Image digitization works by dividing the image into dots, giving each dot a value, and then storing all the dot values in sequence in a file. The computer reconstructs the image by reading the dot values and displaying them on the computer screen, with the dots in the same order and with the same value as they had in the original. These dots are called pixels, or picture elements. The range of values each pixel can have is called the pixel depth. One aspect of pixel depth is color.

All of these images below were scanned from the same black and white photograph. Notice that although the photograph is "black and white," it does not include only the colors black and white, and it looks strange if you try to scan it that way.

If you are viewing this page in Netscape, notice that your browser will not show the binary black and white image. That is because this is a bitmap file (extension .bmp), and this particular format is not recognized by Netscape.

For the Holyoke-Reaves collection, the images were all scanned in true color (16.7 million). If you look at the variations in the images on this page, you will see that the true color image looks much nicer than even the 256 color image.

binary image	Binary images: Every color and shade of color (including black, white, and grey) is indicated by a unique code. A '0' (absence of light) pixel equals black and a '1' (presence of light) pixel equals white. These images are also called one-bit images, as only one '0' or '1' (a single bit) is needed for each pixel in the image. Binary images are satisfactory when the original object is black and white only (such as printed text or line art). This is an example of a binary image.
dithered binary image	Dithering: Even when computers record only black and white, as in binary images, they can create the illusion of grey by grouping black and white pixels into cells, a process called 'dithering,' so that the eye thinks it sees one grey and not some black and some white. This is an example of a dithered black and white image.
greyscale image	Greyscale: In addition to plain black or white, pixel values can express different levels of grey. For example, each pixel could have four values (levels): in binary code, pure white ('11'), light grey ('10'), dark grey ('01') and pure black ('00'). This is a greyscale image. This example is a four greylevel (having four levels of grey), two-bit image, with a pixel depth of four. Because we need two bits for each pixel for a four greylevel image, as opposed to one bit for each pixel for a binary image, a two-bit image file will be twice the size of its one-bit counterpart. Typically, a greyscale image might use up to 8 bits (e.g. '00000000' for black, '11111111' for white) for each dot. An eight-bit image will be eight times the size of its one-bit counterpart.
8-bit, 256-color image	Color: In addition to using the extra values for each pixel to represent levels of grey, they can also be used to represent colors. Eight bits for each pixel equals 256 colors, 24 bits for each pixel equals 16.7 million colors. In 24-bit images, the color of each pixel is made up of three eight-bit values: in the common 'RGB' format one value represents the 'redness' of the image on a 256 level scale, another value its 'greenness,' another its 'blueness.' The computer reconstructs the color of each pixel by giving it so much red, green, and blue as each of the three eight-bit values specify. The range of 16.7 million colors given by 24-bit images is usually enough to satisfy the human eye, and so 24-bit images are sometimes referred to as real color or true color. The two images on the left represent the 256-color image and the 16.7 million color image. Spot color is used for images with large areas of uniform color, such as a pie chart for a presentation. Below is an example of what happens if you scan a normal image using spot color.
24-bit, 16.7 million-color image	spot color image
Read about Resolution Go back Home