image. Then, when you alter the print size, the print resolution (ppi) will be automatically updated, and vice versa if you change the print resolution. With the Resample Image option on, the pixel dimensions can be varied. Scaling the pixel dimensions of an image is referred to as resampling because the number of samples (pixels) is changed. Upsampling (increasing the number of pixels) will create interpolated pixels- these are not based on data from the image, they are created by the software. Downsampling (reducing the number of pixels) will result in the loss of some information. However, if you are printing at a smaller size, it may not be that noticeable. Printer resolution is measured in dots per inch (dpi) the number of ink dots per inch. Color inkjet printers produce a microscopic spray of ink which appears to be very tiny dots of ink. These ink dots of different colors but a limited number of colors produce the required color by optical mixing. The color of a single pixel of an image is represented by a group of printer dots.
Do not confuse the image print resolution (ppi) with the printer resolution (dpi). As see in the scenario 2 above, the print resolution (ppi) affects the print size of an image. However, the printer resolution (dpi) does not affect the print size of the image. It affects quality of the print. Generally, the higher the dpi, the smoother the colors appear on the print. However, higher dpi uses more ink and requires longer time to complete the print job. Optimizing Images for the Web The three image file formats currently supported by Web browsers are JPEG, GIF and PNG. Each employs different compression algorithms to compress the file size. Because of the different algorithms, each file type works best with a different type of image. JPEG: continuous tone, 24-bit, lossy compression GIF: solid blocks of color, 8-bit, transparency PNG: PNG-8 supports 256 colors; PNG-24 24-bit color and transparency, lossless compression. Working with Vector Graphics Vector graphics deal with objects or shapes made up with paths, points, strokes and fills. A path is a mathematical description of an abstract line or curve. It does not have a physical appearance line width or color until the path is stroked with a specified line style, width and color in the vector graphic program. In most vector graphic programs, one of the tools for creating paths is called the Pen tool. You use the Pen tool to define points (anchor points) for the path. The curvature of the path is controlled by the handles (direction handles) of those points. An open path is a line, a closed path defines a shape. There are two basic selection tools in a vector graphic program: one that lets you select the whole object, and another that lets you select the points and their handles. * The material in this handout is excerpted from Chapter 3 of the Primer: Capturing and Editing Digital Images, a work supported by the National Science Foundation under Grant No. DUE-0127280 written by Dr. Yue-Ling Wong (ylwong@wfu.edu), Department of Computer Science and Department of Art, Wake Forest University. http://digitalmedia.wfu.edu/project/digital-media-curriculum-development/