Scanning Software Keeps Improving
Major strides are being made in affordably priced systems that can scan a page of text or numbers and convert it into a computer file.
Optical character recognition software, or OCR, is used in combination with scanners, available from a number of manufacturers. The scanner makes a graphic image file of the page being scanned. The OCR software reads the image file searching for patterns it recognizes as letters, numbers, punctuation marks and other symbols.
When it works well, the result is a text file that can be used in a word processor or a data file that can be inserted into a spreadsheet or database.
The newest player in the OCR arena, Type-Reader from ExperVision of San Jose (800) 732-3897, gave the best results in my testing, especially with difficult pages, such as faxes and pages with unusual sizes and styles of type fonts. At $895, it was also the most expensive.
OmniPage Direct from Caere Corp., Los Gatos, Calif., (800) 535-7226, was just as good with clean copies of typical office correspondence and reports. OmniPage Direct is both less expensive and easier to use than TypeReader. But it can’t recognize the uneven typography in a fax.
The products were tested on Hewlett-Packard’s newest low-end scanner, the ScanJet IIp, which retails for $1,295, including an interface card for any model of IBM-compatible computer. The $300 special for OmniPage Direct is available to HP ScanJet IIp buyers. The regular price is $595.
The ScanJet IIp is smaller and lighter than its predecessor. More importantly, it includes automatic contrast adjustment technology that Hewlett-Packard calls AccuPage. Both TypeReader and OmniPage Direct are able to make use of AccuPage to obtain much better scans of pages that have colored backgrounds.
The down side of the ScanJet IIp is the difficulty that may be encountered installing it. The interface card requires access to specific ranges of computer memory. Despite installation software that is supposed to tell you how to set the board and your computer, it took a lot of trial and error to get it working on my computer.
The HP ScanJet IIp is excellent for graphics and text scanning. It is capable of various resolutions up to 600 dots per inch. Photographs, either color or black and white, can be rendered in 256 shades of gray. The ScanJet comes with excellent software of its own, PhotoFinish, which allows you to modify scanned graphics and photos in a wide variety of ways.
Both TypeReader and OmniPage Direct are Windows programs and run on IBM-compatible computers only. Both are compatible with Windows 3.1. (Both also work with most desktop scanners, not just Hewlett-Packard models.)
OmniPage Direct is easier to use because it installs in such a way that it can be used inside other programs.
In both AmiPro 2.0 and Excel 3.0, for instance, OmniPage Direct becomes a function under the normal program menu, allowing you to scan text or numbers directly into the document or spreadsheet open on your screen. Scan times per page of text were usually in the two-minute range depending on the typographic complexity of the page.
TypeReader must be run as a separate Windows program. The results of its optical character recognition are stored in a separate file, which can then be opened in other Windows or DOS software or imported into existing files in other programs.
It is able to create files compatible with AmiPro, WordPerfect, Word, Word for Windows, Excel and standard text and data files compatible with virtually any word processor, spreadsheet or database program.
Processing times were generally comparable with OmniPage Direct. But OmniPage is faster with one- or two-page documents because of its ability to scan them directly into an already open document. For longer documents, that advantage diminishes because the scanning time becomes much longer than the time involved in switching from TypeReader and a word processing program.
For clean, typeset or laser printed text in standard type fonts, both programs were occasionally capable of scanning pages with no errors and consistently capable of near-perfect scans.
Furthermore, both do a good job of matching the font and style of the original document. Italics in the original will be italics in the scanned file. Type sizes will also vary according to the original.
If you want to convert faxed documents into word processing text or data files, TypeReader is the only choice. In one test, it produced a virtually error-free file from an ordinary fax document received on a standard fax machine with roll paper. Other fax samples tested, while not error-free, were accurate enough to be easily understood and corrected.
