Alex Lee wrote:
On 2009-05-04 06:24:15 +1000, "Michael J. Mahon" <mjmahon@aol.com> said:Jerry wrote:]3ob <bob [ @ ] onemanband.com> writes:On 2009-05-01 20:57:21 -0400, Alex Lee <alelee@mac.com> said: I'd love to hear any tips/suggestions you have. My scanner can do some pretty high-quality scans, but I'm always open to new techniques. Any suggestions for getting around the spine curl (without destroying the issue)?If you use a cradle and lay the magazine on it: ... ... ... ... ... ... side view ... ... magazine spine in the center and then lay glass over top of the pages, you can then use a digital camera on a tripod to take good images of the pages. You do need to make sure the camera is far enough away from the pages to reduce the non-linear warping of the page rectangle, and you need to have strong light (preferably white light) directly above the spine. You want the position of the camera and the lights such that there is no reflection of the light off the glass and into the lens of the camera. Furthermore, you want the area behind the camera to be flat black (black cloth works) so there are no spurious images reflected onto the glass and into your image. I set up my camera in custom mode, so that the white-balance, focal length, aperture, etc are all consistent from page to page. I'm (slowly) working on building a jig like this where I can move the glass up and down easier, but from my experience thus far, I can "scan" a book much, much faster using a camera than with a scanner. I think it's somewhere around 7-10 seconds per page. There are newer scanners that are much faster so maybe that is less of an issue. If I had a working jig, I would also have much greater control over the position of the page within the camera shot, so it would be easier to crop away the non-page parts of the image. I use some free software tools to post-process the images (adjust background to pure white, improve contrast, deskew) and turn them into PDF. ImageMagick is quite useful.The one thing I would add is that if a sheet of flat black paper is inserted behind the "top" page on each side, there will be essentially no "bleedthrough" of the reverse of the page into the page images. -michael ******** Note new website URL ******** NadaNet and AppleCrate II for Apple II parallel computing! Home page: http://home.comcast.net/~mjmahon/ "The wastebasket is our most important design tool--and it's seriously underused."This is a major point raised by Michael - having a black sheet behind the page you're scanning will greatly reduce the other side of the page's contents coming through when scanning the page you want. This will, of course, mean the scan will come out a little darker, but with some Photoshop restoration, you can make what should be white, white again, along with black and the complete range of tone between white and black.As for using a camera to scan...I've never tried it myself, but somehow I think you're always going to get better scans with a flatbed or sheet fed scanner.
"Better" may be true, but "good enough" is the point, and a 10Mpixel camera will capture enough information that a Photoshop script can compress the blacks and whites, dropping out the paper texture, and sharpen it so that it can be saved as, say, an 8-level grayscale .GIF that has plenty of "ink" info and none of the "paper" info, ready for a PDF with OCR for searchability. Some prefer a 1-bit BW representation, but that often requires resampling to 300 or more dpi just to get clean edges on the fonts, while 8-level .gifs look fine at 150 dpi. The speed penalty of a flatbed is a serious problem unless you are willing to shear the spine and use an ADF overnight. -michael ******** Note new website URL ******** NadaNet and AppleCrate II for Apple II parallel computing! Home page: http://home.comcast.net/~mjmahon/ "The wastebasket is our most important design tool--and it's seriously underused."