4
4.1 Digital Images
A fundamental requirement of the digital intermediate process is that all images involved in the process must be digital. Because of this requirement, the entire digital intermediate pipeline is prone to exactly the same limitations and potential pitfalls inherent in digital media. Just as a cinematographer needs at least a basic understanding of the photochemical process, so people using the digital intermediate process need to have a basic understanding of the properties of digital images.
The concept of the digital image has been around since the early days of computing. Back in the 1960s, getting a computer to display even a small picture took incredible resources. Nowadays, we are so overwhelmed with digital images that we barely notice them. From pictures on websites, photos from digital cameras, to interfaces on mobile phones, digital imaging helps us interact more intuitively with technology, as well as provides perfect copies of pictures that can be transmitted across the world almost instantly or stored on disks for future use.
Digital images may be created in a number of ways. Perhaps the most common method—converting (or digitizing) images from another media (such as photographic film)—is covered in Chapter 5. Another common method is to use a digital camera to photograph something that can provide a set of digital images. Finally, it is also possible to create images from scratch, completely within a computer. Images created in this way are called “computer-generated” (CG) images.
4.3 The Anatomy of a Digital Image
All information that can be processed by computer technology (i.e., data) is binary. This means that if you were to look at a piece of computer data in its simplest form, it could be described as a combination of ones and zeros—for example, 10110010. It can also be thought of as a bank of switches, each of which can be either “on” or “off.” This is true for any type of information that flows through any digital computer system. Given the correct context, the computer knows how to interpret this stream of numbers, whether it is a digital image, a spreadsheet, or even a piece of software code. The combinations of ones and zeros form the building blocks of any type of data. In theory, digital images could be represented in a multitude of different ways; however, in practice, almost all digital images comprise the same basic features.
Bits and Bytes
The smallest unit of computer data is a bit (short for “binary digit”). Each bit can be either a zero or a one. The most common unit of data is a byte, which is a group of 8 bits. Each byte of data can have 256 (28) possible combinations of ones and zeros. Therefore, each byte can have any of 256 different values. These values can refer to anything, such as letters of the alphabet in a text document. The kilobyte (KB) is 1024 bytes. Most simple text documents are several kilobytes in size. One megabyte (MB) is 1024 kilobytes (approximately one million bytes). A floppy disk can hold around one-and-a-half megabytes of data, enough for most documents, or small or compressed digital images. One gigabyte (GB) is 1024 megabytes (approximately one billion bytes), which is enough to contain about five minutes of DV (digital video) footage. Finally, the terabyte (TB) is 1024 gigabytes (approximately one trillion bytes). It takes approximately one-and-a-half terabytes to store a 90 minute film at 2k resolution. Refer to the Appendix for a list of file sizes for common applications.
4.3.1 Pixels
The most common type of digital image is bitmap (or raster) images, which are conceptually similar to video images. With this type of digital image, each image comprises a number of building blocks known collectively as “pixels” (short for “picture elements”). Each pixel is a binary number, representing a single square of a solid color. Put enough of them together, and you have a picture, in the same way a detailed mosaic is made from many small ceramic tiles.
Figure 4–1 Individual pixels can be seen when an image is viewed close up
Most imaging systems work in a similar way. You can see the individual grains that form a photograph when you view it close up. Because the grains in photographs are randomly shaped, they are harder to distinguish than pixels of the same size, but the concept is the same. Pixels are regularly shaped and arranged in a gridlike manner for the sake of efficiency, making them easily displayed and simplifying all the “behind-the-scenes” mathematics performed by the computer (and making these processes therefore faster).
Digital images have square-shaped pixels, and so do most devices for viewing digital images, such as computer monitors. When a computer monitor displays an image, it traces a small spot that corresponds to each pixel.
In broad terms, increasing the number of pixels in an image also increases the image’s level of detail and increases the file size, because more bytes of information are needed to record each pixel. Doubling the number of pixels in an image will double its file size. Doubling the width and height of an image will quadruple its file size (by doubling the length and width, you quadruple the area, and hence the number of pixels).
