CDL Guidelines for Digital Images
Chapter 3. Requirements
3.1. Number
of Images
Producers are strongly encouraged to submit at least one copy of
a digital master file for each digital object. Producers must submit,
at minimum, at least two derivative file types for each digital object:
- An access image (a service or reference image for more detailed
viewing).
- A thumbnail image (for the fastest access during the search,
browse, and retrieval process).
3.2. File
Formats and Compression
The following image formats are currently supported by the CDL for
the Enhanced Service Level. See Table 3.5
and Table 3.6 for specific image
format requirements.
- TIFF ITU-T.6: A 24-bit storage format commonly
used by Adobe PhotoShop and other bitmap editors, this TIFF format
may be used to store color images. This format is also suited for
bitonal text documents; it provides a high level of detail (up to
600 dpi, or 4,800 x 6,600 pixels for a letter-sized page). TIFF
ITU-T.6 format should be used for archival files. With lossless
compression, the data that results from compressing and then uncompressing
the image is exactly the same as the original, uncompressed file.
CCITT Group 4 compression is lossless and, therefore, permissible
for masters and derivatives. While compression is allowed for archival
files, it is discouraged, as it adds complexity to the format migration
issues of long-term preservation. When compression is used, it must
be lossless and not proprietary.
- JFIF (JPEG File Interchange Format): A specific
implementation of the Joint Photographic Experts Group compression
format (JPEG), commonly used by bitmap editing programs, viewers,
and web browsers. JPEG is a 24-bit, lossy (some information lost)
compression format that is well-suited for screen and print presentation.
JPEG is supported by all major computer platforms and by Internet
web browsers. With lossy compression, the picture quality of the
compressed file is reduced when compared to the original file, and
can not be restored, except by going back to the original. The advantage
is that the file sizes are much smaller, and image quality is acceptable
in most cases. It is not acceptable as an archival file format.
- GIF: An 8-bit, lossless compression format that
is well-suited for low resolution screen display of images. GIF
is often used for image thumbnails and screen versions of text documents,
and is supported by all major computer platforms and Internet browsers.
3.3. Watermarking
Watermarking technology of any kind should not be used on digital
master files. Visible watermarks should not be used for derivative
access image files, as they can lessen the legibility, quality, interpretability,
and integrity of the image. The use of electronic watermarks (or "invisible"
watermarks) is suggested as an alternative to visible watermarks.
In general, other methods for attaching rights information to derivative
access images (such as through rights management administrative metadata)
is encouraged.
For a summary of the visible and electronic watermarking technology,
see the Western
States Digital Imaging Best Practices (pp. 24-).
3.4. Digital
Cameras
Use of cameras to capture digital files should comply with the NARA
guidelines.
3.5. Guidelines
for Derivative Files
Table 3.5
| Access Image |
File format:
- JPEG (medium to high quality compression, sRGB profile for
color and Gamma 2.2 for grayscale).
Pixel array:
- 800-3000 pixels across the long dimension.
Resolution and bit depth:
- 8-bit grayscale, 24-bit color: 72-200 PPI.
|
| Thumbnail Image |
File format:
- GIF (adaptive/perceptual palette, diffusion/noise dither).
Pixel array:
- Image should fit within a boundary of 150-200 pixels across
each dimension (200 pixels preferred).
Resolution and bit depth:
- 4-bit grayscale, 8-bit color: 72 PPI.
|
3.6. Guidelines
for Digital Master Files
All digital master files should be in the following format, with associated technical metadata in the header file:
File format:
Technical metadata:
- Color and grayscale TIFF files should have ICC color profiles embedded in the file header, to indicate how the color and tonal values in the file are to be interpreted. Technical details for embedding ICC profiles in TIFF files can be found in the International Color Consortium's (ICC) Specification ICC.1:2004-10 (Profile version 4.2.0.0) (see p. 69).
TIFF files without embedded profiles will be assumed to be in sRGB color space if color, and Gray Gamma 2.2 colorspace if monochrome. More information about ICC color profiles can be found the ICC homepage. The Adobe Photoshop software (and associated literature) also includes considerable information about creating, embedding, and using color profiles.
Refer to the tables below for pixel array, resolution, and bit depth specifications, based
on the features of the original object being digitized.
Table 3.6.1
| Textual Documents, Graphic Illustrations/Artwork,
Maps, and Plans |
| Features of Original
Object
| Digital Master
File |
| Clean, high-contrast documents with printed type (e.g. laser
printed or typeset) |
Pixel array:
- 6000 pixels across the long dimension for 1-bit bitonal
mode.
- 4000 pixels across long dimension for 8-grayscale.
Resolution and bit depth:
- 1-bit bitonal mode or 8-bit grayscale: adjust the scan resolution
to produce a Quality Index (QI) measurement of 8 for the smallest
significant character. For more information about QI, see
the NARA
guidelines.
-or-
- 1-bit bitonal mode: 600 PPI for documents with smallest
significant character of 1.0 mm or larger. The 600 PPI 1-bit
files can be produced via scanning or created/derived from
400 PPI, 8-bit grayscale images.
-or-
- 8-bit grayscale mode: 400 PPI for documents with the smallest
significant character of 1.0 mm or larger.
|
| Documents with poor legibility or diffuse characters
(e.g. carbon copies, Thermofax/Verifax), handwritten annotations
or other markings, low inherent contrast, staining, fading, halftone
illustrations, or photographs |
Pixel array:
- 4000 pixels across long dimension.
