Let’s look at the two scanners…

First of all, both scanners are top fed and scan both sides of a page in a single pass. They are infinitely faster than the scanners built into the “all-in one” printers and suitable for high volume scanning…the Epson is rated for 1200 pages per day.

The Epson GT-S50 comes with both TWAIN and ISIS drivers, giving it with a clear advantage in interfacing with software from other companies. Almost every program that can read from a scanner uses one or the other. The ScanSnap, on the other hand, uses its own proprietary driver, which only ties into their ScanSnap Organizer program…a program that displays small images of the scanned pages. Both scanners come with OCR programs, for converting scanned images to editable text, but, with the TWAIN driver, only the Epson directly interfaces with different programs – OmniPage Pro, Textbridge, or whatever else you prefer.

The Epson driver offers more choices in output resolution. Whereas the SnapScan has four levels – Normal (150dpi), Better (200dpi), Best (300dpi) and Excellent (600dpi) – the Epson has eight levels ranging from 75 dpi up through 600 dpi. For some strange reason, neither company’s software allows you to name the output file before scanning. After scanning, you’ll need to manually rename each file with something recognizable. (The default date/time stamp doesn’t give you any useful information if you need to find a file a month later).

The Epson ships with software for both Macintosh and Windows systems. I’ve tested both versions in a mixed Windows/Mac network and they both work smoothly. The Fujitsu ScanSnap, on the other hand, comes in dedicated Mac (S510m)and Windows (S510) models. I’ve only used the Windows version.

Both scanners are similar in their physical shape, although the Epson looks quite a bit larger…it’s about 1″ longer, 2″ taller and an inch or two deeper than the SnapScan. It holds 75 pages in the input slot, versus 50 in the SnapScan. The Epson has a two-line display on the front. The SnapScan only has “scan” and “power” buttons. Both units are solidly built and give the feeling that they’ll last for a long time…I’ve used my SnapScan for 30K+ pages and it’s still going strong. Both units have user replaceable pads and rollers to keep them running smoothly.

How do they operate in real life? To begin my testing, I fed the same 42 page document through each scanner. The document was complex as they come, with both text and color graphics on most pages. It was also laid out to be used in a vertical flip chart (letter-size, but bound on the short edge)…from the scanner’s perspective, the back-side of each page was upside down. My theory was that a scanner that could make sense out of it could handle almost anything…

The Epson scanned my test document faster than the ScanSnap, but software processing of the scanned image took longer. The ScanSnap took 3min 10sec from the start of the scan to PDF file creation, without using text recognition; the Epson took 4min 6sec. The OCR software runs in the background on the SnapScan, but runs in the foreground on the Epson. Because of this, the SnapScan can be used to scan the next document while the text is being recognized, while the Epson remains busy until it’s completely finished. My test document tied up the Epson for 12min 20sec from the time that I started the scan until the OCR finished and saved the file. My computer’s no slouch, but a faster computer would obviously shorten that time. If you use an outside text recognition package, you can always run it at a later time and eliminate this time waster.

The ScanSnap features automatic rotation of upside down and landscape pages. (Everything is fed in vertically on both scanners.) The Epson doesn’t…you get what you feed in. In my test document, the ScanSnap flipped the upside pages right-side up. The Epson left them upside down. Even with the upside down pages, the OCR worked fine.

The Epson can handle mixed page sizes in one scan. You can drop all your receipts in and have them scanned at once. The ScanSnap needs all the pages to be the same size in each scan…a longer page is assumed to be a misfeed and aborts the scan.

The Epson has a straight through paper path. It can handle credit card and ID Card scanning. The ScanSnap doesn’t and can’t. This makes the Epson perfect for a medical office where health insurance cards are always being copied.

The Epson can scan documents up to 36″ long. The ScanSnap stops at legal size (14″). This is useful when you are scanning cash register receipts, but as a practical matter, I seldom run into anything else that’s longer than 14″…at least not without also getting too wide for the scanner.

Using the default settings, the ScanSnap had a sharper output at 600dpi. Clicking the “unsharpen” box in the Epson software produced a scan that almost exactly matched the ScanSnap’s output…why unsharpen actually sharpens is a mystery. In solid-color blocks, the Epson was clearly superior, producing a solid-color output. The ScanSnap had a lot of off-color noise that was clearly visible when enlarged.

The Epson also has the ability to drop out colors, which is useful when scanning something printed on colored paper. It would also be useful if you need to remove a red “confidential” or “top secret” stamp from a borrowed document.

Besides the drivers and other utility software, the Epson includes PaperPort 11 SE, which is a visual file organizer. It operates similarly to the ScanSnap Organizer. I didn’t test it out…the last time I installed PaperPort 11 Pro (NOT the SE version), it messed up my computer so badly that I ended up reinstalling Windows to get rid of it. (If you have a minute to kill, go read the reviews on Amazon for PaperPort 11 Pro!) I assume that Epson’s version is more reliable, but having been burned more than once, I’m not about to try any version of PaperPort ever again! Both scanners use the ABBYY FineReader software. It works nicely. The ScanSnap also includes a full version of Adobe Acrobat Standard…a $200 program if you were to go out and buy it separately.

So….which scanner would I buy? The short answer is that it depends on your application. If you’re using both Macs and Windows machines, the choice is clear…the Epson works with both. Both scanners are built for years of use and each has unique features (e.g. card scanning) that point it in one direction or the other. The Fujitsu ScanSnap is designed to be a more “push the button and it goes” type of device. The Epson GT-S50 is more flexible. I like them both.

Update – August 3, 2009 – I’ve now had both machines running side-by-side for about five months. I now find myself almost always using the Epson instead of the ScanSnap. The reason is simple – the Epson NEVER double feeds. This may be a function of the newness of the Epson, but my Fujitsu isn’t really that old, nor, at 19K pages has it seen much service. The double feeding has become a real problem…I have to carefully watch everything going through the SnapScan to make sure that it’s not skipping a page. Not good.

Update #2 – November 14, 2009 – I upgraded one of my office computers to Windows 7 and the GT-S50 is working fine. There no drivers yet available for the ScanSnap S510…Fujitsu is estimating the end of December. Even their newer model, the S1500 won’t have drivers for another couple of weeks.
I also need to add that I’ve have a less-than-satisfactory with Epson service. My GT-S50 keeps truckin’ along, but another Epson scanner, a V500 Office started misfeeding after two weeks. Epson wouldn’t ship me a replacement unit…their scanner service policy apparently requires that the scanner be delivered, at customer’s expense, to a repair depot. That’s not exactly the most customer-centric way of doing things…

Update #3 – January 27, 2010 – Fujitsu now has Windows 7 drivers available on their website. My s510 is now up and running.