4.3.2 Pixel Aspect Ratio
Although it’s not very common, digital images can have nonsquare pixels. This means that rather than being perfectly square, each pixel is meant to represent a more rectangular shape. There are a few reasons why you might want to do this, but the most common is to match pixels to the CCD elements in video cameras, which are rectangular rather than square. The actual shape of the pixel is described by a pixel aspect ratio, which is simply the ratio between the height and width of the pixel. Square pixels have a pixel aspect ratio of 1.0. PAL DVCAM images have a pixel aspect ratio of 1.07, indicating that each pixel is 7% wider than it is tall.
Where the ratio becomes relevant is in displaying the images (e.g., on a monitor). Almost every digital display device is configured to display square pixels only. Displaying nonsquare pixel-based images becomes more of a problem the bigger the image is. To compensate for this, the display device must add (or remove) a proportion of pixels to the image (for display purposes only) so that the overall image has the correct proportions. The Appendix includes a list of pixel aspect ratios for different imaging formats.
4.3.3 Print Resolution
Some digital images encode a “print size” or the number of dots per inch (dpi) or pixels per inch (ppi). Even more confusing, this is often referred to as the “resolution” of the image, which is a somewhat unnecessary (and confusing) measurement. It basically means that if the image is printed, there is some way to correlate each pixel to realworld measurements. For instance, a 100 × 100 pixel image, printed at 100dpi (meaning that 100 pixels are in an inch, or alternatively, that each pixel is 1/100 of an inch in diameter), will be exactly 1 inch by 1 inch in size. Reducing the dpi to 50 will result in a 2 inch by 2 inch printed image. However, the amount of information (and hence the detail) is unaffected by differences in the dpi value. Whether there are 10 pixels per inch, or 10,000, makes no difference to the total number of pixels, and hence the spatial detail and file size of the image. The dpi value is an arbitrary figure used in reproducing an image in print form. Therefore, the dpi value of an image usually can be ignored (unless you need to print the image, of course).
The Problem of File Size
Increasing the file size of an image has a variety of side effects. When you increase the file size of any piece of data, you are increasing the amount of information contained in that file. The result is that it takes longer to access a bigger file, longer to move or copy it, and more disk space is needed to store it. In addition, computers use memory known as “random access memory” (RAM) to perform operations (such as displaying or resizing) on an image. Like disk space, RAM is a limited commodity and can only hold a finite amount of data at any one time. Also, regardless of disk space or RAM limitations, more computations are needed to modify larger files (which means that larger files are manipulated more slowly).
A useful analogy is that of a library. Imagine that your computer is a library, and your images, rather than being pictures, are books, with each book describing a picture. A bigger book (or a larger file) contains a more detailed description of the picture and requires more pages to do so. That’s why the book is physically bigger. If you imagine that the shelves in the library are analogous to storage space (such as disk drives) within a computer, you can see that bigger books require more shelf space to store them. Similarly it takes longer for someone to read (or access) the bigger books, and because they are heavier, it takes longer to move them around.
Now imagine a table in the library, where you can spread out all the pages from the books and look at them all at once (in the same way that files can be loaded into the RAM of a computer system). Again, your table only has so much space, and thus it can hold a limited number of pages. If you want to reorganize, make corrections, or copy any of the books into a new book, it will take physically longer to write out each new book.
This analogy also provides clues as to methods that can solve some of these file-size issues. For example, a library may have a vault or similar external storage facility for infrequently used books. In the same way, a computer system might have an offline archive device, such as a tape backup system, so that infrequently used files can be stored externally. You can build one or more additional libraries, move some of the books to the new libraries, and mail books between them. With computer systems, you can do something similar; network separate computer systems via cables (or even wireless transmissions).
Ultimately though, there must be a balance between speed, capacity, and cost in any digital intermediate pipeline. Many facilities aim to choose file sizes that are optimal for the color-correction system, and they then base the rest of the pipeline around that size, because color correction is by far the most interactive stage of the process.
4.3.4 Tonality
Every single pixel has a color assigned to it. A group of pixels can all be the same color, but a single pixel can’t have multiple colors associated with it. In a very simple image, each pixel has a choice of being either black or white. This limitation is necessary, for example, for LCD displays on mobile phones, which can only display pixels as black or white (i.e., each pixel has a binary value for color—0 or 1, black or white, respectively). While this may not seem like much, with a highresolution image, it is possible to get a fairly detailed black-and-white image.
