User:Jhellingman/Scanning

Jeroen's Scanning Tips

Scanning books is one of the content provider's job. It is also the start of the DP production chain. Doing the job right is important for all following stages. To do the job right, I have collected a few tips for scanning books with flat-bed scanners.

Avoid duplicate efforts. Before you scan a book, check the various on-line sources for scans. A scan-set found on-line can save you several hours of work (and if you have the book, you can add any missing or bad pages as you go). Note that for illustrations, I often prefer to scan myself on a flat-bed scanner: the scans made with overhead scanners tend to be distorted somewhat, and the lossy compression used on the images makes image processing much more difficult.

Beware of dust. Scanning old books can be a dirty business. Books are often dusty, and little pieces may break of brittle pages. These may all end up on the glass of your scanner, and obscure parts of your scans. Have a piece of soft cloth ready to clean the glass regularly. Use a vacuum cleaner regularly to keep your surroundings clean, but be careful with such big machines on delicate scanner hardware.

Get it flat. To produce good quality scans, the book has to be completely flat on the scanner. This may be problematic for tightly bound books, and pushing too hard may either damage the book or the scanner. Sometimes, you can get a page flat by putting some item of appropriate thickness behind it. (I use various pieces of cardboard for the purpose). If you scan a lot, consider a specialized book scanner, such as the Plustek OpticBook.

Reduce Shine-through. Sometimes, the text on the other side of the page shines through, especially with thin paper or cheap printing procedures. This effect can be limited by placing a black sheet of paper behind the page you are scanning.

Optimize Resolution. Most books can be handled adequately by scanning them at 300 DPI, which can be considerably faster to scan than 400 or 600 DPI. So, if your book allows it, scan at 300 DPI. For books that need it, because of small print or odd diacritics, you should use higher resolutions, though.

Optimize Brightness. Paper can vary from bright white to deep brown, and printing inks from pure black to gray, and older books may have brown (foxing) stains on the page. If you scan in black and white, as will be most common when scanning for DP, you will need to set the threshold level carefully, such that all letters (including for example the horizontal bar in the letter e) appear clearly, while foxing stains still remain white. A few books may require page-by-page adjustments.

I have scanned a few books that where both the pages where considerably discolored, and the ink used faded. The only way I could handle this was to scan the entire work in gray-scale, apply a high-pass filter in Photoshop to make the stains go away, and readjust the levels to make the end result black-and-white again.

Jeroen's Scanner Reviews

I've used a large number of scanners for scanning books, which I will shortly describe here.

HP ScanJet 3c/4c

A 15 year old workhorse. I've owned 3 of these. One broke down with mechanical problems (scanner carriage end-detector failed), and another caught a severe case of misaligned colors, but I still like them, and can still be had second hand for a small price. The major drawback is that they require a SCSI connection, which is not standard, the major benefit is that it supports paper sizes slightly bigger than legal, whereas most scanner do no more than A4 or letter. As a result more books can be scanned two facing pages a time, and you save a lot of time.

One thing to be aware of: older lamps become unstable, which means you scans will not be evenly lighted.

Since I wrote this, my last HP ScanJet has retired. No more SCSI for me on my new computer.

Epson Perfection 1660 Photo

A relatively cheap scanner that always gave me very good results in color scans.

Unfortunately mine broke down when water came inside.

Epson Perfection 1670

A smaller scanner than the 1660. Supposedly a later model, but the results are not as good as its predecessor.

Plustek OpticBook 3600

A scanner specially designed for scanning books, and an absolute need if you scan books that are brittle, you are not allowed to damage and cannot afford a planetary scanner. It will also give you a speed boost on other books. However, the quality of gray-scale and color scans is way behind that of most modern scanners. The speed drops considerably when you go beyond 300 DPI.

The glass and lamp of this scanner extends to very close to the edge of the paper. However, the first centimeter the contrast is slightly different, probably due to the bend in the lamp, that allows it to get so close to the edge.

The software included with this scanner is not the most ergonomic, but, since I only use the book scanner application that comes with it, and works with three buttons (B&W, gray-scale, and color) to make scans. A nuisance with this application is that the compression settings compresses B&W to CCITT G4 Tiffs (what you want) and the other modes to JPEG in Tiff (what you don't want). Changing the settings requires interaction with the application.

Dust can easily find its way under the glass of this scanner, which may affect the calibration (most flat-bed scanners have a bright white calibration strip under the lid and when this gets dirty, you will have very ugly lines on your scan), and is difficult to clean, as you will have to open the cabinet to reach the inside.

Since I wrote this, my OpticBook died, due to a lamp failure.