Guidelines: Content Categories and Subcategories Objectives:
Reformatting Historical Printed Matter, Documents and Manuscripts,
and Pictorial Materials — Content Categories and Subcategories Table
Still Image Working Group

Use Cases: Master Images and Image Sets

• Digitizing organization uses archival or production master image(s) to produce derivative images for the use cases listed under the tab to the right.
• Digitizing organization uses the master (or migrated copies) to create a virtual replica or a physical replica of the original item in the event of its loss, deterioration, or de-accessioning.
• Digitizing organization (or successor/receiving agency with an archiving mission) sustains the master (or migrated copies) for the long-term without loss of essential features.
• Digitizing organization uses masters for disaster recovery in the event of impairment of digital asset management systems.

Quality Notes

• Image-quality characteristics of the archival or production master image(s) must be sufficient to support the production of the various image types listed under the Derivative Images tab. Dependencies include appropriate image specifications and a production activity that applies appropriate process controls, e.g., the use of targets to monitor output, a quality assurance process that includes the use calibrated monitors and viewing environments, and various automated tools.
• For future exploration: When should master images be in an original-referred image state?
• Image-quality characteristics as above plus framing (capture of what curators define as the whole object) must be sufficient to permit the use of the master (or migrated copies) to create a physical replica of the original item in the event of loss or deterioration. That is, the master will fill the niche formerly filled by copy negatives and copy transparencies.
• Image-quality characteristics and completeness of coverage, plus selection of sustainable digital-content formats must be sufficient to permit sustaining or migrating the master over the long-term without loss of essential features.
• Disaster recovery in the event of the impairment of digital asset management systems depends upon the availability of metadata in standardized formats, including descriptive, administrative, and structural metadata, some or all of which may be embedded in individual images.

Use Cases: Derivative Images and Image Sets

• Patron views inline image in user interface.
• Patron makes a hard copy of one or more images for personal use.
• Patron is confident that the content received is an accurate and/or authentic reproduction of the original item.
• Patron (or content delivery system) receives information on rights and restrictions (delivery system may act on that information, if appropriate).
• Patron downloads one or more of the derivative images and, later, uses embedded metadata to identify the content and to determined technical and provenance information about the image.
• Patron studies details in, say, a hat that is part of a Civil War soldier's uniform.
• Digitizing organization may run a process that yields descriptions of "low level features" (colors, shapes, etc.). These descriptions may then be indexed in a special application, permitting patrons to search for and/or analyze images according to color, shape, etc.
• Publisher uses image to illustrate a high quality book.
• Publisher uses image to illustrate a large fine art poster.
• Publisher or service provider (on behalf of an individual patron) uses image to create museum-quality print. In some cases, this entails restoring a work (e.g., a map on yellowed paper with faded ink) to a state "like it must have been when first printed."
• Exhibit designer uses image for display "mural."
• Broadcaster uses image in high-definition television program, possibly with a Ken Burns effect.
• Geographer geo-references or geo-rectifies a map image for integration into a historical geographical sequence.
• Digitizing organization produces and archives an image that documents the condition of the original physical item in order to support its conservation.

Quality Notes

Inline image in user interface:

• Pictorial content is reasonably clear. Zoom in may be required.

Hard copy output:

• Equivalent to good quality photocopy.

Patron confidence in accuracy and/or authenticity:

• Depends upon provenance metadata (attribute of the copy) and trustworthiness of the provider (attribute of the institution).

Example of, say, Civil War details for study:

• Small item visible.

Image used for book illustration or large poster:

• "Studio quality," permitting an accurate reproduction of the shapes and tones of the original. Good enough to make printing plate(s) for a halftone printed with one ink or using a duotone method.*

* What specifics can be offered for the size and resolution (dots per inch) of the printing plates (or, for murals, the output of a large-format digital printer) that would represent success in these use cases? Will the size and resolution increase as the sizes of the original items increase, as is suggested by NARA's incremental recommended specifications for small, medium, and large originals? (See 2004 NARA Guidelines tab.) To illustrate how this might work, imagine creating printing plates for books with the following parameters: from a small original, a plate of 6x9 inches at 120 dpi; from medium, 6x9 inches, 180 dpi; and from large, 10x12 inches, 180 dpi. Similar benchmarks might be established for posters. Meanwhile, the production of exhibition murals might seek the following outcomes: from a small original, a 3x5-foot mural; from medium, 4x6-foot; and from large, 6x9-foot. The success measure would be, "Is the viewing/output digital image produced for these use cases good enough to do the job?"

Image for use in high-definition television program.

• "Studio quality," permitting an accurate reproduction of the shapes and tones of the original. Good enough to fill high-definition screen and zoom in using a Ken Burns effect.

Geo-referencing and/or geo-rectifying a map image may require metadata with geospatial coordinates; geo-rectifying requires high resolution to minimize aliasing.

Note: Derivative images will generally be in an output-referred image state.

Recommended

Note: The 2004 NARA Guidelines are divided by the size of the original item, as indicated below. The guidelines also include alternate specifications similar to those below, but with reduced requirements for pixel density. These alternate specifications are not included here.

Specifications for the format range 8"x10" or smaller and the size range smaller than or equal to 80 square inches:

Pixel Array:

• 4000 pixels across long dimension of image area, excluding mounts and borders Resolution
• Scan resolution to be calculated from actual image dimensions (approx. 400 ppi for 8"x10" originals and ranging up to the appropriate resolution to produce the desired size file from smaller originals, approx. 570 ppi for 5"x7" and 800 ppi for 4"x5" or 3.5"x5" originals)

Dimensions:

• Sized to match the original, no magnification or reduction

Bit Depth:

• 8-bit grayscale mode for black-and-white, can be produced from a 16-bit grayscale file

Specification for the format range larger than 8"x10" and up to 11"x14" and the size range larger than 80 square inches and smaller than 154 square inches

Pixel Array:

• 6000 pixels across long dimension of image area, excluding mounts and borders

Resolution:

• Scan resolution to be calculated from actual image dimensions – approx. 600 ppi for originals approx. 8"x10" and ranging down to approx. 430 ppi for 11"x14" originals

Dimensions:

• Sized to match the original, no magnification or reduction

Bit Depth:

• 8-bit grayscale mode for black-and-white, can be produced from a 16-bit grayscale file

Specification for the format range larger than 11"x14" and equal to larger than or 154 square inches

Pixel Array:

• 8000 pixels across long dimension of image area, excluding mounts and borders

Resolution:

• Scan resolution to be calculated from actual image dimensions – approx. 570 ppi for originals approx. 11"x14" and ranging down to the appropriate resolution to produce the desired size file from larger originals

Dimensions:

• Sized to match the original, no magnification or reduction

Bit Depth:

• 8-bit grayscale mode for black-and-white, can be produced from a 16-bit grayscale file