However, for images with a wide tonal range, describing color in terms of black or white pixels isn’t terribly efficient (or accurate). Each pixel could be any shade of gray. So, in most monochrome (or gray scale) images, each pixel is given a value between 0 and 255, of varying intensity (for a total of 256 shades of gray). Each pixel actually could be given any number of shades of grey, but having 256 possible values (or levels) makes it convenient to store as an 8-bit number in a computer. Because each pixel in the image has an 8-bit number associated with it to describe the shade of gray (or the brightness or luminosity of the pixel), the image is said to have 8 bits-per-pixel, or a bit depth of 8 bits. Increasing the bit depth allows for a greater number of possible shades, and thus a greater tonal range.
Figure 4–2 This image uses a combination of about 650,000 black or white pixels to simulate tonality
Figure 4–3 This image has the same number of pixels as the one in Figure 4–2 and each pixel has 256 shades of gray to choose from
Theoretically, there is no limit to the bit depth of an image. However, increasing the bit depth also increases the file size, because more information is needed to encode each pixel. An image with a 10-bitper-pixel depth allows 1024 shades of grey and would result in a file that is 25% larger, compared to a file with 8 bits per pixel.
4.3.5 Color
When painting, you can make new colors by mixing different primary colors together. For example, mixing yellow and blue paint makes green. This process is known as the “subtractive method of color mixing,” because the more colors you mix in, the closer you get to black. To produce color images on a monitor, different amounts of red, green, and blue light are mixed together. This process is based on the “additive system of color mixing,” where the primary colors are red, green, and blue, and the more colors you mix together, the closer you get to white. Any color can be made by mixing different quantities of red, green, and blue light, which is the process used to display color images on monitors, televisions, and so on.
A color digital image contains a number of channels. Each channel is a single, monochrome image. Most color digital images have red, green, and blue channels, which are mixed together to form the fullcolor RGB image. Each channel contains a possible number of color values (in the same way that gray scale images do), which provides the possible color range for each channel. In a sense, three separate grayscale images are mixed together to make a color image, so there might be three separate 8-bit-per-pixel channels in every image. In most cases, each channel shares the same characteristics of pixel dimensions and bit depth as the others, to enable them to be combined in a fast and meaningful way. For instance, in a typical 400 by 400 pixel color image, there will be three channels (one each for red, green, and blue), having 400 by 400 pixels at 8 bits per pixel. The entire image will therefore have 400 by 400 pixels, with an effective overall bit depth of 24 bits per pixel (or 8 bits per channel). The resulting file will be three times bigger than its monochrome counterpart and have a total of approximately 16.8 million possible colors available to the image.
In addition, other paradigms combine channels in ways different from RGB images. For example CMYK images combine four channels (one each for cyan, magenta, yellow, and black) in the same way that the printing press combines inks on paper to form color images (using a subtractive color model). A HLS model uses three channels, one each of hue, luminosity, and saturation. Other color models use one channel to describe the luminance, and two for the color content, such as the Lab model or the color models used in high dynamic range (HDR) formats (see Chapter 13 for more on HDR images). For the most part, different models can be used to produce the same digital image, but there may be differences in terms of the color space of the image, meaning that certain models can produce colors that others can’t reproduce. This topic is examined further in Chapter 8.
Figure 4–4 Digital images can use any of a number of different models to reproduce colors. © 2005 Andrew Francis. (See also the Color Insert)
It is also possible for an image to have extra, nonimage channels. The most common fourth channel is known as the “alpha” channel and is usually used as an extra “control” channel, to define regions of an image that can be used when modifying or combining images. For example, a pixel with an alpha channel of 0 may mean that the pixel shouldn’t be modified at all, with an alpha channel of 255, the pixel would be affected fully, and a value of 127 might mean that the pixel would be affected at 50%. There may be additional channels to perform specific functions for color correction, masking, and so on. Each additional channel will increase the file size—doubling the number of channels will typically double the file size. Alpha channels aren’t generally displayed as part of the image and won’t affect its appearance.
4.3.7 Transparency
Digital images need not be opaque. By encoding separate transparency information, each pixel in the image may have a degree of transparency.
