[Outis]

I will need to digitize large portions of some books as a part of my work, and I was wondering about the feasibility of using a DSLR to take pictures instead of a flatbed scanner. (Obviously if I used a DSLR I'd have to do something to make sure the books were well-lit.) I know the DSLR would be capable of taking pictures way faster, but how would the resolution stack up against a flatbed scanner?

[Adam]

If you have a sharp distortion-free prime, high-resolution sensor (24Mp or higher will give you fantastic image quality), a stable tripod/rig, and external lighting (flashes and/or floodlights), the DSLR approach could certainly save you a lot of time.

[rrstuff]

Obvious thing to do would be to scan one and then take a picture of another one... since there are many different kinds of scanners and dSLRs.

but what kind of resolution do you really need? Is 16 megapixels from dSLR not going to be enough? I think it will be.
I would try to set up diffuse lighting and take pictures with a dSLR, because it will save you a lot of time.

[jatrax]

My wife does a lot of scanning of old photos and documents for family history. I have her scanner setup to dump the documents into a folder then have Lightroom set to watch the folder and import the file. Honestly, she can scan stuff about as fast as she can put it on the platten. It is a bit faster to shoot with the camera but when you have to then import the files, rename or whatever, I just do not see the whole process being any faster with a camera over a scanner. Unless you put a plate of glass on the book you are not going to get it totally flat and that is going to introduce distortion.

Cameras are great but for some things a scanner does flat out does a better job.

[lmd91343]

You would also need a 35mm or 50mm macro along with good copy stand.

How do you intend to keep the book open and the page flat? A (heavy?) piece of glass?

[narual]

This might be relevant: How good does the quality need to be?

If it's just "we need to get images of this stuff and they have to be readable", the camera would be much faster. It doesn't have to eb lighted perfectly, or perfectly flat, as long as it's readable. Put it on a tripod with a remote, have enough indirect light in the room so the book is lit well, Turn a page, snap a photo. Turn a page, snap a photo, repeat til you're done. Reset the numbering on the camera and the pages will even be pre-numbered (well, at least, in order, since I'd guess you'd be snapping 2 pages at once). If you don't move the book or camera, you can crop one shot, do any sharpening and contrast changes, and then apply the same changes to all the other pages in the book, and you'd be done with a book pretty quickly.

If it's "we need to be able to print new books from the images and have them be indistinguishable from the originals" you could probably still do it with the camera, you'd just have to be more careful.

Either way, the 35mm macro would probably be perfect for it.

If the pages are glossy, you might need to put polarizing film on the light sources and a polarizer on the lens, to cut possible glare.

[stevebrot]

If you have access to a copy stand with lighting standards (45 degree), the dSLR might be a good option. A tripod is less useful. Working distance is important and should be considered when deciding on a lens for the project.

A flatbed scanner will be slow...very slow...

Steve

[stevebrot]

Once you get the process dialed in, it can go quite quickly.

I just looked on B&H and the prices are not bad either for a basic model or better yet, simply find a used 6x7 enlarger for less than $100 and either buy the appropriate copy rod or have one machined.


Steve

[Outis]

Oh, wow, thanks, everyone!

At the very minimum, it needs to be readable, and I need to be able to print pages and have them be readable also. I'd also like to be able to run OCR on certain areas.

[bar_foo]

Scan Tailor can be helpful for cleaning up the images (especially if you're working mainly with text).

[pathdoc]

Put it on a tripod with a remote, have enough indirect light in the room so the book is lit well, Turn a page, snap a photo. Turn a page, snap a photo, repeat til you're done.

I once did this as an experiment with nothing but the book, a table, and a hand-held *ist-DL with kit lens. One hand held the pages flat and turned them; the other hand worked the camera. The text was surprisingly readable, but it's certainly not something I'd settle for if official archiving of the material was intended. I should mention that I used the AC adapter for the *ist-DL; this is not something you want to do with batteries unless you have fully charged spares.

[narual]

I once did this as an experiment with nothing but the book, a table, and a hand-held *ist-DL with kit lens. One hand held the pages flat and turned them; the other hand worked the camera. The text was surprisingly readable, but it's certainly not something I'd settle for if official archiving of the material was intended. I should mention that I used the AC adapter for the *ist-DL; this is not something you want to do with batteries unless you have fully charged spares.

I'd guess it does a good enough job for any good modern OCR software to digitize the text pretty easily.

[nikigunn]

I have a flatbed book scanner (Plustek OpticBook), because I know some of my physical books will never be available as ebooks and I just don't read paper books anymore. Once you get into it, scanning is very fast. I get tired of turning pages and flipping the books around before I get frustrated with the scan speed. The platen goes all the way to one edge and a side of the book hangs down so there's no gutter shadow. I use ABBYY Fine Reader for the OCR, a full version instead of the one that comes with the scanner. I considered building a scanning rig for my camera but buying this was easier, faster, and I'm happier with it. The Mobile Read forums have links if you want to build your own camera setup.

[tim60]

Another possible advantage of the DSLR is if the books are fragile or difficult to open flat. With DSLR you only need to open the book to give clear view of the page imaged, flatbed scanner requires it to be opened flat. Partial opening will require some kind of bracket to support it.