How we created the online Ranfurly Collection
In 1998 the Alexander Turnbull Library acquired from England the private
papers, photographs and artworks of Lord and Lady Ranfurly, from the period
when Lord Ranfurly was Governor of New Zealand (1897-1904). It is a rich
collection that is already providing historians with new source material
for the study of New Zealand's social and political history around the
turn of the last century.
The Ranfurly Collection was selected for for this project was that it
represents a typical cross section of the material normally held in the
Alexander Turnbull Library. This would provide the challenge and learning
experience of digital capture in more than one form.
Using EAD
This is a pilot digital project using Endcoded Archival Description (EAD)
which allows for easy browsing of digital items, their descriptions and
navigation within the collection.
(EAD) is a data structure standard, designed for the markup of archival
records for the purposes of search, retrieval, presentation and navigation.
It was developed to provide a metadata standard for the retrieval, display
and navigation of archival finding aids over the Web. EAD is SGML-based
and XML compatible. It has incorporated the International Council on General
International Standard Archival Description (ISAD(G)) and the Rules for
Archival Description (RAD) standards, and provides MARC and Dublin Core
(DC) equivalencies for the storage of multiple formats.
The 118 elements can be further qualified by 'attributes' (eg source,
audience, level, parent, role) to give an extensive range of definition
to the elements in a description record giving equivalently high levels
of intellectual access. Because of this flexibility, clear implementation
guidelines are recommended for successful sharing of EAD encoded material
within and across jurisdictions. An example of an EAD encoded data structure
and the resulting display that this provides is attached.
EAD provides a hierarchical structure, with lower levels of description
inheriting content from higher levels. It enables sophisticated graphical
Web displays that provide easy navigation within and between collections.
EAD is intended for "communicating with users", and the data
required to facilitate the discovery, retrieval and use of archival materials
in a Web environment - particularly that relating to collections, items
and pieces and their surrogates (and the administration of this data)
- fit the EAD model.
Data relating to collection management functions such as acquisitions,
reader registration, item movement, conservation treatment and the maintenance
of authority terms are not provided for within EAD, although this information
can be linked to from an EAD record and summary information can be displayed
automatically or on demand.
Conservation Treatments
Manuscript Scrapbooks - Dave Ashman, Book Conservator
The manuscript items were contaminated with quite a lot of dirt and dust
as well as a fair amount of dried mould. The main work carried out on
these bound materials and scrapbooks was cleaning using a soft brush and
vacuum cleaner under a dust extraction hood. After cleaning and microfilming,
protective enclosures were constructed. Repair wasn't deemed necessary
at this stage due to the high input of time required and likely low future
usage.
Photograph Album (PA-f-195) - Mark Strange, Conservator of Photographs
The condition of the album was assessed with specific interest in the
handling and the illumination necessary for copying. The prints were albumen
or silver-gelatin and were suitable for copying with the digital camera.
The album was in poor physical condition. It had been exposed to moisture
for an extended period and had mould growth and stains on many pages.
Most pages were buckled and many photographs were peeling away from the
paper supports. Two prints were completely detached and another was missing.
The pages, which are a card/paper laminate, were delaminating as the paste
or glue had lost its adhesion. The facing layer of the pages had discoloured
generally.
Mould deposits were killed and removed with alcohol and cotton swabs.
Because of the size and weight of the volume, it was decided to reattach
the prints and consolidate the peeling pages, to prevent accidental damage
to the prints. All repairs were made with wheat starch paste. When the
repairs were completed, the volume was interleaved with blotters and pressed.
This reduced the buckling in the pages - which will contribute to improved
sharpness in the copies. Interleaving tissue was cut for insertion between
each page. Finally photocopies were made of the pages to assist the curatorial
staff with cataloguing (from the copies) while the album was being digitally
copied.
Microfilming
Hybrid Approach to Reformatting
The first objective was to capture all the manuscript material on 35mm
microfilm to preservation standards. The manuscript material was microfilmed
in 1999 (six reels) by the Alexander Turnbull Library's Microfilm Unit
to Library of Congress preservation standards.
As the National Library has the in-house ability to microfilm, and make
preservation copies at the start of the reformatting process, the Yale
University hybrid model of "film first then scan" was adopted.
A hybrid approach ensures a preservation copy in microfilm is made as
well as digital copies. By microfilming first, intermediate copies of
the master films could then be made and safely sent off site for digitisation.
Similar institutions (National Library of Australia - The Ferguson Project,
Library of Congress - George Washington Papers) have also adopted this
model for some of their digitising projects. Kodak 2470 microfilm was
chosen to make the intermediate negative copies. This film in particular
will allow the reproduction of as much tonal range in the original manuscript
items as possible.