Figure 4–5 The alpha channel is separate from the color channels and doesn’t affect the appearance of the image. In this case, the alpha channel might be used to mask one of the boots. © 2005 Andrew Francis
Transparency can be encoded in several ways. The most common way is to simply use an alpha channel to define the amount of transparency. The value associated with each pixel in the alpha channel defines the level of transparency for that pixel.1 Another way to encode transparency is to nominate a specific pixel value to indicate transparency. Using this method, every pixel of the specified color is taken to be transparent. A final way is to make use of vector graphics (which are covered later in this chapter) to specify regions of transparency.
4.3.8 Layers
A layered image is one that contains several separate images combined or stacked together. Layers allow superimposing (or compositing) images on top of each other, animation, or other special effects. While each layer is a separate image, it typically shares qualities such as pixel resolution, bit depth, and number of channels per image.
4.3.9 Motion
There are a few ways to turn a digital image into a moving image. The simplest is to vary one of its parameters over time—for example, by cycling through the colors in an image. In an image with multiple layers, each of the layers might be moved about over time, independently from each other. These methods are fairly limiting but result in small file sizes, as effectively, only one image (with perhaps several layers) must be stored, along with a small amount of data describing the motion. Displaying the motion correctly depends on the ability of the display system to provide the correct motion when the file is viewed. This method is suitable for special situations and is commonly seen in flash animations on some websites. However, to efficiently display high-quality images, the same method is used for video and film—a series of images is presented to the viewer in rapid succession.
Figure 4–6 Layers may be combined using transparency, alpha channels, or other methods to produce a composite image
Using this method, a number of frames are shown at a predetermined frame rate. As with video and film, this method of working with digital media produces the illusion of motion to the viewer. Unlike video and film, however, digital media isn’t subject to a specific frame rate. Digital media can be played at 24fps to match film projection, 29.97fps to match NTSC video, 10 million fps, or 1 frame per day. Most of the time, the frame rates used in a digital intermediate pipeline are determined by other factors, such as the frame rate of the original or output media, or the limits of the playback system.
It’s also worth noting that unlike other media, digital images can be played in a nonlinear fashion. With video and film formats, footage can be viewed only in sequence (although it is sometimes possible to alter the speed or play a sequence in reverse), whereas with digital media, frames can be accessed and displayed in any order, depending upon the software used.
4.4 Digital Image Operations
One of the main advantages of digital technology is that manipulation and analysis of the underlying data is very simple, compared to analog equivalents. Because digital images are simply numbers, it is possible to affect underlying parameters just by doing simple math. For example, to increase the brightness of a pixel, you increase the corresponding level of the pixel. To increase the overall brightness of an image, you increase the level of all pixels by a fixed amount; to darken them, you just lower the values. Many more operations are possible, by performing simple or complex calculations on individual pixels or convolving regions with a matrix (which is discussed in the following section).
Figure 4–7 Increasing the values of the pixels results in an increase in brightness. © 2005 Andrew Francis
Figure 4–8 A convolution matrix can be applied to a selected area—in this case, sharpening the image. © 2005 Andrew Francis
Convolution is a process where a digital image is modified using a mathematical matrix of numbers to transform the image. One of the more common convolution matrices is for sharpening an image. Changing the size or values of the convolution matrix will increase, reduce, or alter the effect.
Many other options are available, usually referred to as digital image “filters,” that can be used for a variety of artistic, analytic, or special effects. Many digital-image-processing applications are available that allow a vast number of different operations across an image, or even to just a localized area.
4.5 Alternatives to Raster Images
In addition to the many types of raster images, there are other paradigms for representing images digitally, such as vector graphics. Vector graphics are basically mathematical representations of shapes, such as rectangles and circles. They can be filled shapes, can be transparent, and can overlap. The advantages of vectors is that they are resolution independent, meaning that you can zoom into or out of them without having problems such as aliasing. A vector curve is always going to be perfectly smooth, and edges will always be perfectly sharp. Vector graphics typically require less information to store, resulting in smaller file sizes than equivalent raster images.
A disadvantage is that vectors can’t easily represent the complex details that a photograph or pixel-based image can. Also, vector graphics are difficult to generate. You couldn’t, for example, make a vector image by scanning a photograph. It is important to remember that even a vector image must be rasterized (i.e., converted to pixels) before it can be displayed; however, this rasterization needn’t be permanent and is for display purposes only. Although vector graphics don’t intrinsically support the use of color, many applications assign colors and patterns to the edges and contents of the shapes.
Several vector-based imaging programs are available, and many imaging programs combine vector graphics with pixel-based imagery. Many digital intermediate systems use vectors to define regions for selective editing of images.
