I wanted to gather some wills (from around 1730-1800) for various people in my tree from
Norwich Record Office. My hope was that I could end my visit with a full set of digital images of the wills. I knew from the
online catalogue that they had them on microfilm. I also know that the NRO have a very liberal attitude to camera support/tripods. It can be summed up as "as long as it doesn't cause problems, you're OK".
Armed with this...
I knew that microfilm readers tend to have very uneven lighting, so that was an issue to consider. I was a little worried about screen reflection, although it turns out the reading area was dimly lit, so not a problem.
I set up my tripod around 4 feet away from the screen, which left me room to get between the camera and the reader to load/manipulate the film. I used a bit of zoom on the camera to get a full screen shot at this distance.
In order to compensate for the uneven screen lighting, and to avoid the possibility of any part of the image being unreadable due to bad exposure, I took 3 shots for each frame, offset by 2Ev. My camera is capable of auto-bracketing, and I used a timer release to avoid any camera shake.
I captured 12 wills, 32 pages, 96 images in this way, with a few "oops" moments, duplicates etc along the way. The staff at NRO were very helpful with some issues of film-reader manipulation, and locating wills on rather long microfilm reels.
When I came to process and finalise the images, I used
enfuse to make a "best case" version of the 3 bracketed shots for each page.
enfuse --output="merge.tif" --exposure-weight=.8 --saturation-weight=0 --contrast-weight=0.2 shot_normal.jpg shot_plus2.jpg shot_minus2.jpg
This gives an image with is fairly uniform, but slightly lacking in contrast, so I used an automatic process (command line, but Gimp or Photoshop can do this as well) to stretch the contrast to be full from black to white.
Having said all that ... my biggest problems did not arise from digital photography or processing. The original micrfofilms were rather heavily scratched and of low contrast. Since the best I could possibly do it make a perfect copy of the microfilm, this set a low bound on quality.
The other problem was that the reader I was using would quite often not focus the whole image in one go. If one part of the image was in focus, another would be out of focus. Reading the film "there and then" this wouldn't be an issue. You'd just have one hand on the focus, and adjust.
Sadly, I didn't really appreciate the extent of this problem until I was processing and trying to decipher the images, back home.
To pre-empt the obvious question, I didn't try to read the wills at the NRO because I know that deciphering old wills is slow work (and I have use my annual leave days from work to be at the NRO). Worse, I make extensive use of online resources (e.g. old legal "boilerplate", placenames) to guide my interpretation; I also try whenever possible to include all references for my tree in digital form on my database.
I am happy to report that despite my images being imperfect, I have been able to successfully transcribe the particular information I needed from each will, and am in a position to make a full transcription if needed.
I hope this report of my technique, successes, and failures may be of help to anyone else attempting a similar task.
BugBear