The maker, Scansoft, claims that version 11 has the most accurate OCR engine technology with new functionality such as user-friendly PDF input/output capability. Also claimed is "IntelliTrain" Proofing, which uses previous corrections to improve OCR results and reduce correction errors. I intend to fully test these claims and others that are touted in this version and give you my best effort in evaluating them.

My main interest in this program was the fact that I am involved in desktop publishing. I edit and publish an eight-page newsletter for a retiree organization. The inputs that are sent to me are usually typewritten text. I scan and insert these into my word processor for editing and other formatting and corrections. I then put these text files into my desktop publishing program. OCR, or optical character recognition, is therefore an integral tool that I use extensively.

Installation was very conventional. The installation program easily recognized my brand of scanner (MAG Model 3050), even though it was not listed on the supported scanner list. It also recognized my word processor (Word 97) and activated the automatic OCR from within it. Just to make sure, I ran the scanner setup wizard to verify that all of the pertinent data was there, and it was. I then ran the scanner tests, and it passed them and informed me that the program was ready to go.

There are three modes of operation: Auto OCR, Manual OCR and OCR Wizard. The first two methods can be customized for your personal requirements and the complexity of the scanned material through a series of settings. The OCR Wizard, on the other hand, asks a series of questions and depending on your choices, sets the various parameters automatically.

For the majority of users, this would be the user-friendliest method of operation and it works very well. You can also schedule OCR jobs to be performed automatically with the "Schedule" wizard, which allows you to specify settings and starting time for unattended operation.

For complex documents with text and graphics, the program has zoning tools and templates which allow you to customize properties and the output of the job. One oddity of the installation is that it does not ask to install an icon on the desktop, which necessitates running the program from the Start Button. This can easily be fixed by creating a shortcut and dragging it to the desktop.

The Scansoft Web site is very helpful in curing installation problems. (There were very few). The online FAQ section has 165 typical problems that can occur within this product. The manual is a concise 92 pages, of which 66 tell you all there is to know about running the program in each of the modules.

The main desktop displays your scanned document in a thumbnail view, an image view, and a text view, identical to version 10. After proofing and editing in the image view, questionable words and items not recognized by the OCR are highlighted. The scanned document is then placed into your word processor in an editable form. The formatting, images, and other objects are maintained, depending on the preferences set before the scan.

Better OCR recognition? I would say yes in a minimal sense. The literature claims a 40% decrease in word recognition errors. This is hard to measure exactly, although I did see a small improvement in word recognition.

PDF input and output? Since PDF format is the standard of the industry with regards to Web printing and reading operations, this upgrade or program would be a welcome addition to people in the industry. It can convert between PDF and Microsoft Office and 20 other formats.

However, you would still need the regular Adobe Acrobat program, version 3.0 or later, to do many of the Web operations connected with the PDF format. The program also saves the output in 30 additional formats, covering most of the major programs used in today's computing environment, including HTML.

It recognizes images (read text) in all of the major languages in the world. The voice read-back feature has been improved, incorporating the Microsoft voice engine used in the new XP operating system, which was released in October 2001. It can read back in all of the major covered languages. This feature was also in version 10, but without the extensive language coverage.

All of the other reasons given are, in my opinion, cosmetic improvements and most of these features, although improved here, are available in version 10. One example of cosmetic change, as touted in their literature, is the so-called "despeckle module." After spending a lot of time looking for this module to test and evaluate it, I could not find it. Tech support advised me of the following: "The despeckle module is actually built into the program. It is not a setting that you can select. There isn’t a way to turn it on, because it is always working." You have to take it on faith because honestly, I did not see any evidence of this feature working.

The program does a good job in the WYSIWYG (what you see is what you get) category, somewhat improved over version 10 in maintaining the look and feel of the original document.

For these reasons, I would say that if you are a professional user with a heavy volume of OCR and Web requirements, then this upgrade is highly recommended and has all of the "bells and whistles" you would need. If you are a new user, this OCR program is probably the best available today for the price. For the hobbyist or home user with version 10 and with infrequent need of OCR capabilities, I would not upgrade, since the version that you have does 95% of what version 11 does.

If you have basic computer skills and the required hardware, you should have no problem learning and using the program. The OCR accuracy factor as stated previously was much better than Textbridge 8.0, the principal competition, and somewhat better than OmniPage 10. The claim of 99% plus accuracy is borne out generally on a good original copy, which is true in both versions 10 and 11. Graphical representations, such as tables and graphics, the so-called "true page" representations, were much better in Version 11 over the other versions that I have mentioned. The integration with word processors, spreadsheets, and other Microsoft Office programs is very good. When all is said and done, this is a good OCR program. Did the time you spent fixing recognition and typographical errors take less time than you would have spent retyping?

Price: On various Web sites the price ranged from $103 to $129. CompUSA had it for $130, Fry's $100 for an upgrade to version 11. Scansoft's upgrade price is $90 from version 10.

OmniPage Pro 11
Scansoft Inc.
9 Centennial Drive
Peabody, MA 01960

Hyman LaVere is a member of the Los Angeles Computer Society, from whose newsletter, User Friendly, this article is reprinted.