Digital Imaging Basics

An ASMP "Strictly Business" Seminar white paper
by Scott Highton

 

Introduction

Digital technology has become a hot item in the business of photography in recent years. It is something new in an industry that has changed very little since its inception. Digital is hot, it's sexy, it's cool, it's new. Photographers and the rest of the visual communications industry are jumping on the bandwagon in a frantic effort to keep from being left behind.

The reality of this technology is that it's simply a new tool to help us do things we've been doing for years more efficiently. In its current state, it is painstakingly slow, relatively expensive and only just reaching a practical level. It is important, however, to not dismiss it as just another fad or trendy style that the photo industry throws at us yearly.

The creation of surreal photographic effects or the seamless compositing of one photo into another may be an "in" trend for the moment, but digital technology goes far beyond that. It is easy for those of us heavily invested in film-based photo equipment, who have established our careers by being good photographers, to panic at the idea that we will now have to buy thousands of dollars of new equipment and completely change the way we do business. It is even easier for us to ignore the technology and assume that it's only a passing fad, like light painting or tilted horizons.

Digital technology represents a major change in the way we deal with information. It promises to make communication and information processing far more efficient than most of us have imagined. The changes it brings to the business of photography are only a reflection of the changes it is bringing to our lives in general.

This does not, however, mean that we must all go out and immediately invest $30,000 to $50,000 in computers or digital cameras. It does mean that we need to educate ourselves about the technology and understand how it can be used in both our clients' and our own businesses.

We will not presume to teach you everything there is to know about digital photo technology here. There are dozens of seminars and workshops where you can learn intricacies of particular applications. Rather, we will give you an overview of digital technology as it relates to photography, so you will have a basic understanding of pixels, bits and bytes, which will enable you to determine for yourself what directions to proceed.

Before delving too far into the following information, understand that digital imaging, by it's very nature, involves a certain amount of math. Very little of this is actually necessary for the photographer or digital imager to understand in detail. Most digital imaging systems are designed to make their mathematical functions transparent to the user. However, in order to realize both the potential and the limitations of digital technology, the photographer will need to understand a few basic math concepts.

If the numbers or formulas presented here seem excessive, skip over those segments and try again later. It will be important that you understand them eventually, but it is better for you to learn what you can, rather than giving up because some high school algebra trauma haunts your past. If you can calculate f/stops, shutter speeds and can handle inverse square lighting relationships, you can learn this. Besides, the old joke is that the only math a photographer really needs to know is how to multiply and divide by two, and that Profit = Income - Expenses.

 

Overview

Digital technology encompasses many areas of the traditional photographer's business. It can allow us to record our photographs electronically, rather than on traditional silver-based film. It can allow us new power and flexibility to control the content of our photographs, whether originated on film or not. It can provide us with greater control over the access, usage and distribution of our images. We can make submissions, even to overseas clients, within hours or minutes, not days. It enables us to distribute unlimited copies of our images as perfect as the originals. It can eliminate the dangers of losing originals as they move to and from our clients. It is opening up new markets and new types of work for us to explore. It also presents potential threats to the copyrights and licensing of our images.

There are three basic components of digital imaging: Input, Processing and Output. Input involves a digital camera or scanner which actually converts light patterns (what we've traditionally recorded on film) into digital data or numbers that computers can work with.

Processing components are the computers or other electronic systems that we use to manipulate the digital data, whether it be color correction, compositing, sharpening or even preparing for storage.

Output components are the printers, film writers, printing presses and recording devices that we use to once again display the images. These devices convert the numbers of the digital data back into images that we can view with our eyes.

Storage might be included as a fourth component. These would be devices such as computer hard disks, CDs and other such elements of computer "memory" which allow us to store digital data for processing or output at another time.

 

The Basics

Digital data is a numeric representation of tangible information, whether it be the colors and shapes in a photograph, audio pitches and tones, text, numbers, calculations, graphic illustrations or combinations of them all.

In digital imaging applications, photographs are broken down into minute elements known as pixels. They are similar to the individual dots that make up a halftone for a printed photo. Pixels have also been likened to individual grains in silver-based film. Although the analogies are less than perfect, they do illustrate the concept in a way most photographers can relate to.

Each pixel is identified by its two dimensional location in the image and by the color it represents. The more pixels contained within an image, the greater the resolution and the more digital information available.

If a digital image contains only a few hundred pixels, the resolution will be quite low and the image may not even be recognizable. There is simply not sufficient information available for us to make sense of the result. Even with several thousand pixels, we still may not be able to recognize the image, but as we increase the number, we increase the resolution and detail that we can see.

