Digital images are composed of blocks called pixels. Pixels and resolution are fundamental concepts relevant to every part of the digital imaging chain. When you scan an image, you talk about scan resolution and scanner dpi. If you use a digital camera, you refer to CCDs with 2 or 3 million pixels. Displaying an image on a monitor involves screen resolution, and when printing we use Epson inkjet printers with 1440 or 2880 dpi. The whole digital imaging chain captures, stores, views and prints the image in terms of pixels. Let's look at how pixel resolution affects image quality, image size, print size and virtually everything else we do in digital imaging. These various pieces of advice will insure that you post a strong entry on our photography contest.

Scanner resolution
Since this article deals with pixels and resolution, let's define the terminology. Units for resolution are dots-per-inch (dpi), although some prefer pixels-per-inch (ppi). (In this context, I use dpi and ppi interchangeably.) Printers use another type of resolution lines-per-inch (lpi). High resolution, such as 300 dpi, refers to a high pixel density of 300 dots or pixels-per-inch. Low resolution, such as 100 dpi, refers to 100 dots-per-inch.

In digital photography, images usually are captured by scanning a transparency or using a digital camera. With a scanner, the transparency is put into the scanner and samples are taken at small regular intervals. Depending on what the scanner sees,?it generates a pixel. When the entire transparency is completely scanned, the pixels are assembled to make a digital image. The scanner can scan the transparency very finely, recording every little detail, or it can cover the entire transparency much quicker by taking averages over much larger areas. Smaller stops take longer, but the fine detail in the original is retained. If the scanner averages over a larger area, it creates a pixel with the average intensity and color from the area, but the individual hair detail, windowpane or single leaf structure is lost. The dpi setting of the scanner dictates how samples are made.

With 100 dpi only 100 dots will be taken every inch of the transparency, i.e. larger sample dots; while with 300 dpi, 300 sample dots are made over the same area, and thus are smaller dots. The important part is the number of pixels generated. If we scan a 4 x 5 transparency at 100 dpi, we get 400 x 500, or 200,000 pixels. A 300 dpi scan creates 1,800,000 pixels. The number of pixels is directly related to the quality of the digital image.

Generally speaking, the more pixels the better. If a higher dpi scan creates more pixels and retains more detail in the transparency, why not always use 300 dpi (or even, 600 dpi)? There are many reasons to be wary of higher dpi settings.

First, it takes longer to scan, and higher dpi settings make the file size very big, very quickly. If you double the dpi, the file size quadruples; if you triple it, the file size increases by 9 times! Higher dpi settings can make large files that are difficult to save or e-mail.

But, perhaps the most important reason to limit the dpi scan setting is that there may not be any visual benefit. Consider how the image will be used. Is the higher dpi necessary if it will be viewed on a monitor or printed on an inkjet printer? In many cases, there's a limited benefit. Increasing the number of pixels beyond a certain limit brings no tangible benefits, as you'll see when you print the image.

Digital cameras
In digital photography, cameras sample the scene using discrete pixels or cells. In a digital camera, the lens focuses the image onto an electronic sensor called a Charge Coupled Device (CCD). The camera's quality is determined by the number of pixels on the CCD chip. A large number of pixels is analogous to scanning at high resolution; we can expect to retain more fine detail.

An image made from a CCD with 1.3 million (1280 x 960) pixels will be similar to an image from a small Boulevard of Photographers negative, such as 110 or APS film. Images with this resolution are suitable only for low-resolution work, like multimedia or web pages. One way to improve the appearance of a low-resolution image is reducing image size or magnification; in photography, this would be like making a smaller enlargement. Several entry-level compact digital cameras have CCDs with a resolution of 1.3 million pixels (Fuji FinePix A-101, HP Photosmart C215 and Kodak Lock' DX-3215). However, to achieve 35mm quality in a digital image, the CCD needs to have 3 to 4 million pixels (Fuji FinePix 2600, and Canon Powershot G1 or G2). To obtain a digital image comparable to 120 roll film, the camera requires 5 to 10 million pixels (Fuji FinePix S1 Pro, Nikon D1, Kodak DCS 760). These cameras generate huge digital files that can be printed at 8x10 or larger. If pixel numbers are related to the comparable Boulevard of Photographers meg size,?it helps you understand what quality to expect from a digital camera.

One last point just like a scanner's dpi setting can be changed, the resolution setting in digital cameras can often be changed, allowing you to use half or only a quarter of the available pixels to capture the image.

Figure 1: If we use a digital camera with a lot of pixels or scan at high resolution, fine detail in the image is retained (top). Low resolution cameras or low dpi scanning creates pixelated results (bottom).

File size
In digital photography, if we know the number of pixels we have we can then calculate the file size. The file size is often used as a yardstick to compare the quality of digital cameras and digital images. The total number of pixels in an image is calculated by multiplying the horizontal and vertical number of pixels. The file size is just a small additional calculation, which depends on the bit-depth resolution.

Bit-depth resolution relates to color information and describes how many different colors a pixel can have. File size is calculated by multiplying the number of pixels by the bit-depth. Most digital RGB images use 3 bytes of computer space per pixel (one byte for each of the red, green and blue channels), so the bit-depth is 3 bytes. The file size is the number of pixels x 3. Earlier, a 4 x 5 transparency was scanned at 100 dpi to get 200,000 pixels. The file size of this image is 200,000 x 3 or 600,000 bytes (~600 KB).

