Tuesday, February 8, 2011

Tech Tuesday - Digitizing With Cameras, Part I

Digital cameras now come in several forms.  There are a plethora of what are called point & shoot models, cameras that are good basic units that don't require much knowledge of photography or gadgetry.  Simply point them at what you wish to photograph, and press the button to take the picture.  The camera does the heavy lifting of figuring out how to make the picture good.  The other major type is Digital Single Lens Reflex (DSLR).  These are a digital version of the 35mm film cameras, and require more knowledge of photography, both theory and camera operation.  Most of these also have automatic modes like the point & shoots, but they do best when you understand how to get the most out of them for the images you want.  The single biggest benefit brought by digital photography is cost; you can shoot a nearly endless stream of digital images for virtually no money, where old film-based photography cost for every frame.

A week or so ago I wrote about scanners, how they work and some ideas about how to use them best for genealogy.  During the last Scanfest there was some Q&A about how to scan things that don't fit in your scanner  This includes oversized materials as well as three dimensional objects.  The simple answer is to use a digital camera!  The quality of digital camera images has increased very rapidly over the last decade, and current and recent models are more than capable of capturing very useful images.

Some of the items I need to scan are unfortunately glued into old scrapbooks and photo albums.  This means I will need to use my camera to digitize those items, while not damaging them by attempting to remove them from their mountings.  You'll want to get some distance and use a slight telephoto setting (or lens) as using a wide angle lens up close causes visual distortion, making straight lines appear curvy.  You'll also want plenty of light!  Your camera's flash may work well enough, try a number of setting variations to see if you can get usable results.  If not, you will need some kind of auxiliary lighting.  This can be done on the cheap.  Look at your local DIY superstore for clamp on reflector light fixtures and some 75-100 watt incandescent or 13-20 watt CFL bulbs.  A couple of these should be sufficient.  You need to place them at about a 45 degree angle above and to the left and right of what you're shooting to avoid glare and reflections, especially on glossy photos or plastic-covered album pages.  Be aware that if you're using this method, you will need to set your camera's White Balance to Tungsten so it can compensate for the yellowish color of the light.  Another option if you're shooting printed material like newspaper clippings is to set your camera to monochrome mode, which makes the resulting images black & white.

Here's an example I took:
 
Normal color shot with flash
Camera set to Monochrome




















I should note at this point that the same issues with image type that we discussed in the Scanner Technology article hold true here!  If your camera only exports JPG images, you should immediately convert those to TIFF before doing anything else.  This will keep them from getting degraded when you edit, and you will be editing these.  Looking at my samples, you can see there are several obituaries on this page.  I will use the smallest one on the right as an example.  Using my favorite image editor (I use Adobe Photoshop Elements, you can use whatever you are comfortable with) I straightened and cropped the image down to just the little obit I want to work on.  Here's the result:

Betty Steuart obituary, monochrome
Not bad.  However, I've already saved it once when I shot it, and now a second time when I cropped it.  You can already see some issues with image quality.  I did this one as a JPG as an example.  (When I shoot for archival purposes, I will shoot in my camera's native RAW format.  You should use RAW if you can, or convert to TIFF immediately if not, then convert and save as other formats as needed for other uses.)   I would like to use some Optical Character Recognition (OCR) software to transcribe this article to text so I can insert the text into my genealogy software.  OCR software works best with very clean type of high contrast.  This example is pretty low contrast, so I can use my image editing software to adjust that and boost the contrast.  Here's the result:

Betty Steuart obituary, contrast enhanced




This is much better!  Any stray markings can be interpreted by OCR software as additional letters or punctuation, so we want to eliminate as much as we can up front.  Here's the output from my OCR software from this cleaned up example:


MRS. BETTY STEAURT
WALNUT CREEK, Calif.
Mrs. Betty Lommel Steaurt of
here died. She is survived by a
daughter, Mrs. James (Rita)
Stoll; three grandchildren and
brothers and sisters, Mrs.
Hirsch (Ann) Kiebert,
dian, Idaho; Jerome, Modesto,
Calif.; Mrs. August (Gertrude)
Meyer, St. Paul; Mrs. Ray
(Johanna) Weismann, Min-
neapolis; Gilbert, Pearl Lake;
Mrs. Ray (Rose) Ubereeken,
Denver, Colo. Edmund and
Sally, St. Cloud.
There's just one error in this transcription, a common one where a 'c' has been rendered as an 'e' in Uberecken near the end.  That's a very good result for OCR software! 

I've covered a lot of material, and this should be enough to let you try it yourself.  If you have some album-bound material you need to digitize that won't fit in the scanner, try using your camera and see what kind of results you get.  Try a bunch of different settings to see what works best.  Next time I'll cover digitizing photographs and 3-D objects using a camera instead of a scanner.

Tech Tuesday – Have you stumbled upon a piece of technology or new Web-based application that would be of interest to your fellow genealogy colleagues? Post at your blog on Tech Tuesday and show us the ins and outs of this technology and how it can benefit the genealogy community.  This is a new series suggested by Donna Peterson of Hanging with Donna and in the past there have been many iterations of this series: the National Archives and Records Administration (NARA) blog Narations as well as The Family Curator by Denise Levenick.  (Blurb shamelessly stolen from Thomas McEntee at Geneabloggers!)