 

Pixels and Image Resolution


400x267 pixels of 8-bit grayscale = 104 Kilobytes or 104K

 

   
 12x18 pixels grayscale = 216 bytes  24x36 pixels grayscale = 864 bytes
   
 48x72 pixels grayscale = 3.4K  96x144 pixels grayscale = 13.5K

 

The numeric representation of digital data is done using a binary system, which consists simply of ones and zeros. Each 1 or 0 is referred to as an information bit. Simply put, bits work like light switches, which are either on or off at any given moment. One is on, zero is off. The bit is the foundation of digital information. For the photographer, one bit of information at a particular point (pixel) in a photograph would mean that that point is either black (no color = off) or white (full color = on).

By combining groups of these bits together to represent a single pixel, we are able to represent levels in between the extremes of white or black. If we combine two bits to represent shading or color, we can represent four (2^2=4) different light levels (0-0; 0-1; 1-0 and 1-1). Thus, instead of having only black or white available, we add two levels of gray in between.

If we group three bits together to represent a single pixel, we have eight (2^3=8) combinations possible. They would include black, white and six levels of gray in between.

With eight bits per pixel, we get 256 possible combinations (2^8=256). This is the standard for current digital imaging systems. Using grayscale (also called "two-tone" or "black & white"), eight bits of information allows us to represent 256 levels of gray per pixel, which is quite adequate for photorealistic representation of black and white images.

 

Color Levels: Grayscale bit depth

Full grayscale: 8-bit black - 256 levels

 1-bit black: 2 levels

 2-bit black: 4 levels

 3-bit black: 8 levels

 4-bit black: 16 levels

 

Coincidentally, eight bits of information is known as a byte in computer jargon. A byte is the amount of computer memory required to represent one pixel of 8-bit grayscale information in a digital photo.

Including color information in a digital image requires more data for each pixel. Using the standard RGB (Red, Green, Blue) color model, most visible spectrum colors can be created by combining different amounts of the primary additive colors ­ red, green and blue. Computer monitors display these colors by combining 256 levels (8 bits) of each of the three primary colors (totaling 24 bits of color per pixel).

For each pixel, if you have only one bit of information per color, you can represent eight colors (2^(1*3) colors = 2^3 = 8). These would be as follows:

Red

Green

Blue

Result

Off

Off

Off

Black

On

Off

 Off

 Red

Off

On

Off

 Green

Off

Off

On

 Blue

On

On

Off

 Yellow

On

Off

On

 Magenta

Off

On

On

 Cyan

On

On

On

 White

 

If we had two bits per color per pixel, we would get four brightness levels for each of the three colors, giving us 64 possible color combinations (2^(2*3) colors = 2^6 = 64). This is referred to as the color depth and, in this case, would be referred to as "2 bits per color" or a total of "6-bit color."

Extending this to 8-bits per color, we would have more than 16 million possible colors (2^(8*3) = 2^24 = 16,777,216 colors). This again, is the standard for color digital imaging systems. It is similar to the color rendition capabilities of film, and is referred to as "24-bit color" or "8 bits per color" color depth.

Higher end digital imaging systems provide even more color depth. Many digital manufacturers tout the increased color depth of their equipment, claiming 10, 12 and more bits per color. A billion different colors can be represented with 10 bits of RGB color (2^(10*3) = 2^30 = 1,073,741,824). Sixty eight billion colors are available with 12-bits per color (2^36). Keep in mind however, that this increase in color information results in a corresponding increase in computer memory and storage required to work with the digital files. Even with today's high end desktop computer systems, 24-bit color files can get extremely large, demanding lots of expensive computer memory and long processing times for even the simplest operations.

With memory and storage concerns in mind, some manufacturers of input devices (digital cameras, film scanners, etc.) build their equipment to scan or "capture" the image with more than 24-bit color and then "optimize" it to 24 bits. These systems can recognize billions of colors, but when storing the image file, save it with the "best" 16 million colors.

If this seems a bit confusing, don't worry. Even an experience photographic eye will have difficulty telling the difference between a 24 and a 32-bit image, particularly since most systems use the 8-bit difference for the internal calculations required for compositing and special effects. Let the equipment manufacturers worry about the technical aspects and pixel processing algorithms. Consider simply that the range of colors which can be rendered by traditional transparency film is very similar to the range of 24-bit color on a computer.

 

Storage and Memory

Recording or manipulation of digital data requires that we be able to perform calculations and record the numbers representing the image pixels. This requires computer memory, or computing power and storage capability.

As an example, let's consider a digital image measuring 1,000 pixels wide by 1,000 pixels high, with each pixel containing 24 bits of color information. This gives us a total of 24 million bits of information (1,000 x 1,000 x 24 = 24 million). Computer memory and storage are generally measured in bytes, so with 8 bits per byte, our image requires 3 million bytes of memory.

