This project, providing access to the important Humanist work of First Unitarian Society of Minneapolis ministers, includes addresses from Rev. John Dietrich, a pioneer in Religious Humanism and Rev. Carl Storm. There are a few addresses from Rev. Raymond Bragg who followed Dietrich in our pulpit and was a co-author of the Humanist Manifesto.
With rare exception, we've included only addresses given in Minneapolis, though the FUS archives include a few addresses from other locations. Seven hundred addresses were digitized. Well over 10,000 individual pages were shot with a camera, corrected in software, OCR'd and bound into PDF documents which were then uploaded to this site.
Special note should be made of Rev. Douglas Peary. The addresses you see here are images of the address pages. Though these are PDFs and we have used OCR spoftware, the results of the OCR analysis on these old typewritten and aged papaers is often less than useful. Peary has therepore spent many years transcribing the Dietrich addresses by hand. FUS is likely to move this site in the furure and our plan is to then include Peary's work to expanded the collection's offerings.
Crew
Our crew, Matthew Greene and Ellen Lenagh, are worth a mention as both crew members are on the high functioning end of the autism spectrum. This means that their intellects are quite intact -they are both actually brilliant in their own ways; the autism just prevents them from working with a lot of people. Ellen does computer work exclusively as she doesn't like to touch paper and Matthew does camera and computer work. Both have this opportunity to earn money doing important work. (If you have a project for us please contact George Greene geo@unclegeo.net)
A great advantage of people on the high functioning end of the spectrum is that they tend to be perfectionists -obviously a useful trait in this kind of work! George Greene (Matthew's dad) manages the projects and has an extensive background in image transfer of historic, scientific, medical and academic collections.
Gretchen Clemence and Ray Shuerers helped prep materials. Shuerers provided the lions share of the funding.
Originals
The original media include typewritten pages with, in some cases, handwritten marginal notes and corrections as well as a small number of reel to reel and audio cassettes of varying condition.
Paper
The paper ranges from thin, almost translucent paper to light card stock. The paper has aged to a lovely cream color (the files though are mostly in monochrome to reduce file size and increase readability) but the sermons are in remarkable shape otherwise.
There appears to have been a convention used by both Dietrich and Storm that as much of the paper as possible must be filled with type, which meant that words sometimes run off the edges of the paper and off the bottom of the page.
It is a marvel how the last line of a typewritten page could be typed at all when it is no longer held firmly by the typewriter. In the Storm sermons you see that reflected in the last lines curving off the page at an angle (as above), but the far older Dietrich sermons have perfectly aligned last lines, so much so that we wondered if the paper was cut from rolls after typing, but the pages are absolutely identical in length and are perfectly squared. The Dietrich addresses are half sheets, as shown below, and appear to have been typed on the half sheets, not cut.
We've taken great care to crop the image so only the apparatus is obscured; if the type appears cut off, it was cut off on the page.
Apparatus
Our copy stand is a Bencher Illumina (no longer available but can be found used) which has been modifed so the fan that normally cools an incandescent backlight for transmissives (in the well beneath the stage) now blows out rather than in to create a vacuum which pulls pages flat. You can see a bit of it in the images above. The vacuum stage is made of varnished pressboard routed to accept a perforated insert of what was once a LiteBrite kids toy.
Our illumination is from LED lights (Youngno YN 300 II) that are tunable for color temperature and brightness. These LED lights replace the stand's incandescents which would have been far too hot, not only for the original material, but for our crew! Also, the life of bright incandescent lights are measured in the low hundreds of hours whereas LEDs are in the tens of thousands.
The lights were designed for on-camera use and are battery powered with no provision for adapter power -obviously not good for sustained production, so 7.5V regulated DC power supplies were added. So as not to cut into the body of the lights, inexpensive realoadable AA Battery holders were modified with a Dremel tool to accept a small black plastic Radio Shack bulkhead DC power jack which fit perfectly in the battery housing and have internal switching that disaconnects any AA batteries in the holder when the power is plugged in. The lights are held onto the stand arms with nifty and inexpensive hot shoe adapters.
Our Nikon D5200 24Mpx DSLR was perfect for this project. A full frame camera was unnecessary as the OCR application simply does not need the extra quality.
