What's the best way to capture images: a flat bed scanner, a digital camera, or a video camera?

A flat-bed scanner is easy to use and can make very high-quality images of flat copy and graphic artwork, but most are limited to covering a standard-size sheet in one scan. Scanning very large artwork or solid objects is not practical. A digital camera is more portable and can take pictures of large objects of any shape, but the results depend on good photographic technique (adequate light and careful focusing and camera alignment). A digital video camera (or an analog video camera or VCR connected to a computer with a video digitizer) can capture still video frames, but these typically give an image size no better than 320 X 240 pixels.

Why is it that sometimes I can't open my images files - I just get a "The application couldn't be found" error message when I double-click on my file.

Double-clicking on a document lets the computer decide which program to use to open the file. It can be a convenient short-cut for opening files that were created with software than still resides on that same machine. But image files are commonly created by separate hardware (e.g. on a digital camera) or by software that resides on another computer (e.g. the computer with the scanner attached). In that case the computer can't open the file with the original program that created it, and it will either choose another program that can open files of that type, or it will give you an error message.

A more predictable way to open image files is to decide what software you want to use to view the image, then launch that program and use the Open command in the File menu. Alternatively, drag the icon of the image file and drop it on top of the icon for the program that you want to open it.

Why are image files so darned big compared to text files?

Because images are much "denser" - after all, text is mostly white space (i.e. the white paper or screen on which the text appears). Computer images are formed of thousands of closely-spaced tiny dots called pixels. Each dot takes one, two, or three bytes of memory, depending on the "pixel depth" (maximum number of colors) in the image:

Number of colors	pixel depth
256 colors (8 bits)	1 byte/pixel
Thousands (16 bits)	2 bytes/pixel
Millions (24 bits)	3 bytes/pixel

You can easily calculate the memory requirements of any image by multiplying the height times the width (in pixels) times the pixel depth. For example, a 480 X 640 image in "millions" of colors (24 bit) mode would take 480 X 640 X 3 = 921600 bytes (nearly 1 megabyte!).

1. Make the picture smaller by cropping or shrinking. Reducing a image by half reduces the memory requirement by a factor of four. For example, a 240 X 320 image in "millions" of colors (24 bit) takes only 240 X 320 X 3 = 230400 bytes (230 K). Even a modest 30% reduction in image size reduces the memory requirements by half.

2. Save in a compressed file format, such as JPEG format for photographic images or GIF format for line drawings, graphs, and cartoons. For example, a 480 X 640 color photographic image might take up nearly one megabyte if saved without compression, but if saved in JPEG format might take only 40 or 50 Kbytes of disk space (depending on the complexity of the image). A 475 X 323 pixel KidPix drawing would take 151 Kbytes if saved without compression but only 6.5 Kbyes as a GIF file. Those are big savings!

3. Adjust the compression settings of JPEG photographic images. When you save a JPG image, you will usually have some way of adjusting the amount of image compression. For example, in Paint Shop Pro, click on the Options button in the Save dialog box and use the slider to select a compression number between 1 (highest quality image and largest file size) and 99 (lowest quality image and smallest file size). Try different settings yourself and choose the highest compression number that gives an acceptable image quality.

4. Reduce the number of colors: Thousands of colors (16 bits) is enough for most photographic images. 256 colors (8 bits) is enough for KidPix drawings. 16 colors may be enough for computer-generated charts and graphs. Fewer colors often reduces the image file size and memory requirements. All image processing programs have a way to reduce the number of colors in an image.

That's because the term is used in two different ways. In some cases it means the total number of pixels in an image (really image size). For example, digital cameras are rated this way. In other cases it means the number of pixels per inch or dots per inch (dpi). The resolution of scanners and display devices such as computer screens and printers is rated in this way.