If this image is black and white (grayscale) rather than color, then the memory or storage required is only 1/3 of that of a similar color image. A color image would require eight bits of information per pixel for red, green and blue, whereas a grayscale image needs eight bits of only one color (black). Thus, a 1,000 x 1,000 pixel grayscale image contains 8 million bits of information (1,000 x 1,000 x 8) and requires only 1 million bytes of memory.

A kilobyte (KB) is equal to 1,024 bytes, and a megabyte (MB) is equal to 1,024 kilobytes. (Kilobytes are often abbreviated as "K", such as "an 800K file" and megabytes are often called "meg", as in "it's saved as an 18 meg file.") Therefore, our 1,000 x 1,000 24-bit color image would be a digital file of 2,930 KB or 2.86 MB. This will be important later when we look at the usability of digital files.

In general, the larger a digital file is, the more detail it contains and the larger it can be successfully reproduced.

 

Reproduction

One of the benefits of digital data is the fact that it can be perfectly copied. Since a digital photograph is a long series of 1s and 0s, duplicating these number strings with a computer is quite simple. Every copy of a digital photograph can be as perfect as the original. There are no inherent contrast, grain, sharpness or color shift problems that we encounter when duplicating on film.

This obviously has advantages for photographers, because it allows us to submit digital files to clients without our original images ever leaving our offices. Since digital data can also be transferred over phone lines and electronically, we can transmit our images directly with our computers and modems or by utilizing online networks.

This ability to make perfect copies can also work against us. Unscrupulous individuals who may want to copy our images without permission can do so with ease once they have a complete digital file. There are ways to control this, however, which we'll discuss later. Computers are very good at keeping track of things and filing detailed information. It is relatively simple to keep what amounts to a digital paper trail, which tracks the history of a digital file. There are also data encryption and watermarking schemes that can offer some protection of digital images from unauthorized use.

Traditional color reproduction is both an art and a science. For decades, it has required experienced technicians in prepress and printing houses. Every press is different, as is the performance of different inks on different papers. Knowledge of separation, screen angles, dot gain, trapping, screen resolution, under color removal, registration and a plethora of other concerns particular to the print business have been required for successful print jobs. Suffice it to say that photographers haven't really concerned themselves with most of this, since prepress and production firms have dealt with it in the past. However, more of this work is now done on the computer desktop and a basic understanding of how the process works becomes more important for photographers.

Black and white printing is relatively simple, as it only requires one ink (black) and one impression of a printing plate on the paper. Reproduction of a black and white photo requires that we use a halftone screen, which optically converts the subtle gradations and tones of a photograph into small dots. The larger the dots, the more ink goes on the paper and the darker that portion of the image prints. Likewise, the smaller the dots, the lighter that part of the image prints.

Printing resolution is dependent primarily upon the frequency or number of dots representing the image ­ just as the resolution of a digital image is based upon the number of pixels available.

Halftone screens are available in many resolutions and are measured in lines per inch (lpi). For our purposes, lpi and dots per inch (dpi) are essentially the same thing. Typically, newspaper reproduction is done with an 85 to 100 line screen. Most magazines use 133 lines, while books are usually printed at 150 or more lines.

Converting data from the pixels of a digital image into a line screen for reproduction is relatively simple, at least for computers. Most page layout and image manipulation programs include this utility. However, it is important for the photographer to understand the relationship between the amount of data in their digital files and the amount required to reproduce their images on a printing press.

If you have too little information, or too few pixels, the printed piece shows jagged edges and the individual pixels become visible to the naked eye. If you have too much information, you wind up wasting unnecessary processing time and put undue demands on the prepress systems, which usually costs money. Thus, it is necessary to know what the size and resolution of the final output is before you can determine how much information you need in your digital image files.

The first rule of thumb is that you will generally want about twice the resolution that you will need for your output. This is not an absolute figure by any means. In fact, some experts claim you need as little as 1.25 times. You can experiment for yourself, but for our purposes, we'll assume that we need 2 times the resolution (called a "sampling" or "scaling" factor).

 

Example:

If we want to reproduce a black and white photograph sized to 4" x 6" in a typical magazine, we can calculate the digital file needed through the following formula:

2 x screen resolution (lines/in.) x reprod. height
x 2 x screen resolution (lines/in.) x reprod. width
_______________________________________
= file size (bytes)

 

(2 x 133 (lines) x 4 (inches))
x (2 x 133 (lines) x 6 (inches))
_____________________________
= 1,698,144 bytes (1.6 MB)

 

Thus, we would need a file size of 1.6 MB. This would be an image containing 1,064 x 1,596 pixels (8-bit grayscale). If we were reproducing the same photograph in color, we would need to triple the file size, because we would need this same amount of information for each of the three RGB colors (24-bit color).