Digitising Methodologies
Manuscripts (2,970 items)
The task of creating digital files from the microfilm was outsourced
to the Library's micrographics contractor, New Zealand Micrographic Services
Ltd. who utilised a "Sunrise" microfilm roll film scanner to
automate the digitising process. The intermediate negative film was scanned
to produce both bitonal and greyscale digital files. The greyscale files
produced better tonal representation of the original material, which included
sketches, newspaper clippings, photographs and various density of ink
handwriting. The intermediate microfilm was scanned as greyscale 8 bit
TIFF files at either 200 or 300 dpi to produce digital master copies of
the film.
The next stage was to check all digital and crop the images tidily using
"PhotoShop 5.5". A PC and "microfilm reader" were
set up alongside each other at a workstation for the checking and preparation
of the image files. These files were then carefully resized to allow efficient
Internet access as well as still giving full legibility when delivered
on screen. The software application "ThumbsPlus" was used to
allow batching of the resizing and file reformatting process where possible.
Two copies of each file were saved, one a high resolution TIFF file for
archival storage and the second a resized JPEG file for Internet access.
Where applicable, these screen resolution JPEG files were grouped and
saved as PDF files. This allows users to access groups of images that
could display images in a physical context similar to that of the original
items.
Photographs (450 images)
The Alexander Turnbull Library's Image Services Unit scanned all photographs.
All images were scanned using a "Leaf Lumina" digital camera
at its maximum resolution of 2253 x 3380 pixels as greyscale images. Each
cropped and prepared image is approximately 6 megabytes. The 24 separate
photographs were scanned by a flatbed scanner (Agfa Horizon+) at various
resolutions to produce digital copies of at least 8 megabytes.
In both cases this was to allow good reproduction quality of the digital
images.
Watercolour paintings (40 images)
The Alexander Turnbull Library's Conservation Unit photographed the paintings.
All images were photographed on 4"x 5" colour transparency
using a view camera, the transparencies were then scanned (Agfa Horizon+)
at approximately 750dpi. These files were saved as 18 megabyte 24bit TIFF
files.
All the digital images from photographs and paintings were prepared in
"PhotoShop 5.5". Two files of each image are stored, one an
archival TIFF file and one a resized JPEG file at a minimum of 800 x 600
dpi.
In addition to the above 40 images, 17 pages from one manuscript scrapbook
were selected to be photographed on colour transparency. They were colour
maps and paintings that would be better reproduced in colour. Again these
images were prepared as for the paintings above. Here the greyscale files
digitised from the microfilm were deleted and replaced by these full colour
images.
Challenges and issues during the project
Microfilming
Microfilming the manuscript scrapbooks proved challenging. One of the
manuscript scrapbooks required partial re-filming as some pages were identified
as being out of order. The size, format, and density of items within this
scrapbook varied widely making it comparatively complicated and difficult
to film.
The Library had not previously microfilmed material in readiness for
digitising. This required training and adjusting some filming processes
to allow for the scanning of the negative intermediate film. Research
and choosing an appropriate film stock was required to retain as much
texture and tonal representation of the originals.
Digitising Microfilm
All microfilm frames were scanned in both bitonal and greyscale to compare
results. As suspected, bitonal scans of the microfilm did not reproduce
all the manuscript items particularly well. In some cases the bitonal
scans failed to reproduce all the information, yet their greyscale counterparts
did. Typically this would be in areas of poor ink density in the handwriting,
sketches, photographs and paper tones. This does not mean bitonal scanning
is not effective for microfilm, but consideration of the original material
and the final information required to be reproduced is required when choosing
the scanning process. The quality of the final result supported the additional
expense and extra data space for storage required by the greyscale scanning
of manuscript material.
When filming the material, a "blip" in the form of a small
white card was filmed with each item. This was to aid the microfilm scanner
to pick up the leading edges of the manuscript material to allow for as
much automation of the digitising as possible. But on two reels of film
the orientation and size of the material changed frequently which did
not allow automated scanning operation and finally individual image by
image scanning was required. These reels were not scanned on the "Sunrise"
roll film scanner and the tonal range in the final images did not match
the quality of the "Sunrise" output. This required the rescanning
of some pages using the "Leaf Lumina" digital camera to make
files to replace these.
The digitised set of images from the microfilm required frame by frame
checking against the images on microfilm to ensure accuracy, effectively
a "double handling" in the checking process. Overall the digitisation
of the microfilm was more time intensive than first thought.
Digitising Photograph Albums
The digital camera used to capture pages and individual photographs from
the albums is very slow (scan time of three minutes). It is unable to
render accurate colour results therefore the resulting images are reproduced
in greyscale. An improved camera system with a quicker scan time would
improve efficiency and enable accurate colour capture and full colour
reproduction.
When digitising the photograph albums, it was thought that if only the
photographs were digitised and cropped the physical context of the albums
would be lost. Therefore it was decided to digitise full pages and covers
of the albums to allow visitors to "read" through the album
as they would with the original album in front of them. To supplement
the full page images, some albums have had individual photographs digitised
to allow more detail to be seen in each on-screen image.
|