By Al Giovetti


Support The Computer Show and get paid to surf the web. Click on this Big Bang ad!

How to buy a Scanner and not make an idiot of yourself

by Al Giovetti

Just one short year ago many scanners were priced at $1000 and more. They were priced right out of the budget of most computer users. Scanners prices have become more manageable following the trend of decreasing prices and increasing features. Many consumers and businesses are finding scanners affordable solutions to life's and business' challenges.

Just looking at the adds and boxes for scanners may make you think the buying decision is a simple one. It is not this is a computer peripheral. You need to know more than the words compact, color, flatbed, and easy to use. You, once again, need to make sense out of an army of acronyms and jargon, like OCR, ASCII, USB, SCSI, bits, dpis, and lets not forget TWAIN.

Scanners have a variety of different categories. One category is the way the scanner handles paper, which divides scanners into the flatbed, handheld and sheet-fed types. The flatbed scanner can only take one sheet at a time, unless it comes with or you can purchase an optional sheet-feeder. The flatbed scanner has emerged as the scanner of choice out of the three.

Flatbed scanners handle image documents better than text while the sheet-fed scanners work fine with text. Flatbeds do not distort the image like many sheet-fed scanners do. The optical distortion in sheet-fed scanners comes when the paper is bent when it moves through the scanner or paper handling rubber rollers.

Sheet-fed scanners are generally smaller, and they get their names from the way they can feed in a stack of papers automatically. The sheet-fed scanners can jam and they will not work with all types of paper, and books are impossible to scan with them. Thinner paper stock can be destroyed by the rollers and metal clips used to move the paper. The sheet-fed scanners used to be much cheaper than the flatbeds, which are now the most convenient and more desirable type.

Of the three categories the handhelds are the hardest to find. Handhelds were inexpensive alternatives to the more expensive flatbed and sheet-fed varieties. Since handhelds were awkward and difficult to use, they are quickly vanishing from the marketplace.

Most computer scanners are different than those used for copy machines. Copy machines use lenses and mirrors to transfer the image to a drum that picks up magnetized plastic ink to form the image. This method tends to be more precise and yield a much better image than a computer scanner. The computer scanner has to digitize the image that it picks up electronically, and depending upon the quality of the scanner, some of the image may be lost when it is printed.

Many people who bought scanners to be used with their computer printers as a cheap alternative to copy machines, were disappointed when the image produced was inferior to that obtained from an optical copy machine. Some of that technology has been improved in the more modern and more advanced digital scanning devices. The more information collected in terms of lines and dots, the better the picture, making some of the newer scanners approach the quality available in the optical scanners used in copy machines.

Another issue is the output of a digital scanner. The output is divided into bitmap pictures and ASCII (American standard code for information interchange) text. Some scanners are made to take a picture of the page which is reduced to a bitmap of the page. Bitmaps are known to take up 50 to 150 KB per page of disk storage. Bitmaps can be in color or black and white and can be converted to other file types such as GIF, tiff, or JPEG used on the Internet.

(Skip over this more technical discussion if you don't want to get buried in jargon.) A bitmap is a binary representation of picture where bits or a set of bits from a computer represent part of an image or a font from letters or characters. When using a monochrome scanner one bit (binary digit) in the computer represents on pixel (picture element) on the screen. A bitmap is usually associated with graphics, in which the bits are a direct representation of the image.

Other scanners convert the image data into ASCII text through Optical Character Reader (OCR) technology. Once converted the text files for an entire page are significantly smaller, about 25 times smaller than one page of bitmap information. A page of ASCII text takes up 2- 3 KB per page of disk space. Typically computers that perform OCR functions well, do not do good bitmaps and visa versa, further compounding the selection process.

(Skip over this more technical discussion if you don't want to get buried in jargon.) OCR or Optical Character Recognition analyzes the light and dark areas of the bitmap that is produced when you scan a document into your computer. The OCR software then converts the areas of light and dark in the bitmap into letters and digits. Recognized characters are converted to ASCII text.

A problem with OCR technology is that it does not recognize 100% of the characters (letters and digits) of a text document when scanned. ASCII text produced by OCR software must be edited by a human being to make sense out of what results. In spite of this deficiency, OCR can save you from retyping 95% of a document.

Additionally, some type faces, especially the smaller ones cannot be converted by OCR software. The resulting ASCII text does not resemble the original document. OCR scanners work best with familiar type faces like courier and larger type such as that which would come out of a typewriter. Often newsprint and magazine print is too small to use with OCR software.

ASCII text can be edited and reformatted in a word processor or publishing package, while bitmaps of text cannot be manipulated. Bitmaps of text are virtually useless for anything but archival storage. OCR of pictures is even more useless in that you will get gibberish instead of your picture when restoring. So it becomes important that you pick your scanner based upon its intended use.