1,698,144 bytes x 3 colors = 5,094,432 bytes (4.9 MB)

 

Notice how our required file sizes change when we change the reproduction size or resolution. If we size the same photo to 3.5" x 5" (a rough equivalent to a 1/4 page photo in the typical consumer magazine) instead of 4" x 6", we get the following:

Black & white:
(2 x 133 x 3.5) x (2 x 133 x 5) = 1,238,230 bytes (1.2 MB)

Color:
(2 x 133 x 3.5) x (2 x 133 x 5) x 3 = 3,714,690 bytes (3.5 MB)

 

If we increase the reproduction size to 8.5" x 11" (full page or cover bleed), we'll require a far greater file size:

Black & white:
(2 x 133 x 8.5) x (2 x 133 x 11) = 6,615,686 bytes (6.3 MB)

Color:
(2 x 133 x 8.5) x (2 x 133 x 11) x 3 = 19,847,058 bytes (18.9 MB)

 

Note that if we lower the screen resolution to an 85 line newspaper level, the file size for the same 8.5"x11" reproduction decreases to about 40 percent of that required for the magazine resolution:

Black & white:
(2 x 85 x 8.5 ) x (2 x 85 x 11 ) = 2,702,150 bytes (2.6 MB)

Color:
(2 x 85 x 8.5 ) x (2 x 85 x 11 ) x 3 = 8,106,450 bytes (7.7 MB)

 

This information is important when we are acquiring the image in digital form, whether it be from scanning a film image or shooting with a digital camera. The size and resolution of the digital file limits our successful use of it.

 

Reproduction Sizes

   Scaling Factor = 2

 Description

File Size:
B&W

 File Size:
Color

 Pixels

 (85 lines)
Newspaper

 (100 lines)
Brochure

 (133 lines)
Magazine
Photo CD Master:          
Base/16 24K 72K 128 x 192

.8" x 1.1"

.6" x 1"

.5" x .7"
Base/4 96K 288K 256 x 384

1.5" x 2.3"

1.3" x 1.9"

1" x 1.4"
Base 384K 1.1MB 512 x 768

3" x 4.5"

2.6" x 3.8"

1.9" x 2.9"
Base x4  1.5MB 4.5MB 1024 x 1536

 6" x 9"

5.1" x 7.7"

3.8" x 5.8"
Base x16 6MB 18MB 2048 x 3072

12" x 18"

10.2"x 5.4"

7.7"x11.5"
Pro Photo CD:          
Base x64 24MB 72MB 4096 x 6144

24" x 36"

20" x 30"

15.4"x23.1"


Double Density Disks:        
Grayscale  800K    739 x 1109

4.3" x 6.5"

3.7" x 5.5"

2.8" x 4.2"
Color RGB    800K  427 x 640

2.5" x 3.8"

2.1" x 3.2"

1.6" x 2.4"
           
High Density Disks:          
Grayscale  1.4MB    978 x 1466

5.8" x 8.6"

4.9" x 7.3"

3.7" x 5.5"
Color RGB    1.4 MB  564 x 847

3.3" x 5"

2.8" x 4.2"

2.1" x 3.2"

Notes:
1) Reproduction sizes are approximate.
2) Maximum sizes on floppy disks will actually be slightly less than the full capacity of the disks.

 

Many people will make the mistake of scanning or acquiring every image at the highest possible resolution, thinking that it's better to have too much information that too little. While this approach does provide more options for the use of an image, there is little point to having to deal with a 72 MB file when you're only reproducing it as a 2"x3" photo in a black and white catalog. The excess data demands more computing power (read $$$), more memory, more storage space and slows down every step of the computing process. Pushing pixels around in a digital file is slow enough as it is, even with today's fast desktop computers. There is little point in making these systems work harder (and slower) by demanding millions of unnecessary calculations.

Sometimes, you may want to reverse the above calculations in order to determine what size you can successfully reproduce a given digital image. For instance, a high density floppy disk can hold almost 1.4 MB of data. How large could you reproduce a file that size in a newspaper or magazine? Let's start out printing a color image in a standard magazine:

 Area of printed image = Bytes ÷ (2 (scaling factor)
÷133 (line screen)) ÷ (2 ÷ 133) ÷ 3

 

 

To convert MB to bytes:
1.4 MB x 1,024 (KB/MB) x 1,024 (bytes/KB) =1,468,006 bytes

1,468,006 bytes ÷ (2 ÷133) ÷ (2 ÷ 133) ÷ 3 = 6.9 sq. inches (approx. 2.1" x 3.3")

 

