I've been waiting for an affordable scanner to digitize my hundreds of great shots yearning to be set free from their archival preservers. So just for the heck of it, I recently brought a few 120 and 35mm chromes into a service lab to be scanned to have a benchmark for comparison. At 1440 DPI, the results looked impressive enough to print a few 8 x 10s on my Epson 820. They look nothing short of amazing.
So, with the number of 2400 x 4800 scanners coming out recently at bargain basement prices, I decided to jump in for a cost of about $230. After receiving delivery of my Microtek 5900 a second time (the first unit was defective), I was a man on a mission. The hardware and software installation went smoothly - it comes with Adobe PhotoDeluxe 4.0 for Windows, and Photoshop 5.0 LE for the Mac, amongst other freebies. I own a Windows machine with 2000 Professional, and there were no issues getting the scanner up’n running.
With alot of anticipation, I dove right into scanning 35mm and 120 prints, negs, and chromes, spending hours and hours trying to get the scan I was originally impressed with, expecting “2400 x 4800 DPI” to deliver. But everything looked out of focus both on-screen and from my ink-jet printer. Playing with every setting in the book, every permutation and combination of options settings, I tried over and over to duplicate what was delivered from the service lab - a tack-sharp image.
To compare apples to apples, I scanned one of the same 120 chromes mentioned earlier at the same DPI as the lab’s scanner. I realize it is recommended that the DPI setting be a mathematical factor of the scanner’s optical resolution, nevertheless I wanted to get a close approximation of the quality. The two images are incomparable. The scan from my flatbed is still out of focus in comparison to the detailed and sharp image from the service lab. Scanning positive and negative 35mm film strips yielded worse results because of the smaller image, even setting the DPI to the manufacturer’s recommended settings (300/600/1200, etc).
More testing: I did multiple scans of another 120 chrome that has good tack-sharp detail, containing such things as lettering on an old antique gas pump. Starting at 300 DPI, I then scanned at 600, then 1200, and finally 2400 DPI. Seemingly the optics on the scanner are only good up to 1800 or so. There's very little difference between the 1200 and 2400 DPI scan, except for the file size: the 1200 scan is 6 megabytes compared to the 2400 scan at 25 megabytes! I also tried saving in different formats such as gif, tiff, and jpeg. Still no improvement. Even more fundamental to this is the lack of sharpness in the image itself: edges of letters are fuzzy enough that even sharpening the lightness channel after converting from RGB to LAB colour didn't improve the image to any great degree. (See article “Sharpening and the Unsharp Mask Function” by Ian Neeson, page 5, May 2002 edition of Photolife Digital.)
Product support at the company's help-desk viewed the comparison between their scanner's output and that of the service lab's, and said I was expecting too much from a $230 scanner. They stated the output from their scanner was up to snuff, so I had indeed squeezed out all the performance possible from this model.
Although technology does progress at an enormous rate, I was naive to think a $230 investment would return $4,000 worth of results. Trying to make a silk purse from a sow’s ear never did work! Perhaps a more accurate analogy would be to expect a 3” telescope to resolve a distant binary (double) star less than 1/10th degrees of arc. It simply doesn’t have the resolving power, resulting in a less-than-perfect image. In the end, it seems that although the DPI is high, the optics fall short of delivering a crisp image to the CCD.
So if your goal is to digitize images from print, negative, or transparencies for quality printing, do not bother with cheap scanners claiming high resolution such as 2400 x 4800 DPI. It is my suspicion that any scanner priced under $300 will deliver similar results. They simply do not have the same quality optics as the more expensive dedicated units. The service lab that scanned my chromes said that flatbeds starting aroung $500 CDN aren't too bad…