Figure 4–9 A vector image is made up of mathematical coordinates that form shapes
Figure 4–10 Even when the vector image is zoomed into, the sharp edge is retained
Figure 4–11 In some cases, it’s possible to apply colors and patterns to the individual shapes
Figure 4–12 A fractal image has detail no matter how far you zoom in
Similar to vector images is the idea of fractal images. Fractals are complex formulas that generate infinitely complex, resolutionindependent patternse. In a fractal image file, only parameters that describe the formulas must be recorded. For this reason, the file sizes can be incredibly small, smaller even than vector-based images, and yet the information they contain can be very detailed. However, it is very difficult to generate anything other than strange patterns with fractals, and intense computation is required to display them. For these reasons, fractal images are less common than vector graphics. Fractal technology is applied to other aspects of digital imaging, such as special types of image compression.
4.6 File Formats
As there are so many different ways to represent a digital image, there are literally hundreds of different digital image file formats that can be used for different purposes. The formats can vary based upon the number of channels and layers, the bit depth, and the color space. Some formats are highly standardized, and thus are guaranteed to work with a large number of applications; some are less standardized but still highly popular. (The less-standardized formats can cause headaches when you try to move them between various applications.) Some have been optimized for various functions, such as the ability to load a small part of an image at a time. Some formats are even more flexible, allowing any number of options to be determined separately for each file; however, these formats are normally limited to being readable by a specific application.
For moving picture images, each frame can be stored as a separate digital image file, or an entire sequence can be saved as a single, selfcontained file (which applies to video tapes and reels of film). Again, there are many different formats for storing such media, some of which have a high level of compatibility with a great number of processing applications.2
The file formats that are relevant to the digital intermediate process are covered in later chapters.
4.7 Compression
One of the main problems with digital images is large file sizes. The size of a single raster image is roughly
number_of_pixels × number_of_channels × bit_depth_per_channel × number_of_layers
It is easy to quickly produce large files. For example, a film-quality image is typically taken to be 48MB. This size is too large to put onto a website or to email. Furthermore, the high volume of images produced for feature films (typically 50–100 million frames per film) means that even relatively small images require a large amount of disk space.
One of the ways to solve this problem is to reduce the amount of data. For example, you can lower the resolution of an image by a process known as “interpolation,” which merges or discards a proportion of pixels to produce an image with fewer pixels. (Interpolation typically involves the use of some form of anti-aliasing to preserve detail and maximize quality.) Other methods may be to limit the color palette to a much smaller number by using an “indexed-color” image.
Indexed-color images have a preselected palette of colors available to use in the image. Rather than assume that you need 16.8 million different colors in a single image, a much smaller number of colors is chosen from the 16.8 million possibilities, and the pixels are assigned a color from this palette. This process results in a much smaller file size but at a cost of a reduced number of available colors in each image. For example, GIF files allow a color palette of about 256 colors, preselected from a range of about 16.8 million, thereby reducing the file size to a third of the original size.
The most common method for reducing file size is to use a form of compression. There are two main ways of compressing a digital image. The first, known as “lossless” compression, simply rearranges the saved data in a more efficient way (e.g., by indexing all the colors in such a way that the most popular colors have a smaller index number), using less disk space.3 On average, it is possible to reduce the file size by half (without any alteration to the content or quality of the image), but doing so is dependent on the content of the image. For all intents and purposes, a file compressed with lossless compression is identical to its uncompressed counterpart.
An alternative way of compressing digital images is to use “lossy” compression. This method optimizes image size by discarding some of the data—ideally, data that is superfluous. For example, JPEG (the popular format for the distribution of photographic images over the Internet) compression works by reducing the number of colors based on the color response of the human eye. In theory, it discards colors the eye wouldn’t detect easily anyway. Other methods are even more esoteric—for example, converting areas of the image to equivalent fractals.
4.7.1 Visually Lossless Compression
The term “visually lossless compression” is something of a buzzword among manufacturers. Essentially, it refers to compression methods, such as the JPEG compression method, that reduce quality, based on the human eye’s limits of perception. In theory, there is nothing wrong with using visually lossless compression because the imagery is perceptually the same, which is why it’s very suitable for consumers. However, its strength lies in the assumption that all you are going to do with images compressed in this manner is view them. In the digital intermediate pipeline however, this assumption is incorrect because frames undergo many mathematical operations, such as resizing, color-correcting, and so on, all of which benefit from the available quality outside of the perceptive range of the human eye.