If this is a black and white (grayscale) file, then we use the same formula but don't divide by the three RGB colors. Thus:

1,468,006 bytes ÷ (2 ÷133) ÷ (2 ÷ 133) = 20.7 sq. inches (approx. 4" x 5")

 

If we change the printing screen resolution to 85 lines (newspaper), we can print the image files as follows:

Color:
1,468,006 bytes ÷ (2 ÷85) ÷ (2 ÷ 85) ÷ 3 = 16.9 sq. inches (approx. 3.5" x 4.8")

B&w:
1,468,006 bytes ÷ (2 ÷85) ÷ (2 ÷ 85) = 50.8 sq. inches (approx. 6" x 8.5")

So just like most other aspects of photography, everything is a tradeoff. If we increase resolution, we decrease reproduction size. Reproducing color requires three times more data than black and white. The better or larger we want to display our work, the more digital data we need.

 

Compression

Digital data, particularly from digital photo files, can be reduced in volume through a variety of mathematical algorithms or compression schemes. These systems, which have acronyms such as JPEG, ADCT, LZW, MPEG, etc., all reduce the numbers in a digital file into more efficient equations that a computer can store in a smaller amount of memory. Some are called "lossless," which means that there is no loss of quality or data in the compression/decompression process. Others, which compress files into the smallest sizes, are called "lossy." Generally, the more a file is compressed, the more information is lost and the greater the degradation of the image.

For those who use lossy compression schemes, the tradeoff in quality is usually worthwhile because of the value of reducing the file size. Some compression schemes can reduce file sizes by almost a 30:1 ratio, with little noticeable loss of image fidelity. For the user who needs to store large image files in small spaces of computer memory, or who needs to transmit images over phone lines, the advantages are obvious. Using a 9,600 baud (bits per second) modem, an 18 MB image file would take almost four and a half hours to transmit over a phone line. Compressing the image file reduces transmission time, and the dollar savings are often worth the relative losses in quality.

Compressed images need to be decompressed before they can be opened or used again. The decompression process replaces the data that was lost in the compression of the file with the exact data that was removed (lossless) or with estimated data (lossy). For the most part, photographers don't need to be too concerned with the complexities of compression, except to understand the difference between lossless and lossy systems.

 

Photo CD

In 1990, Eastman Kodak introduced Photo CD, a system that can store digitized photographs on a 43/4" diameter compact disk (CD). There are currently a number of Photo CD formats, and the technology has become an industry standard. Basically, Photo CD is designed to allow the scanning and digital storage of film-based images at a relatively low cost. Although originally targeted for the consumer market with the hopes that consumers would want to look at family snapshots on TV, it has yet to be widely accepted there. However, the photographic, publishing, computer and communication industries have embraced it wholeheartedly.

The Photo CD Master format can store up to 100 scanned images at five different resolutions each (called an "Image Pac"). It uses a proprietary technology (Photo YCC) developed by Kodak and stores the equivalent of 1,800 MB of photo digital data on each 600 MB compact disk. Each "Image Pac" includes five different resolutions of the same image. A sixth resolution is available on the newer and more expensive Pro Photo CD format.

 

Photo CD Image Pacs

Description Pixels File Size Use (Kodak described)
Base/16 128 x 192 72K Thumbnails & index use
Base/4 256 x 384 288K Low resolution preview, FPOs & comps
Base 512 x 768 1.13MB Viewing on standard TVs and monitors
Base x4 1024 x 1536 4.5MB Proposed HDTV resolution
Base x16 2048 x 3072 18MB Full resolution (35mm), almost full page magazine reproduction
(Pro Photo CD)    
Base x64 4096 x 6144 72MB Medium format and 4"x5" film scans, high resolution reproduction

 

While many other manufacturers sell scanning and photo digitizing equipment, Photo CD is important to mention because of it's relatively low end-user price and because it is now a standard for both the distribution and storage of digital photo files. Although the Photo CD Image Pacs can only include the above resolutions, Kodak has designed the format so the resolutions available meet most user needs, including those of traditional publishing and the newly developing multimedia fields.

Hundreds of Photo CD service centers are available throughout the country, which can scan 35mm film images (both slide and negative) for between $1 and $2 each. Compared with desktop publishing service bureaus that charge between $10 and $25 for similar scans, the price break is significant. Additionally, the storage media (the CD) is usually included in the price of Photo CD scans. With service bureau scans, you generally have to provide your own SyQuest cartridge ($50 - $100) upon which you can only store a few 18 MB images.

The difference between these services is that the Photo CD centers have invested $100,000 or more in PIW (Photo Imaging Workstation) equipment and are dealing in high speed and high volume scanning, whereas other service bureaus tend to use desktop scanners in the $2,000 - $20,000 range and work at a lower, more specialized volume. There are advantages to both, although service bureau prices have fallen dramatically since the arrival of Photo CD.

