Image size and resizing

"Change the size of the image" can mean either of two things: set it to print with a different number of pixels per inch, or change the number of pixels. The first of these is innocuous; the second almost always removes information from the picture.

13.3.1 Dots per inch

Consider dots per inch (dpi), or pixels per millimeter, first. A sharp color print requires 100 to 200 dots per inch (4 to 8 pixels/mm). Recall that a full frame from a 6-megapixel DSLR is about 2000 x 3000 pixels. At 150 dpi (6 pixels/mm), such an image makes a full-frame print about 13 x 20 inches (333 x 500 cm).

A much smaller, cropped portion of a picture can look sharp when printed the size of this book; one megapixel (about 1000 x 1000 pixels) is generally

Figure 13.2. Resampling an image to shrink it. The pixels that are to be combined are averaged together. (From Astro/photography for the Amateur.)

Figure 13.3. Resampling an image to enlarge it. The computer must fill in the missing pixels by averaging their neighbors. (From Astrophotography for the Amateur.)

considered the minimum for a full-page print. Clearly, with DSLRs we have plenty of pixels to work with.

13.3.2 Resampling

To save file space, or to make the picture fit on a Web page, you may wish to reduce the actual number of pixels in the image. This reduction should be done as the last step in processing because it throws away information. Once you shrink an image, you cannot enlarge it back to its original size with full sharpness.

Changing the number of pixels is called resampling, and Figures 13.2 and 13.3 show how it is done.

13.3.3 The Drizzle algorithm

The technique in Figure 13.3 assumes that every pixel covers a square area. In reality, pixels are spots; on the image sensor, pixels do not fill the squares allocated to them. Particularly when an image is to be rotated, it is unrealistic to

Figure 13.4. Histogram shows how many pixels are at each level of brightness.

Black Brightness level White

Figure 13.4. Histogram shows how many pixels are at each level of brightness.

treat a pixel as a square segment of the image. That is the key idea behind the Drizzle algorithm of Fruchter and Hook.1

To enlarge an image, the Drizzle algorithm takes only the central area of each input pixel and "drops" it (like a drop of rain) onto the output image. This reduces the effect of the square shape and large size of the input pixels. Missing pixels can be filled in by adding additional images or by interpolating.

Was this article helpful?

0 0
Learn Photoshop Now

Learn Photoshop Now

This first volume will guide you through the basics of Photoshop. Well start at the beginning and slowly be working our way through to the more advanced stuff but dont worry its all aimed at the total newbie.

Get My Free Ebook

Post a comment