Manual Scanning

The objectives of this exercise are:

To really, really understand how an image can be analyzed and measured, and the measurements saved as a collection of numbers, and those numbers can be used by a display device to recreate a facsimile of the original image. In short, how digital images work.

This exercise requires three people to play the roles of Scanner, Recorder, and Imager.

Step 1 The Scanning Person will divide the image into a grid of square picture elements (pixels) by laying a wire grid on it. This gridded image is called a Raster Image, taken from the Greek and Latin word for rake. The wire grid we place on our image has 1" squares so it will produce one pixel per inch horizontally and vertically.

Step 2 The Scanner will use the cardboard grayscale to estimate a gray tone for each pixel. The grayscale has eight steps from white to black, numbered 0 through 7. You will start at the upper left corner and evaluate each pixel horizontally to the end of the row, then return to the left side and scan the next row of pixels.

Step 3 As the Scanner calls out the value of each pixel, the Recording Person will write the numbers into a spreadsheet-like grid on the whiteboard. The numbers will be stored in rows and columns mimicking the grid laying over the image being scanned.
Step 4 The Imaging Person will use a set of gray squares to build an image from the data stored on the whiteboard. These squares are numbered and match the eight tones on the Scanner's grayscale. As the Recorder writes the numbers on the whiteboard, the Imager will lay down the gray squares to match the recorded numbers.
The result: the image built by the Imager will crudely resemble the original image. How could you improve the Imager's picture? You could divide the original image into smaller chunks. This would be called "increasing the resolution" of the scanned image

Since the scanned image has only eight shades of gray, it doesn't accurately capture the many possible gray tones in the original image. You could improve the scanning by using a grayscale with more tones, and giving the Imager the same increased number of gray toned squares.

By increasing the resolution, you would get a more accurate reproduction but the process would require much more time for the Scanner, more time and more Whiteboard space for the Recorder, and much more time for the Imager to lay down the larger number of squares.

This simple exercise describes the basics of digital images and provides a visible model of the processes that take place in our computers, scanners, monitors and printers.

More Details

Scanning Resolution:

If you need more image detail, you will need more pixels in the scan. You control the resolution when you select a ppi or dpi value in the scanner software. If you are using a digital camera you select different levels of resolution, or total pixels in the image. If you create an image with software, such as Photoshop, you set the resolution when a new file is created or you can resample the pixels with the Image Size controls in the Image menu.

Storage:

You need to be familiar with a computer's method of data storage and the terms used to describe quantities of data.

The most basic unit of computer information is the Bit. A bit is a Binary Number; it is either a one or a zero. It is like a light switch, either on or off. In fact, the chips in computers contain millions of transitors which are tiny electronic switches. Every process in a computer is accomplished by sending electricity to those tiny switches to change their on-off status or to read their on-off status.

Most information we need to work with is more complex than the two choice a binary number can provide. To record and use bigger numbers, we combine more than one bit into a group to describe more possible variations on a piece of information. We might have a font with 200 characters, or we might have a scan which has assigned one of 256 gray tones to each pixel.

Sidebar:

The Indexed Color file format uses very basic pixel and color tricks. It reveals the working color lookup tables.

 

To handle these bigger numbers we add bits. Here is how it works:

 

Important facts and concepts

Bit Depth

1 bit provides 2 choices for information; i.e. one or two, black or white, Seahawks or Broncos, etc.

2 bits combine to give you 2 times 2 choices, or 4 choices. As you add each additional bit to a group, it multiplies the groups choices by 2. The math term is "powers" of 2; the number of times 2 is multplied by 2.

So groups of bits can describe the following number of data choices:

1 bit 2x1=2 choices

2 bit 2x2=4 choices

3 bit 2x2x2=8 choices

4 bit 2x2x2x2=16 choices

5 bit 2x2x2x2x2=32

6 bit 2x2x2x2x2x2=64

7 bit 2x2x2x2x2x2x2=128

8 bit 2x2x2x2x2x2x2x2=256

...and on and on as more bits are added.

In raster images, the number of bits grouped to describe each pixel is called the "Bit Depth" of the image.

A very common group of bits is 8 bits. The 256 choices are enough for all the characters in a type font or all the gray tones needed to reproduce the smooth tones in photographs. In honor of this usefullness, the 8 bit group is called a Byte. Think of a byte as a computer word. (There are some computers which use a different number of bits in their bytes, but we won't be using them.)

Since we use so many bytes, the Roman numeric shorthand has been adopted to describe the big numbers we see listed in our file size windows.

One thousand bytes is a Kilobyte

One thousand Kilobytes is a Megabyte (a million bytes)

One thousand Megabytes is a Gigabyte (a billion bytes)

One thousand Gigabytes is a Terrabyte (a trillion bytes)

Terms related to this exercise:

Raster (derived from Rastrum, Latin for toothed hoe or rake)
Raster image - bitmap
Raster image processor (RIP)
Pixel - short for Picture Element
Resolution - desrcibes the image resolution, or clarity of detail. More pixels give greater resolution.
Terms for resolution: Pixels per inch (PPI), Samples per inch (SPI), Dots per inch (DPI), Spots per inch (SPI, again), and Lines per inch (LPI).
Halftones - read the Haltone Story, and study the Halftone Creation tutorial.
Adjust output values with the Laser Printer Calibration exercise.
PostScript - read how Postscript makes halftones
Bit depth
Line Art, 1-bit, or Bitmapped images
Vector images
Resolution Independent