Since Photo CD has become an industry standard, stock agencies, software and clip art publishers, online photo networks and most of the publishing industry have adopted it for storage and electronic distribution of their images. Using Photo CD files requires little more than a desktop computer with a CD-ROM drive and low cost software from Kodak. (Popular digital imaging software packages, such as Photoshop, generally also include access to Photo CD files).

Photo CD base resolution provides sufficient data for full color, full screen display of the photos on most televisions and computer monitors. Base x 16 resolution is adequate for 8"x10" reproduction in most magazines ­ almost a full page. In fact, Kodak has claimed that base x 16 resolution contains the same amount of picture information as many of its 35mm consumer films do.

 

Security

In 1985, the United States Office of Technology and Assessment (OTA) released the results of a survey about America's perceptions of right and wrong when it came to copying the work of others. The report, entitled "Intellectual Property in an Electronic Era," indicated that 70 percent of the general population thought it was acceptable to copy other people's work. Forty percent of the American business population did, too.

The report also said that "public relation strategies (to counter these practices) are likely to be most effective if they focus not on the rights of copyright holders, but on the relationship between the copyright holder and the users of the work. Messages about unauthorized copying may be more effective if they emphasize the ongoing value of a partnership between creators and users."

Copying and unlicensed use of photographs have been made far simpler, and the quality of such reproduction has become far better with the advent of digital technology. Once a digital file leaves the photographer's possession, it can be copied repeatedly without any loss of quality. This presents a threat, particularly those who distribute high resolution digital files on CD-ROM or via online picture services.

However, the traditional methods of photographer self promotion ­ sourcebook ads, promo pieces and reprints ­ are even more vulnerable to theft due to the tremendous quality and availability of desktop scanners. There have been hundreds of cases reported where clients simply scanned an image from a photographer's printed promotion piece and published it without ever needing original film. Desktop scanners and digital imaging software make it relatively simple to acquire any image electronically, to enlarge it, sharpen it or alter it if necessary.

While unauthorized use of photographs has always been a problem, the fact that it is so much easier today means that unscrupulous types have an easier time committing their crimes. Even more of a concern is that 40 percent of people in business and 70 percent of the general population aren't even aware that it's not OK to copy our images.

There are however, ways to protect our work. One unsuccessful method, which a number of CD producers and online providers mistakenly trust, is to only provide "screen resolution" or smaller data files. These distributors believe that their files are small enough that they can only be reproduced at sizes less than two inches square. Their assumptions are based on the same formulas that we used earlier.

The standard computer screen displays images at 72 dpi, or a total of 640 x 480 pixels. For a 24-bit color image, this is a 900 K file. Photo CD base resolution is 512 x 768 pixels, resulting in a similar 1.12 MB file. Using the formulas listed earlier with a scaling factor of 2 and a standard 133 line screen for a consumer magazine, we get the following:

Screen resolution (640x480 pixels):
921,600
bytes ÷ (2 ÷133) ÷ (2 ÷ 133) ÷ 3 = 4.3 sq. inches (roughly 2.5" x 1.75")

If we take this same file and change the scaling factor to 1.25 (which some experts use), we can increase our reproduction size over 150 percent:
921,600 bytes ÷ (1.25 ÷133) ÷ (1.25 ÷ 133) ÷ 3 = 11.1 sq. inches (roughly 3.8" x 2.9")

This is getting close to the size of a 1/4 page magazine ad. If we use the same formula on a Photo CD base resolution file, we get a 4.7" x 3" reproduction.

Furthermore, most software applications like Photoshop offer the ability to do data interpolation, which expands the size of the digital file and fills in the data between the original pixels with with similar data. A variety of interpolation algorithms are available and, when used with unsharp masking and other sharpening algorithms, can successfully expand the usable dimensions of a digital image up to 400 percent (or 16 times the file size).

Thus, in a worse case scenario, if an end user were to take our original "screen resolution" file, use a 1.25 scaling factor and quadruple the dimensions of the image, they would wind up with a 14 MB file and could reproduce it 15.4" x 11.5" in size ­ almost a two-page spread in most consumer magazines. Using the more common scaling factor of 2, this file could still be reproduced 9.6" x 7.2" in size ­ close to a full page. Fractal-based interpolation algorithms currently under development promise even greater interpolation amounts.

It is also important to note that the new electronic markets, such as multimedia and other electronic uses, only require a maximum of screen resolution (640x480 pixels), since they will only be used for display on television and computer monitors. Delivering a screen resolution file (900 K or Photo CD base) to an electronic or multimedia client is the equivalent of delivering original film to a traditional print client.