Some of the problems with OCR software and scanners can be resolved with higher optical resolution. Scanners are rated with optical resolution, which is the actual resolution of the scanner, and with their digitally enhanced resolution. Optical resolution should be the measure of the machine when comparing prices not the digitally enhanced resolution, which is called interpolated resolution. Interpolated resolution is where the image is enhanced by the software that comes with the scanner.

Interpolated resolution is the enhanced resolution of a scanning device and not the "true" resolution. The image is enhanced with a software algorithm. Images appear as if scanned at a higher resolution. Interpolated resolutions are often significantly higher than optical resolutions. Both the optical and the interpolated resolutions are given in dots per inch or dpi. You need to find out if the 2400 dpi resolution shown on the box is optical or interpolated before comparing prices.

Another comparative feature is the number of bits supported by a scanner. Scanners are rated at 8-bits, 24-bits, 30-bits, and 36-bits. The more bits the more image can be handled and processed. The scanners with the highest bits are generally of the best quality, all other things being equal.

Bits relate the number of colors that a scanner can interpret or scan. An 8-bit scanner can scan 256 colors, while a 24-bit scanner can scan 16.7 million colors. The more bits the more colors. The more colors the better the picture. When looking at a black and white scanner the bits relate the the shades of grey supported.

The biggest mistake that scanner buyers make when selecting a scanner is to get one with too small optical resolution and too few bits. Most scanner buyers are back in the store within three months buying another scanner with higher capacity in these areas. Make sure you buy a near the top of the line scanner in these areas. Software can always be upgraded more cheaply than hardware.

Scanners often come bundled with software. Almost all scanners have the image enhancing software of interpolated resolution. Sheet fed scanners and some black and white scanners come with OCR software bundled in. Flat-bed and color scanners often come with image editing software.

Many of the software packages are watered down versions of the full-blown commercial version of the software. Be sure that you understand the features of the included software before you upgrade to the full-blown commercial version. Many commercial software packages today have a plethora of available features. Most computer users cannot and do not use more than 5% of those features, so that the included "Casper milquetoast" version of the software may be fine for your needs.

Scan rate is the speed at which a page of paper is scanned. Scan rates are usually set at pages per minute. or seconds per page. Scan rates can become very important when you are sitting and waiting for a document to scan or if you have many documents to scan, especially when you use the flat bed type of scanner for multiple documents. Just think of how boring and time wasting feeding on document at a time into a copy machine is and you will get the picture on scan rates.

Another important feature to understand is the footprint, which is the size of the scanner and the room it will take up on your desk or computer furniture. Flatbed scanners have the largest footprint, but are often worth the space. Sheet fed units take up very little desk space and may be fine for those who primarily need the scanner for OCR or text conversion to ASCII use.

Ease of installation and use is another important consideration. Scanners attach to the computer via three most popular routes: the Parallel port, the SCSI interface and the serial bus. The parallel port usually has a printer attached, which must be detached and reattached unless you have a bypass device. The SCSI (Small Computer Systems Interface) port which requires an interface board be installed with an external port, which can be expensive, time consuming and require much expertise. The serial bus, which should be the new Universal Serial Bus or USB port or you have an out of date scanner.

Of the three the USB port is the best option. USB or Universal Serial Bus is a hardware interface for peripherals. Examples of peripherals included the keyboard, mouse, joystick, scanner, printer, and modem. USB has been designed to support MPEG-1 and MPEG-2 digital video. The maximum bandwidth or speed at which information can be exchanged is up to 1.5 MB (megabit) per second and up to 127 devices can be attached.

The SCSI port is needed for those who want the fastest speed available. SCSI devices run at a date transfer rate of 10 MB per second. SCSI devices are reliable, but are difficult to install and often require software drivers which may not work without help from the vendor's call in support line.

And no discussion of scanners would be complete without a discussion of TWAIN compliancy. All scanners should be TWAIN compliant, check for it on the documentation when you buy. TWAIN stands for Technology Without An Interesting Name. I think that is enough for this installment of "Make Technology Simple" which we call MTS. So get your TWAIN and MTS ducks in a row and I will return in the next issue for more MTS. Bye.


  1. From: FIREFED@aol.com
    Gents, this is excellent information. I'm in the market for my first scanner. I have seen everything you wrote about in your article and now I think I understand part of it at least. What I think I see out there now is a decline in price on most scanners but not one on them is USB compatable. I think I will have to wait until the current inventory is exhausted and then I'm sure we will be seeing USB scanners. I hope I'm on the right tract.
  2. From: pcc
    Subject: scanners: GOOD JOB!












Please send us your comments and suggestions.