The D5200 has no onboard jack for an AC adapter so we ordered an adapter that has an end that looks like a battery and goes where the battery goes.
We needed a remote viewfinder to check placement and focus as well as for a confidence image after each shot. The Nikon live view feature is output on HDMI to a 23 inch HDMI monitor. We needed a mini to standard HDMI cable.
We also needed remote shutter control which can be accomplished with the handy WiFi adapter that Nikon included in our camera package. A free Nikon phone app triggers the shutter... which would be perfect except that you can't use the HDMI and the WiFi adapter at the same time. Not only can't the connectors be physically connected at the same time, but Nikon says it wouldn't work even if you could (probably a power issue having both functional at the same time). Newer D5300 Nikons have built in WiFi and maybe the new software will allow HDMI and wireless to be used simultaneously.
In any case, an unexpected and even more elegant solution to our remote shutter problem was actualy provided by the light manufacturer which included a wireless remote to turn the lights on and off and control brightness and color temp. It also happens to have buttons for wireless shutter release for Nikon, Canon and Konika/Minolta cameras!
Post Processing
Pages were shot as JPEGs in the camera and processed in Adobe Lightroom. Images were converted to B&W, cropped to preserve as much of the page as possible, rotated as necessary and adjusted for lens distortion, clarity and tone as needed. A touch of sharpening was added as the type on the page was often light or indistinct due to the slight feathering made by the typewriter ribbon and the lightening of the ink over time.
Images were combined into PDF/A files (/A includes all assets needed to display the document, including fonts) and processed in ABBYY Fine Reader Pro 11 for Optical Character Recognition. I.e., the pdf's are are not just the pictures they appear to be; they also include searchable and copyable text.
Due to the limits of project funding and the enormous amount of time it would take to correct over 10,000 pages of text, the OCR is only Level One:
Level 1: Sub-OCR Quality. The text may have many typographical errors; essentially, it is unproofed text from automated OCR, probably of a less-than-ideal original.
...and unfortunately, with the typewritten documents, the accuracy is much lower than the typeset pages found in "The Humanist Archive" items within the Dietrich collection.
Why OCR at all? The accuracy may be good enough that the documents can be searched and relevant words are likely to be correct somewhere in the document. You may also want to highlight and copy text, say for a quote, and make corrections later in your own document. We did correct all address titles and dates in the OCR to aid in searches.
Audio
The reel to reel tapes of Carl Storm's addresses were between 50 and 65 years old, though many of the tapes had previous lives as musical and radio recordings, so the age of the tape is certainly older. Also, the quality of the original tape was on the lower end of available media at the time.
Storage conditions were dry and constant so the tapes were not sticky from binder layer degredation but some shed a bit of oxide. We elected not to "bake" the tapes. A few tapes had paper rather than plastic backer but held up well. There were some splices in the priginals and the tape did suffer in most cases from stretching and cupping.
The quality of the recording varies from fairly good to rather bad. In some cases there is print through from the previous recordings. The tapes also appear to be copies of other tapes as many recordings required a reel flip, though no content seems to have been lost at the flip point. At times you can even hear the bells from the Minneapolis Basilica in the background, but there is only one crying baby!
The original recording equipment was poorly grounded (or not grounded at all) so most tapes have 60 cycle hum (with many harmonics) which we've attempted to filter out as much as possible. High end response is lacking on most tapes, probably from the quality of the original deck and perhaps also a failure to demagnify the recording heads. Pops and crackiling, perhaps from bad connections are evident on a few tapes. (A tape exhibiting all of these issues is entitled Agnosticism.) A few tapes were oversaturated during recording resulting in muddy distortion in loud passages.
In some cases the tapes were recorded at 1 7/8 ips. Reel to reel decks that will do 1 7/8 ips could not be found so they were played back at 3 3/4 and the speed was reduced in software.
Ray Schruers transferred some reel to reel tapes to cassette (which also had speed problems) in 1997. Wherever possible, however, we digitized from the original reel to reels.
A Tascam 34B reel to reel deck and a Tascam 122MKII cassette deck were fed through a Mackie Onyx 1220 mixer to a PC running Tracktion DAW software.