Once again, it is not necessary for photographers to commit this math to memory. Rather, it is important for you simply to understand that the technology to reproduce and improve your "low resolution" digital images is readily available, and that "low resolution only" is not sufficient protection against unauthorized use.

 

Better Security

One method of better protecting digital images is to use a combination of watermarked and encrypted image files. The watermark, such as a large copyright © symbol, is a semi-transparent overlay on the "preview" or screen resolution image, which allows the viewer to see the entire image clearly, but makes it impractical for publishing. Generally, a watermarked image will be accompanied by a "clean" image in another file, which is data encrypted so it can't be opened without passwords. This allows the photographer to submit preview images to clients with a watermark embedded, and then if the client decides they want to use the image, they can agree on a usage fee over the phone with the photographer, who then provides the passwords to unlock the "clean" encrypted file.

While it is possible for an experienced Photoshop user to remove the visible watermark from the preview images, it would take a number of hours to do yet would still only result in a screen resolution image. Thus, it's generally cheaper for a user to buy the needed rights and get passwords for the encrypted file, which is usually large enough to be reproduced up to 1/4 page. If larger reproduction is needed, the photographer can send film. It's not a perfect system, but it keeps the honest people honest, and makes it very clear that the digital photos are not provided as clip art or free-use images.

One criticism of this system has been that art directors and clients looking at lots of pictures, feel that the watermarks are distracting and degrade the quality of the images. They argue that when scanning hundreds of thumbnail sized photos on a computer, the visual distraction of the watermark detracts from an image. They say they are less likely to choose a watermarked image over a similar unmarked image. Whether this criticism is valid depends upon whom you talk with, but it has been expressed as a concern.

A more recent advance in the watermark arena is the Digimarc invisible watermarking technology, which introduces a virtually indistinguishable (to the human eye) pattern of identifiable noise into the pixel patterns of any digital image file. This noise pattern can be detected with Digimarc reader software, and can survive duplication, halftone screening or color separation and reprinting of an image. It has tremendous potential as a copyright protection tool in the digital world for photographers and other visual artists, and allows protection of the images without the distraction of a visible watermark. This technology was introduced in 1996 and was incorporated that year into industry standard software packages such as Photoshop 4.0 and CorelDraw7. Digimarc, Portland, OR

One of the great promises of the Pro Photo CD format was to be its ability to store encrypted high resolution image files. However, Kodak has yet to include this ability because they've not received National Security Administration (NSA) clearance on their encryption scheme. The NSA requires that any encryption system available in a product that might be used outside the U.S., be cleared by them, as part of their responsibility to protect national security.

It is critical to keep a photographer's credit, copyright notice and reference information with their photographs once those images have been stored as a digital file. Many software programs provide text files to accompany digital images. In the various text fields, data about each photo, including copyright notices, captions, photographer, file number, model release and much other information can be stored. These text fields are generally used for database searches, particularly by online picture services and electronic stock agencies. Unfortunately, the text files are easily stripped away from the digital image when the files are opened with software other than that used to create it. This is a very common occurrence.

It is therefore, critical that the photographer's name, copyright notice and a source ID be included as part of the bitmapped image of photos stored for electronic distribution. This can appear as a credit bar above or below the actual image area. It can still be cropped out by the end user, but it at least offers more assurance that they will see and note the copyright information before using the image. Digital files are opened and re-saved by clients as many times as traditional film is handled by individuals. The copyright information must be kept as part of the digital image data, just as a copyright notice remains on the slide mount for a piece of film.

 

 

Digital Paper Trails

Another advantage to digital technology is the fact that computers are designed to deal efficiently with countless minute details that would drive humans nuts. This includes cross referencing of files and the automatic recording of transaction data. This can be a boon to picture professionals concerned about tracking their images.

It is relatively simple to create data files accompanying our digital images that include information about the photos. This can include caption, exposure, copyright, publication history, sales price or electronic transmission information. Essentially, we can create electronic information trails for our digital images, which are far easier and faster to use than paper records.

Most online picture services use such practices. They can keep track of who previewed what images on their system, what choices they made, how their search pattern worked and what their preferences were. Also, they can record how the picture was downloaded or sold, what the usage was and what keywords the buyer used to search for it. Profiles of clients can be developed based on their past choices ­ whether they prefer their sunsets to be orange or red, in the upper left or lower right of the frame, whether they prefer horizontals or verticals. All this information can be used by the photographer to better understand his or her markets, and can also be used to track image files.