The resolution of computer screens is only 70-90 dpi (dots per inch), so if you scan a picture at more than that resolution, it will look big when all of its pixels are displayed in another program. Solution: scan at 75 dpi. (You can calculate the dpi resolution of your display by dividing the horizontal screen width in pixels by the physical horizontal width of your computer screen measured with a ruler: for example, 800 pixels/11.5 inches = 70 dpi). Most computer models allow you to change the numbers of pixels on your screen: (Macintosh: Control Panel > Monitors > Resolution. Windows: Control Panel > Display > Settings > Screen area).

It is sometimes useful to create an enlarged image of a small graphic (e.g. a postage stamp) by scaning at high resolution (e.g. 300 dpi). When displayed in other programs, the resulting image will be larger than life and will show all the details.

I took a photo with my new digital camera and emailed it to a friend, but my friend said the picture was way too big.

Modern (2 and 3 megapixel) digital cameras produce images that are 1600 - 2000 pixels wide. but most peoples' computer screens are only 640 - 1000 pixels wide. Solution: reduce the image to 640 pixels width or less so it will neatly fit on their screen.

I scanned a photo, but when I zoom in on small details it looks all "blocky".

Those blocks are the individual "pixels" in the original image. You may need to re-scan the photo with higher resolution (150 or 300 or even 600 dpi) in order to preserve the fine details in the picture.

Some formats are platform-specific (PICT on Macintosh and BMP on Windows) and some are platform-neutral (TIFF, JPEG, and GIF). Some formats include file compression (JPEG and GIF) and some save without compression. Some programs accept only graphics in certain formats (e.g. The Writing Center on Macintosh takes only PICT format). Web pages need graphics in either GIF or JPEG format. GIF format is best for hard-edged computer drawings (e.g. KidPix pcitures) and charts and graphs. JPEG format is best for photographs. If you use the wrong format, your graphics will look poorer and take longer to load (some examples). You can use a shareware graphic utility program (such as "Graphic Converter" for the Mac or "LView" or "Paint Shop Pro" for Windows) to convert from one format to another by Opening in one format and Saving in another.

The larger the image file size, the longer it takes to load. Also, the larger the number of graphics on one page, the longer it takes the whole page to load. Therefore it's a good idea to avoid putting too many pictures on one page - break them up between several pages.

You should seek to reduce the file size of your Web graphics by (1) making the picture smaller by cropping or shrinking, (2) using the proper file format. For digitized photographs, JPEG format is best. JPEG images can be saved with different amounts of compression. The greater the compression, the smaller the file but the poorer the image quality. Here are some compression examples. For computer-generated drawings, like KidPix drawings, GIF is the best. Here are some examples of JPEG vs GIF. In some cases reducing the number of colors of simple images can reduce file size substantially.

It is not a good idea to use a Web page editor or HTML editor to re-size graphics. It is better to re-size the graphics beforehand using a graphics editor. Displaying an image smaller than its real size does nothing to speed up image loading because the file size in not reduced, only the displayed image size.

Finally, avoid putting large numbers of graphics in a single table, because the entire table has to load before it is displayed on the Web page. This creates a very annoying blank-screen delay for the user. If you must use tables, break up large multiple-row tables into several separate tables of one row each. That way the first rows will be displayed while the later rows are loading. Here is an example.

Position the mouse pointer on the picture, hold down the (right) mouse button, and select "Save image as..." from the pop-up menu. Graphics will be saved in their original format (JPEG or GIF) but can be changed into other formats by using a shareware graphics utility program.

Try using a "screen capture" utility such as HySnap for Windows (which is shareware) or the Capture function that is built into Paint Shop Pro version 6 or higher. This allows you to capture the image of any program's windows, menus, dialog boxes, or even the entire screen.

On the Mac, there is a built-in screen capture: press Command-Shift-3. This will save an image of the entire screen on the hard disk as "Picture 1". "Picture 2", etc., in PICT format. You can crop it to the portion that you want using Graphic Converter.

