Bitmap vs.Vector
Image Formats

Making a home page on the Internet can be fun and exciting (sometimes even profitable). To make the best home page, you will need to have the best tools. A scanner is an important part of making a distinctive and recognizable home page and Mustek is one of the most affordable scanners around. Before you go out and buy a $1000 scanner, look at what exactly you will need in terms of resolution and color depth.

Here are a few guidelines to help scan better pictures. Some ideas of things to scans are pictures of people, real buttons for web pages, and corporate logos.

Resolution

Unlike money, hard drive space and processor speed, with resolution, less is more. Graphics on the Internet take forever to load. That's because people often scan at resolutions and color depths much bigger than they need.

Most scanning software tells you what your scan is going to be in inches and dots-per-inch. A typical scan might be 5" x 5" at 600 dpi, which would create an image 3000 x 3000 dots. Standard VGA supports a 640 x 480 resolution, better video cards 800 x 600 and 1024 x 768 resolutions. A picture of 3000 x 3000 wouldn't fit and would take 70 times longer to load than the same image at 72 dpi! Web publishers don't think in terms of inches and dpi, they think in terms of pixels.

To fill approximately half of a standard VGA display, the graphic would need to be around 320 pixels wide. Lower the resolution until the picture is approximately 320 pixels wide. On a 5" picture that would be around 60 dpi. If the resolution doesn't go low enough to scan an image 320 pixels wide, scan a 640 pixel wide picture and use the software to scale the image.

Number of Colors

The rule of thumb when displaying pictures on the Internet is the less information you can give and still have it look good, the better off you are. Lets say you scanned your company logo which has four colors, green, gray, black and white. You could easily save this as a 16 color file. If you were to save this as a 24-bit color file, it could be as many as 6 times bigger, display no better while still taking much longer to load.

When scanning pictures of landscapes, people, and even product shots, the number of colors is important. 256 colors wont display these types of images very well, for optimal display, 24-bit color is needed. Does a 30-bit scanner help? Yes, 30-bit scanners help get the best possible 24-bits (assuming your picture is either too dark or too bright).

File Formats

There are two different file formats popular on the Internet, both are very different. The question isn't, "Which file format is better?" That question is much like, "Which is better a hammer or a screw driver?" It isn't so much which is better, its which tool should I use for which job. Use the hammer to drive nails, the screw driver to remove the screws. Warning: using a hammer to remove the screws from your computer could void your warranty. ;^)

GIF File Format

The GIF file format is a lossless file compression format. Information on areas with similar colors are compressed. For example most of our company logo is white and could be compressed in the GIF format, where a picture of a person has subtle tones which could not be compressed.

Some GIF file formats support a handy feature called interlacing. Interlacing saves the file with the odd line information first, then the even line information. This allows the user to get a good idea what is coming up, so they don't have to wait until the entire button is shown before moving on. These graphics don't really load any faster, they just seem to.

Another handy feature of GIF is the ability to save transparency information. Round buttons would have background areas that you would have to specify for each different background; however, with GIF and transparency, the browser can automatically overly it on any color or textured background.

The software shipping with Mustek scanners do not support these two features, in fact even very expensive software like Adobe PhotoShop doesn't directly support it. However, your scanner works with other image editing software which does. Some of which are even available for free on the Internet.

GIF files are usually 256 and 16 color files called "Palette Color." So you wouldn't want to save images that had many more colors. The GIF file format is most often used when saving buttons and other small items, because they have very few colors and lossy compression would make small text unreadable.

JPEG File Format

The JPEG file format is a Lossy image compression format. It was designed by a group of very smart people to compress pictures of natural objects. It does some fancy mathematics to allow images such as people and flowers to be compressed from sizes like 1 megabyte to sizes like 50K. When compressing an image that much, some information is lost. Pictures of people and flowers aren't changed much with JPEG; however, items like company logos often look fuzzy and distorted.

JPEG has a cool option called Progressive rendering. Its much like GIF Interlacing but its much fancier and looks more interesting when its loading. Very few programs support progressive JPEG (for example Adobe PhotoShop 3.0.5 doesn't support this feature), but if you have one that does, it is a great addition to any Internet toolbox. Some handy utilities exist on the Internet for converting images into Progressive JPEG.

Because JPEG saves files with 24-bit color information, use this file format for picture of organic items like people and landscapes and use GIF for things like buttons and logos.

Optical Character Recognition

When converting from paper to HTML, you don't want to use programs which offer format retention. For best results, it is best to scan the text and lay it directly into an HTML editor. The formatting used by most documents is hard to undo in most word processors and isn't support by the HTML standard (i.e. multiple columns and snaking text).

The OCR software that ships with Mustek scanners offers an excellent recognition rate and doesn't retain non HTML formats like multi-column and snaking text.

For digital display

When scanning text for display, do not scan at high resolution. The lower the resolution the more text you can fit on a page. Text is very readable at 72 dpi when scanned in grayscale. Don't scan entire documents and just put them directly onto your web page. A single 300 dpi line art page takes one megabyte which can take 10 minutes to display on some Internet connections.