Of utmost importance is the fact that digital technology brings the need for new clauses in photographers' delivery memos, estimates, contracts and other forms. We must clearly state what rights are being granted and for how long, particularly in uncharted multimedia waters. Some photographers include clauses that prohibit any electronic alteration or scanning of their images without prior written permission. All should specify that the client will destroy, erase or return to the photographer any digital files that were created from the photographer's work. This prevents future unauthorized reuse (accidental or not) by the client and is just as appropriate as requiring that the client return original film after their contracted use.

 

Summary

There is a great deal for all of us to learn with the new digital tools available to us. There are several basics, however, that are important to keep sight of:

1) Digital technology is not a passing fad. It represents a major change in the way our collective business and personal lives are run. It should not be ignored.

2) It is not necessary for any of us to rush out and blindly jump into the digital world. Investing thousands of dollars into new computer equipment or digital cameras without first understanding the needs of our own businesses, can be suicidal. Many of us may never even need to touch a digital camera or scan a piece of film. What is necessary is that we understand the basics of how the technology works ­ that we stay informed about how our clients are using it and what their needs are. This information will help us make our own decisions and help us run our businesses appropriately.

3) Technology changes fast. What is state-of-the-art today, may be obsolete in six months. Do not fear the changes, because technology improves through change. Be aware. Stay informed. Make your decisions based upon the best information available at the moment.

4) Most importantly, don't lose sight of why you became a photographer. Our business still requires us to be visual communicators, and creating interesting, informative and inspiring images is the basis of that work. Don't get sidetracked into becoming a techno-weenie and lose sight of why you are a photographer. Be creative. Make good pictures.

...And enjoy what you do.

 

 ©1994 Scott Highton
All Rights Reserved

Reproduction of any portion of this volume requires written consent of the author.

 


 

Appendix

The following formulas are also useful for file size and reproduction calculations:

Conversions:
1 Kilobyte = 1,024 bytes
1 Megabyte = 1,024 Kilobytes = 1,048,576 bytes

 

To calculate required file size for a particular reproduction size:

Color (24-bit) images:
File size (bytes) = (Scaling Factor x Line screen x Reproduction height) x (Scaling Factor x Line screen x Reproduction width) x 3

Grayscale images:
File size (bytes) = (Scaling Factor x Line screen x Reproduction height) x (Scaling Factor x Line screen x Reproduction width)

 

Other file size formulas:
1 pixel = 1 byte (grayscale files)
1 pixel = 3 bytes (24-bit RGB color files)

Total pixels in image = pixels in length x pixels in height
Area of reproduced image = length x height

 

To calculate maximum reproduction size of a file:
Area of printed 24-bit color image = bytes ÷ (Scaling factor ÷ Line screen) ÷ (Scaling factor ÷ Line screen) ÷ 3

Area of printed grayscale image = bytes ÷ (Scaling factor ÷ Line screen) ÷ (Scaling factor ÷ Line screen)

Max. length of reproduced image = pixels in length ÷ Scaling Factor ÷ Line screen
Max. height of reproduced image = pixels in height ÷ Scaling Factor ÷ Line screen

 

 

Bonus formula for techno weenies:
Suppose you have a digital file of a certain size and you want to determine the maximum reproduction length and width (not just the area or product of the length and width, but the individual length and width measurements). Assuming you only know the file size and the format of the original (35mm, 6x7, 4x5, etc.), along with your line screen frequency for reproduction, you can do it as follows:

1) Calculate the area of your printed image as described above.

2) Determine the height to length ratio of your original. Examples: 35mm = 2/3, 6x6 = 1, 6x7 = 6/7, 4x5 = 4/5, etc.

3) Length of printed image = Square root of total printed image area (1) ÷ square root of height to length ratio (2)

4) Height of printed image = Total printed image area (1) ÷ length of printed image (3).

Example:
Your local service bureau has provided you with 14.8 MB scans (24-bit RGB) of your recent prize-winning 6x7 color transparencies, which were used for an awards program. You want to reproduce the photos again as part of your own promotion calendar for next year. Knowing that your calendar separations will be done with a 150 line screen, what are the dimensions of the largest reproduction you can do from each of these digital files?

Convert MB to bytes:
14.8 MB x 1,048,576 = 15,518,925 bytes

Calculate area of printed image: (Scaling factor = 2)
1) Area of printed image = 15,518,925 bytes ÷ (2 ÷ 150) ÷ (2 ÷ 150) ÷ 3 = 57.5 square inches

Calculate max. length:
2)
6x7 image = 6/7 ratio

3) Reproduction length = square root of 57.5 ÷ square root of 6/7 = 7.58 ÷ .92 = 8.2 inches

4) Reproduction height = 57.5 ÷ 8.2 = 7 inches

Total reproduction size = 8.2" x 7"
(6x7 cm original scanned as 14.8 MB file printed with 150 line screen)