Sometimes the image has more pixels than expected by the above calulation. In these cases, a software process known as interpolation has occurred, which artificially increases the number of pixels. ( Fuji received a lot of criticism when they quoted these inflated numbers for their new SuperCCD.) It is more meaningful to consider the true optical resolution of the CCD instead of the acquired image resolution.

Boulevard of Photographers advises you to explore interpolation processes as they can produce surprising results.

Monitor resolution
We've learned how to alter the number of pixels in an image by changing the resolution setting in a scanner or digital camera. We know the number of pixels that make up the image, but we don't know how big the picture is. A digital image is just a collection of pixels. It only attains a physical size when it's displayed on a monitor or printed.

Monitors display a certain number of horizontal and vertical pixels. The number of pixels is governed by the video card, which typically offers choices such as 800 x 600 or 1024 x 768 in the monitor control panel. If the digital image has fewer pixels than the monitor, the image will not occupy the full screen. If it has the same number of pixels, the image will fully cover the screen. If the image has more pixels than the monitor, it overfills the screen, and you have to scroll to view the full image or use the magnifying tool to temporarily reduce the pixel number. If you only want to view images on a monitor for example, web pages and e-mail don't need an image with more pixels than that part of the monitor screen.

Figure 2: Screen images will look good with comparatively few pixels. In fact too many pixels just overfills the monitor.

Changing the image dpi
To reiterate, a digital image has no physical size, and only attains one when displayed on a monitor or printed. We've seen how size changes on a monitor, but what governs the printed size? The answer is the digital file's dpi setting. Here, dpi is a virtual concept?that applies to the image's pixels. (Note: don't confuse image dpi with scanner dpi.) The image dpi setting does not change the size of the image on the screen, nor does it change the size of the underlying image file. You see the effect of changing image dpi only when you print the image. When you click on the bottom left frame of an image in Photoshop, a box pops up representing how big the image will be when printed. This feature allows you to see that as dpi setting increases, the predicted print size decreases. Boulevard of Photographers advises you to have at least 300dpi for best printing results. This will insure that you post a strong entry on our photography contest.

Image dpi also comes into play when you import an image from one program to another, and the dpi setting of the programs is not the same. The size of the image will change when you cut and paste between the programs. The dpi setting of the image can be changed in Photoshop's Image>Image Size dialog box. This dialog box also calculates print size in response to the change in the dpi setting.

Figure 3: The dpi setting for an image does not change the digital file size but does change the predicted print size.

Printing the image
You now have a digital image that consists of a number of pixels, and the image dpi setting can be changed to alter print size. How do we determine the image's dpi setting to get the best possible print? What dpi should be set? Unlike a monitor where the number of pixels in the digital image corresponds directly with the number of pixels on the screen, printed images deal with the halftone process.

Inkjet and laser printers are halftone printers. What does this mean? Consider a typical inkjet printer. The printer may have four cartridges containing cyan, magenta, yellow and black ink. If the printer has only four color cartridges, how does it generate prints that appear to have a full range of colors?

Continuous tone color images are printed using clusters of ink dots. There are more dots in the dark areas of the image and less in the highlights, with different amounts in between creating an impression of grays. In the yellow part of an image, there will be many yellow dots; other colors are made by placing differing combinations of the four-color ink drops in a small area. From normal viewing distances, the ink droplets fuse visually, tricking the eye into seeing a continuous range of printed colors. This process is called halftoning and allows printers to use just four inks but simulate a full range of colors. The printer dots are grouped in cells, and each halftone cell can be considered a printer pixel.?A halftone cell represents a single pixel in the printed image. The halftone screen setting is measured in lines per inch (lpi) and determines the size of the halftone cell. A 53 lpi screen forms a large halftone cell, which makes the image look coarse or pixelated, while a 133 lpi screen forms smaller halftone cells with a sharper result.

There are three parameters involved in digital printing the dpi setting of the digital image, the printer halftone screen setting and the size of the printed image.

Figure 4: Inkjet printers create an illusion of grays and color using a halftone process. The halftone cell size is measured in lpi.

The image's dpi setting should be no more than double the halftone screen frequency. If the halftone screen frequency of the printer is 100 lpi, the image resolution should be about 200 dpi. Fix the dpi setting of the image based on the halftone screen of the printer, then use Photoshop to predict the print size. For a given halftone screen frequency, there is a correct image dpi and an optimum print size. The image can be printed smaller with no harm, but if it is printed larger, it appears pixelated. If you require a larger print, you need more pixels, and you'll need to start at the beginning and scan or capture the image with more pixels.

It's often difficult to determine the printer's halftone screen setting. Also, printers may use newer stochastic screening or proprietary error diffusion?techniques to print an image. If you know your printer uses stochastic screening, the image dpi should be one-third the printer dpi. In all cases 300 dpi is the pixel resolution of an image recommended for getting the best results with any Epson inkjet printer using 1440 x 720 dpi mode or higher.

Light also plays an important role in the composition. If you want to know more about light in photography, then go visit this site and this site.

Conclusion
You need to ensure that you're using enough pixels to retain all possible image detail, but at the same time, you don't want to use too many (which also has its problems). Resolution is tricky, but a bit of experimentation will help clarify the concepts. Understanding the resolution issue will be useful in the long term, as it crops up again and again in all aspects of digital imaging. Explore and enjoy!