Digital Imaging Standards

National Anthropological Archives
Smithsonian Institution

The NAA's imaging standards were developed to meet the archives' needs at a particular point in time and reflect its staffing and funding opportunities. Although these standards have served us well, other imaging programs may require different solutions. Our standards are therefore not recommendations.

Image Resolution

All digital images created from reflective analog materials should have a minimum of 4,000 pixels along their longest dimension, and a minimum resolution of 400 dpi. All digital images created from transmissive analog materials should have a minimum of 5,000 pixels along their longest dimension, and a minimum of 500 dpi. For example, an 8 x 10-inch photograph should be scanned at 400 dpi; a 5-inch photograph at 800 dpi, etc.. However, items should never be scanned at less than 400 dpi — even if the longest dimension of the object is greater than 10 inches. When measuring the longest dimension of an image, round down to the nearest half inch. When calculating resolution, round up to the closest multiple of 25.

Calculations should reflect the dimensions of actual item to be scanned, not its matting or support. For example, a 5-inch print on an 8-inch mat should be scanned at 800 dpi (5 x 800 = 4,000), even though the longest dimension becomes 6,400 pixels (8 x 800 = 6,400). This exception is made so that the NAA can later reproduce the print or artwork at 400 dpi.

Text on the reverse side of an object may be scanned at 200 dpi.

File Format

Files should be saved in 24-bit RBG color TIFF format, with an IBM PC Byte Order.

Test Targets

All reflective images should be scanned along with a Kodak Q-13 Color Separation Guide. All transmissive images (i.e. slides, negatives, film, etc.) should be scanned alongside a Kodak #1A Uncalibrated Step Tablet. Test targets should not be considered when calculating the dimensions of the digital image.

Color and Greyscale Modes

All items (including black-and-white photographic prints) should be scanned and saved as RGB color images. Black-and-white photographic negatives should be scanned as RBG and saved as grayscale images. There may be exceptions when negatives are saved as RGB files.

ICC Color Profiles

TIFF image files should include the ICC color profile of the input equipment used to create the image (i.e., the flatbed scanner or digital camera).

Pre-scan Cropping

Crop each image to the edges and include the entire color bar.

Items that are matted on secondary material should be cropped to include a portion of the matting (and scanning resolution determined by the dimensions of the actual image, not including the mat).

The upper portion of the color bar may be partially cropped to achieve a smaller overall file size.

File Naming

Use lower-case letters only.

The same unique object identifier (name or number) should be used for archival items and their digital surrogates. For example, a print or drawing identified in the archives' catalog as (INV) 00000001 should be named 00000001.tif.

The verso or back of the item will be called 00000001v.tif or 00000001verso.tif.

Two parts of a broken item will be called 00000001_pt1.tif and 00000001_pt2.tif

Covers should be called collectionname_cover_front.tif (for example, 2003-22_cover_front.tif").

Oversize objects that include multiple smaller items (for example, a large Winter Count with multiple smaller images): shoot the entire object at maximum resolution, then also shoot the smaller items at ordinary (i.e. 300 dpi) resolution. If the smaller associated items have no number, use a suffix to indicate that they are a detail of the larger image. For example: if the large item is called 01234567, number the associated items 01234567_detail1, 01234567_detail2, etc.

Previously unnamed or unnumbered objects will be named after their NAA Accession Number, hence: ms99-01_01.tif, ms99-01_02.tif, etc. Hint: count the number of objects before you begin to scan a collection, so you can determine whether to use series such as "_01," "_001," or "0001."

Occasionally, collections will already have been numbered by the donor or collector. Retain these numbers where possible as a suffix to the (newly supplied) NAA accession number. For example, the Fruithead Filbert Collection (Photo Lot 99-87) arrived in the archives bearing photographs numbered 1-375. The digital surrogates for the collection should be numbered ms99-87_001.tif, ms99-87_002.tif, etc.

Donor-supplied numbers occasionally conflict with our own schema. For example, the Charles Edward Doty Glass Negatives (Photo Lot 73-26a) included symbols that conflict with operating system file-naming conventions (e.g., #, ?). In place of these, the NAA recommends:

"no" denotes #

"q" denotes ?

"x" for unnumbered items in an otherwise numbered collection.

These "new" inventory number should be written on the folders that house the objects.

Volume (Disk) Names

Disk names should bear the collection's Accession Number (or MS number ), bearing in mind that a CD-ROM volume label (i.e. its name) cannot include more than eleven (11) characters.

"Ms99-23 Disk1" or "99-23no1" for CD disk 1

"Ms99-23 Disk2" or "99-23no2" for CD disk 2, etc.

Hint: to keep volume names under 11 characters, skip the prefix Ms (manuscript) and substitute no (i.e., number) for the word disk.

Volume (disk) labels and directory names must conform to DOS file naming conventions and may not include these characters:

/ . : * ? " < > |

CD-ROM Recording

CD-ROMs must be ISO 9660 format. However, the file naming setting used in Toast should be "Allow Macintosh Names."

Additional Criteria

Each file shall contain a .tif extension.

Horizontal alignment of images shall be off no more than .25 degrees.

The correct aspect (portrait or landscape) of the original shall be retained in the digital image.

Record Keeping

Each completed collection should be entered in the imaging project database. Each record should indicate:

  • Accession number and title of collection
  • Number of original items, number of digital images
  • Items not scanned (and the reason; e.g., broken, fragile, oversize, etc.)
  • Whether item- and collection-level records exist in the Smithsonian's online public access catalog (SIRIS), as well as which item-level records are missing
  • Notes on imaging, such as as exceptions to file-naming conventions
  • Imaging hardware used
  • Names of imaging specialist and the specialist who proofed the images, with dates for each.


Last revised: August 11, 2008
Comments to Robert Leopold
or Stephanie Ogeneski