By way of an analogy, let’s say you have a scene containing a stack of gold bullion on a table. If you were to set up a camera and tripod to photograph the scene from a specific angle, you might find that some of the gold would not be visible through the lens. You could safely remove those extraneous gold bars from the scene and take the photograph, and no one would know the difference. However, doing so means that you can’t change the camera setup without revealing the missing gold. In the same way, visually lossless compression removes parts of the image that aren’t seen but inhibits editing ability later on. The destruction of information by the compression method becomes clear as you make changes to it, such as by color grading.
Each compression method introduces different artifacts into the image. You can often get a better understanding of how a particular compression method affects an image by loading a digital image and then saving it with maximum compression. This pushes the compression algorithm to such extremes that the quality reduction becomes visibly apparent in many cases. This problem is often further exacerbated by repeatedly recompressing files.
In addition, many compression artifacts become apparent only when watching the footage at speed. Any form of lossy compression is, by definition, a destructive process, and its use is therefore recommended only as a final process before displaying the image.
Figure 4–13 Too much compression can lead to visible artifacts
4.7.2 Motion Compression
With a large number of stills, a lot of data has to be moved around to view the files in real time, without any lagging or dropped frames. With low-end computer systems and/or high-quality images, compression becomes very important for the viewing of such images. For example, a typical 90 minute program at SD video resolution requires approximately 150GB of space to allow it to be stored uncompressed. However, a consumer DVD can only store up to 9GB, and consumer equipment can’t move around such a high volume of data at speeds fast enough for real-time playback.
Data Corruption
When a piece of film is damaged, it can become scratched or torn. With video tapes, the noise level may increase, or drop-out or interference may occur. When a digital file is damaged (i.e., corrupted)—either due to physical damage to the storage device, or a software error in accessing or creating the files), individual pixels, lines, or even frames are destroyed—they are inaccessible or otherwise randomized—partially dependent upon the file format used. Chapter 9 details ways of correcting some of these problems. Data has one significant advantage over other media though, which is that a backup copy is always as good as the original.
For digital moving picture formats, compression methods can work across a range of frames. Lossy motion compression typically works by comparing the content of each frame to the previous or next frame (or both frames in bidirectional compression) and looking for differences. Because a relatively small amount of difference exists between adjacent frames, space can be saved by only storing the differences, rather than storing each individual frame. The degree of sensitivity in detecting these changes can be adjusted and combined with lossy still image compression. To aid playback of these files, a parameter can be set on certain types of files to provide a fixed bit rate of each file, so that one second’s worth of images occupies a fixed amount of disk space (the quality level being continuously adjusted to meet this target), resulting in smooth playback. Variable bit rate compression methods continuously adjust the footage’s bit rate, so that more detailed images lose less information than less detailed images.
Render Artifacts
A whole class of problems can occur during the process of rendering images. Rendering is the process of converting one set of (usually computer-generated) data into another, suitable for display. Rendering is commonly used to convert a multilayered, composited image or shot, or a 3D scene, into a single-layered (i.e., flattened) sequence of images. In fact, almost all color-grading systems will involve rendering to output the color-graded footage. During this process, many errors can occur—errors caused by a lack of precision in calculations, glitches in the rendering software, or even sporadic hardware faults. Such faults can create noise or corruption artifacts. Fortunately, most rendering processes aren’t destructive—they create new files rather than overwriting originals—so if spotted early enough, re-rendering the images will usually fix the problem. This is why it is important to carefully check every rendered image as early as possible in the digital intermediate process.
