I wonder whether this would be a good way to go for affordable high-quality book scanning.
Atiz Booksnap ($1,595)
2 Canon Powershot G10, $410.00 each (free shipping) at B and H Photo
2 Canon Powershot S5 IS, $250.00 (free shipping) at B and H Photo
With G10, grand total (before tax and shipping) is approximately $2,600. With two Powershot S5 IS cameras, grand total (before tax and shipping) is $2095.00.
I’d appreciate the advice or comments of anyone with experience. Can one get from camera megapixels and focal length to DPI? How does one go about the calculation?
I ended this last post with a plaintive question. But I began to think, and this ain’t rocket science. I remember telling Les Harrison that one needs to think about books in order to think clearly about digital humanities. If this is beyond the reach of my humanities brain–and I spent ten years in desktop publishing including a bit of pre-press–I need to get rid of that claim that I’m doing “digital” things.
The largest object that I need to capture is a page in the Jewett’s 1853 illustrated edition of Uncle Tom’s Cabin. The page size is 5 1/2″ by 9″. For capture of the pages, one actually needs to think of the pages as being about 6″ by 9 1/2″ because (in my opinion) a scholarly image capture should also include the cover and the page edges. A color chart and a ruler might also be desirable, so one should allow for a width of 7.5″ and height of 11.” So far, so good, because BookSnap can capture 10″ by 14″. But I also need to capture at an archival quality rate, which is higher than 300 DPI. From my days at Special Collections, 600 DPI was considered archival quality.
Note that my copy of National Era is captured at 450 DPI, which at 26″ by 19″ was at the upper limit of powerful image capture software of that day and took about 4 minutes per scan and produced 85 MB TIFFs. This was not funny when I tried to transfer 160 files from one server to another across the LAN. But that just meant I did not think clearly enough to know that a portable hard drive connected to a FireWire port was a more sensible option. The LAN connection between Special Collections and IATH didn’t cover a distance of 50 yards, so sneaker power was more sensible.
But, to get back to business, to capture a page image of Jewett’s 1853 illustrated edition I need to capture 600 dots on each inch of a 7.5″ X 11″ rectangle to have an archival quality scan, that is, 4500 dots by 6600 dots. Go to Google and search for “dpi and megapixels,” and near the top is this article from camera reference called photofolio, which even includes a handy guide for testing in PhotoShop. I have GIMP on home computer, so I guess I’ll test there. 7.5″ X 11″ at 600 DPI is a huge file, 269 MB, and thus straining the limits of computer power at my disposal.
But why be sensible? We want to push the limits to create a high-quality and (hopefully) a long-lasting project. Can we do it with a G10 or an S5 IS camera? G10 has effective resolution of 14.7 megapixels and S5 IS has effective resolution of 10 megapixels. In any case, I would ideally want to archive a rectangular image of 4500 dots by 6600 dots, which is 29,700,000, which in megapixels is 29.7. 30 megapixel cameras cost as much as a car. So that won’t happen on my project budget. Time to recalculate, after I sleep and think. But I do think the G10 would be preferred.
I think I made a mistake. I did not distinguish between how much space GIMP wants to manage the file and how big the saved file is. So here is the revised calculation.
7.5″ X 11″ at 600 DPI (29.7 megapixels in memory) is an 85 MB file.
7.5″ X 11″ at 450 DPI (16.7 megapixels) is an 47.8 MB file.
Let’s trim the empty margin and reduce space for color swatch and ruler.
7″ X 10″ at 600 DPI (27 MP in memory) is a 75 MB file
7″ X 10″ at 450 DPI (14.1 MP in memory) is a 56 MB file
Hey, I think that this can work. I’ve moved from a high-end 20+ thousand dollar camera four years go to an off-the-shelf solution. As I have a smaller page size, I believe that the project can achieve comparable results with $2,600 in equipment. It’s on the edge, but it’s on the cutting edge of affordable off-the-shelf equipment. More importantly, it’s what can be done within the budget.
At very least, for the three smaller editions with margins (Houghton Osgood at 5.5 X 8, Paperback at 6 X 9.25), it will be quite doable. A grant to purchase some capture would still be available for large pages. And having equipment in-house will probably suggest other avenues to explore, including Webb’s Christian Slave, which beckons, and perhaps (woe be me) HM Writings Edition. Large-paper edition of that just might work too.