Omnipage Professional 17 Review
Omnipage Pro 17 is a powerhouse OCR program that can perform almost any OCR task at top speed, with options that let you either automate the whole OCR process for hands-off convenience, or fine-tune and correct each stage of the process for the most accurate possible results. The latest Omnipage version builds on two decades of experience with OCR, and I’m more impressed than ever with its power, flexibility, and speed. On large operations, Omnipage Pro 17 ($499 direct) tended to be faster than either ABBYY Finereader 10 or Readiris Corporate 12. Omnipage’s interface, while improved since last year’s review, still retains some niggling frustrations and costs it the overall Editors’ Choice—by a hair. If you’re looking to do high-volume, highly automated OCR scanning, you’ll want to give Omnipage a look.
If you plan to use Omnipage mostly for corporate-level automated processing of stacks of documents that you feed into a scanner, you’ll be impressed by the speed of the operation and the quality of the results. But if you plan to use it to process documents that need their own fine-tuned settings, you’ll be impressed by the quality of the result but annoyed by some of the steps you need to perform to get there. In terms of output quality, Omnipage comes very close to our Editors’ Choice, Finereader, and even surpassed it in one of my test documents. But Omnipage’s interface is too confusing for an Editors’ Choice rating—more on this later.
I tested Omnipage Pro Professional, which includes an advanced barcode processing feature that lets you fine-tune the kind of processing that the app performs depending on the barcode that you use on the first page of an OCR job. Omnipage handles barcode processing impressively well, complete with an option to print out barcode pages directly from the app. The less-expensive Standard version lacks this and other corporate-level features, such as preset controls for formatting legal documents.
Unlike Finereader and Readiris, Omnipage doesn’t offer a start-up menu of common tasks such as scanning to PDF or Word. Instead, you can choose a Quick Convert view from the Window menu and use a set of drop-down menus to select options. You use one drop-down to choose whether to load an image file or using a scanner or digital camera as the image source, another drop-down to decide whether to let the app detect the page layout or whether to specify in advance a layout such as single-column text or spreadsheet format. Similar drop-downs let you choose your output format—Word, PDF, or other standard formats. The dialog is cleanly designed but clumsy to use because you need to navigate drop-down menus to see every option, even though your screen is probably large enough to display modern, easy-to-find buttons for the options you need most.
If you don’t use the Quick Convert view, you’ll use one of Omnipage’s prebuilt, customizable “workflows.” These are sequences of steps that start with getting an image from a file, camera, or scanner; then proceed through recognizing layout, optionally correcting spelling and other errors, and outputting the result to a file. I was able to build a custom workflow by modifying one of the four that come with the app. I added a step that let me review and correct the program’s automatic detection of text and graphic fields, and added another step for that let me manually proofread the output for errors.
So far, so good. But once you’ve begun a workflow that processes a document, you can’t easily go back a few steps and correct problems that arose earlier. I sometimes let the workflow proceed to the proofreading stage, when I noticed that the app had misread a text region in my document as if it were a graphic. So I went back to the layout-selecting pane and removed the mistaken graphic box from the image of the document, and drew a text box instead. Even though the app let me draw the text box on the image, it wouldn’t perform OCR on it, because the workflow had already proceeded to a different step. I had to abandon the whole operation and start over.
Another minor but highly irritating problem crops up in the proofreading dialog that displays suggested alternate spellings for doubtful words. Each suggested alternative is preceded by a number and a parenthesis, for example, “2),” but if you type the number 2 in the logical and intuitive belief that you can select this alternative by typing its number, the program doesn’t select the word. It actually enters the number 2 into your text. If you want to select the alternative word with the number “2″ next to it, you have to click on the word. Both methods ought to work. I’ve been complaining about this for years, and for years Nuance has told me that no one has ever complained, but they’ll consider changing it in the future.
Power Beneath the Skin
Beneath its interface, Omnipage is impressively powerful. In my ultimate test document—an early 19th-century legal textbook—it detected printed sidebars on some pages and correctly interpreted them as text boxes. Not even Finereader could do that automatically. It lagged behind Finereader in its ability to read a fiendishly complex table in a PDF file, and produced output that would have needed hours of work to turn into an editable table. Some of its best features needed to be switched on from deep in obscure dialog boxes, however. For example, I needed to switch on a “Look for headers and footers” option before Omnipage would treat page numbers in a scanned book as part of a page header instead of a paragraph on the page.
I’m impressed by Omnipage, and would be even more impressed if its interface were a bit more intuitive. As a bonus, it ships with the same vendor’s PaperPort 11 document-management software and PDF Create software for creating PDF files from any Windows application (not tested for this review). The interface alone keeps it from sharing the Editors’ Choice rating—that honor goes to Finereader 10. But it’s a much closer race this year, and Omnipage’s bundled software, its impressive speed and power—especially when it comes to high volume, highly automated OCR scans—could well make it your top choice even if it isn’t mine.