4.8 Encryption and Watermarking
Another option granted by digital media is the use of encryption. This allows each digital file (each image frame or video stream) to be encrypted or scrambled with a special code. A user or system then supplies the correct code, and the file is decrypted (i.e., unscrambled). Without the code, the encrypted image doesn’t display properly. Other implementations, such as watermarking, stamp a hidden pattern over an image. The hidden pattern can contain information about the origin of the image, for instance, which can be used to trace transferred images or prevent operations such as editing or copying the image. As with lossless compression, most image-encryption methods shouldn’t affect the quality of an image.4 However, it may affect the performance of the system because additional computation is required to decrypt files each time they are accessed. Encrypted images tend to share the same disadvantages of lossless-compressed images. Further, encryption may also exclude the use of lossless compression as well. In some cases, it is possible to use an encryption method that allows lossless compression to be used in conjunction with encryption, but in most situations, it’s a case of one or the other (or most frequently, neither). Encryption and watermarking techniques will probably see more usage as the performance of imaging systems improves (and thus the additional computation involved in working with and encrypted images becomes negligible), and so does the awareness and paranoia of filmmakers and distributors toward the threat of computer hackers and piracy.
4.9 Metadata
Finally, it is worth noting that almost all digital image formats or storage methods have some proviso for encoding additional information (or metadata), such as a text comment, along with each image or group of images. This functionality is entirely dependent upon the software and file formats used. Even where the use of metadata isn’t natively supported, it can be augmented in other ways, which is covered further in Chapter 6.
4.10 Problems with Digital Media
By their very nature, digital images are prone to a number of problems, or artifacts. These problems may be invisible except under certain circumstances, or they may significantly degrade and distort the image. Artifacts are generated either during the digitization process, or through image processing, such as compressing or colorgrading the image. The most common digital artifacts are covered in Chapter 12.
Pretty much all of the problems with digital media have to do with the “quantization” of information. Unlike analog sources that record continuous streams of information, digital images (by definition) contain discrete (i.e., separate or distinct) units of information. Most of the failings of digital images can be directly attributed to this characteristic.
An important requirement of moving pictures is that the images are processed (and displayed) at a regular, usually predetermined rate. Analog video systems are designed to work in real time when recording or playing back. This means that one second of recorded footage takes exactly one second to display or copy. With digital media however, this isn’t necessarily the case. Computer systems have a lot of variables attached to them (e.g., the specifications of the system, the software running on them, network conditions, and cabling quality) that all affect the playback. Displaying a series of frames on a computer system may result in a different experience each time. Computer systems can suffer from “lag,” which is when a queue (or line) of frames to be displayed builds up, making a fixed-length sequence take longer than real time to display completely (although at no loss in image quality). Alternatively, digital images may be prone to dropped frames, where certain frames are discarded to maintain the real-time speed of playback. Conversely, if the host computer system is displaying frames too quickly, stuttering may occur as some frames are held on the display for longer than others. Dropped frames and stuttering are purely display issues and don’t affect the quality of the data saved in any way (unless you are recording the output of the display, e.g., to video).
Flicker
The human eye can detect changes of illumination at a much higher rate than it can discriminate motion. When film is projected on a cinema screen, a visible “flicker” occurs when the projector shutter closes and advances to the next frame. To overcome this issue, each frame is usually flashed onto the screen twice, resulting in the shutter opening 48 (or 50) times every second, thus eliminating any visual flicker. Televisions and monitors use a similar approach, although flicker is strictly a displayrelated issue and therefore doesn’t affect the creation or manipulation of digital media at all.
4.11 Summary
Digital media offers many advantages over other types of media, such as the capability of making perfect duplicates quickly. However, lots of variables must be considered. Generally speaking, as the quality of an image increases, so do the associated costs. While a larger file might contain more pixels, and therefore more detail, it also becomes less practical to work with, requiring more storage space and additional processing and transmission time.
There are no specific standards when working with digital images. They may be encoded in a variety of different file formats, each suitable for specific purposes. Digital files may have additional options available for tracking and protecting images, or they can be compressed to reduce storage requirements. Lossless compression can reduce the size of each file without compromising quality; lossy compression can produce much smaller files but may degrade the image.
However, digital images can suffer from the lack of sufficient information to produce colors or details, thus exhibiting a number of artifacts. Some artifacts may only become noticeable late in the digital intermediate process (e.g., when performing color grading).
In the next chapter, the journey into the digital intermediate process begins. The first stage in the process is getting digital and analog material into the chain, while maintaining the highest level of quality.
1 Conventions vary as to whether black or white pixels in the alpha channel represent full transparency.
2 Even the highly standardized DV format can be saved on a computer in a number of different ways.
3 In certain situations, some types of lossless compression can create a larger file size than the original!
4 Watermarking an image alters the image slightly, although it does not normally introduce significant degradation.