Enlarging a image in software just makes the pixels larger; it does not create more pixels or improve the resolution. Take (or scan) the picture again, but this time use a higher scanning resolution or zoom in on the part that you are interested in.

The first thing to do is to check you monitor color setting. (Macintosh: Control Panel > Monitors > Color depth. Windows: Control Panel > Display > Settings > Color Palette). Set to the highest color setting you can for the best color display. If the display is set to too-few colors, all photographic images will look grainy and splotchy, no matter how good the picture actually is. (Unfortunately, some programs - like the older versions of KidPix - force you to use 256 color mode.)

Don't forget that your monitor has contrast and brighness controls. You may need to adjust those controls if everything looks wrong.

If you created the picture on a Mac and it looks too dark when viewed on a PC, or if you created the picture on a PC and it looks too light and washed-out when viewed on a Mac, then you probably need to adjust the "gamma correction" of the image. You can do that in Graphic Converter, LView, or Paint Shop Pro.

All of the shareware graphic utility programs will let you adjust the contrast, brightness, and color balance and saturation, which can sometimes improve an image. You can even make images look slightly sharper by using the sharpen function.

However, it may be that your image was flawed to begin with, perhaps scanned with too low a resolution or with two few colors. In that case your only choice is to scan it again with higher resolution and more colors.

No. Photoshop is a heavy-duty professional program that is expensive and requires considerable computer hardware. You can do most of the routine image editing tasks using a shareware graphics utility program (such as "Graphic Converter" for the Mac or "LView" or "Paint Shop Pro" for Windows). Another good low-cost program is Adobe PhotoDeluxe, which is a simplified and less expensive ($49) version of Photoshop for home use.

Save the image in a format that the receiving program accepts. You may find it useful to use a graphic utility program that can Open and Save in many different formats (such as "Graphic Converter" for the Mac or "LView" or "Paint Shop Pro" for Windows).

Try Copy and Paste. Sometimes you can just select a graphic in one program, Copy it to the Clipboard, go into another program, and Paste the graphic in.

Buy a newer printer and use better paper. Ink jet printers have improved greatly in recent years; they are now better than color laser printers for photographic printing. For the best results with photographic images, look for a model with "photo-enhanced" printing - variable dot-size and/or six-color printing. Printers often have two or more settings for print quality - a faster, lower quality setting and a slower, higher-quality setting. Print quality is selected in the print dialog box. For example, on a Windows machines you click on the Properties button. With Hewlett-Packard Deskjet printers, the quality selections are Best, Normal, and Econofast. Use Best or Normal for printing photographs; Econofast for printing Web pages and text documents.

Paper selection is also very important for ink-jet printing - much more so than for laser printing. Use coated ink-jet paper rather than plain copier paper, and for the best results with photographic images, use a paper with a glossy or semi-glossy finish, such as Kodak Glossy Inkjet Paper (#17416); this particular paper also has the advantage that it is coated on both sides, so you so you can print on either side or both sides. (Most other coated paper is coated only on one side, and you must print only on the coated side, meaning that you have to load the paper properly into you printer so it prints on the correct side).

For best results, it's important that you select the type of paper you are using in the print dialog box. The print dialog box for your printer will have a pop-up menu of paper types (plain paper, inkjet paper, glossy photo paper, transparencies, etc). This selection influences how much ink is deposited on the paper. Make sure this selection matches the actual paper type you are using. Either follow the recommendations of the paper manufacturer or make you own series of test prints with various papers and paper settings and choose the combination that you like best. Don't neglect this; it can have a major effect on the quality of your photographic prints. Even low-cost ink jet printers are capable of good photographic image quality with the right paper and printer/paper settings.

---

Sources of Images and Image Processing Software


---

Tom O'Haver
Professor Emeritus
Maryland Collaborative for Teacher Preparation
The University of Maryland at College Park
toh@umd.edu
http://www.wam.umd.edu/~toh