Resolution and bit depth:
- 8-bit grayscale mode: adjust the scan resolution to produce
a Quality Index (QI) measurement of 8 for the smallest significant
character. For more information about QI, see the NARA
guidelines.
-or-
- 8-bit grayscale mode: 400 PPI for documents with smallest
significant character of 1.0 mm or larger.
|
| Documents as described for grayscale scanning and/or where color
is important to the interpretation of the information or content,
or desire to produce the most accurate representation |
Pixel array:
- 4000 pixels across long dimension.
Resolution and bit depth:
- 24-bit color mode - adjust the scan resolution to produce
a Quality Index (QI) measurement of 8 for the smallest significant
character. For more information about QI, see the NARA
guidelines.
-or-
- 24-bit RGB mode - 400 PPI for documents with smallest significant
character of 1.0 mm or larger.
|
Table 3.6.2
| Photographs: Transmissive Originals
(Film, Slides, and Negatives) |
| Features of Original Object |
Digital Master
File |
Format range:
- 35 mm and medium format, up to 4x5 in.
Size range:
- Smaller than 20 in. square
|
Pixel array:
- 4000 pixels across long dimension of image area, excluding
mounts and borders.
Resolution:
- Adjust the scan resolution to meet pixel array specifications, based on the format of the original object - approximately 2800 PPI for 35mm originals and ranging
down to approximately 800 PPI for originals approaching 4x5
in.
Bit depth:
- 8-bit grayscale mode for black-and-white, can be produced
from a 16-bit grayscale file
.
-or-
- 24-bit RGB mode for color and monochrome (e.g. collodion
wet-plate negative, pyro developed negatives, stained negatives,
etc.), can be produced from a 48-bit RGB file.
|
Format range:
- Equal to or larger than 4x5 in. and up to 8x10 in.
Size range:
- Equal to or larger than 20 in. square and up to 80 in.
square
|
Pixel array:
- 6000 pixels across long dimension of image area, excluding
mounts and borders.
Resolution:
- Adjust the scan resolution to meet pixel array specifications, based on the format of the original object – approximately 1200 PPI for 4x5 in. originals
and ranging down to approximately 600 PPI for 8x10 in. originals.
Bit depth:
- 8-bit grayscale mode for black-and-white, can be produced
from a 16-bit grayscale file.
-or-
- 24-bit RGB mode for color and monochrome (e.g. collodion
wet-plate negative, pyro developed negatives, stained negatives,
etc.), can be produced from a 48-bit RGB file.
|
Format range:
- Equal to or larger than 8x10 in.
Size range:
- Equal to or larger than 80 in. square
|
Pixel array:
- 8000 pixels across long dimension of image area, excluding
mounts and borders.
Resolution:
- Adjust the scan resolution to meet pixel array specifications, based on the format of the original object – approximately 800 PPI for originals approximately
8x10 in. and ranging down to the appropriate resolution to
produce the desired size file from larger originals.
Bit depth:
- 8-bit grayscale mode for black-and-white, can be produced
from a 16-bit grayscale file.
-or-
- 24-bit RGB mode for color and monochrome (e.g. collodion
wet-plate negative, pyro developed negatives, stained negatives,
etc.), can be produced from a 48-bit RGB file.
|
Duplicate negatives and copy negatives can introduce problems in
recommending scanning specifications, particularly if there is no
indication of the size of the source object that was photographed.
For copy negatives or transparencies of prints, use the specifications
for that print size. For duplicates (negatives, slides, transparencies),
match the size of the original object used to derive the duplicate.
However, if the original size is not known, follow these recommendations:
- For a copy negative or transparency, scan at a resolution
to achieve 4000 pixels across the long dimension.
- For duplicates, follow the scanning recommendations for the
size that matches the actual physical dimensions of the duplicate.
For scanning negatives with multiple images on a single negative,
see the section on scanning stereographs below.
Any reduction or enlargement in size must also be taken into account,
if possible. If a ruler has been included in the scan, use it to verify
that the image has not been reduced or enlarged before calculating
appropriate resolution.
Often photographic negatives are the most difficult originals to
scan. Unlike scanning positives, reflection prints, and transparencies
or slides, there are no reference images to which to compare scans.
Scanning negatives is very much like printing in the darkroom —
it is up to the photographer/technician to adjust brightness and contrast
to get a good image. Also, most scanners are not as well calibrated
for scanning negatives compared to scanning positives.
To minimize the loss of detail, it is often necessary to scan negatives
as positives (the image on screen is negative), invert the images
in Photoshop, and then adjust the images.
If black-and-white negatives are stained or discolored, we recommend
making color RGB scans of the negatives and using the channel that
minimizes the appearance of the staining/discoloration when viewed
as a positive. The image can then be converted to a grayscale image.
Table 3.6.3
| Photographs: Reflective Originals
(Prints) |
| Features of Original Object |
Digital Master
File |
| Format range:
Size range:
- Smaller than or equal to 80 in. square
|
Pixel array:
- 4000 pixels across long dimension of image area, excluding
mounts and borders.
Resolution:
- Adjust the scan resolution to meet pixel array specifications, based on the format of the original object – approximately 400 PPI for 8x10 in. originals
and ranging up to the appropriate resolution to produce the
desired size file from smaller originals, approximately 570
PPI for 5x7 in. and 800 PPI for 4x5 in. or 3.5x5 in. originals.
Bit depth:
- 8-bit grayscale mode for black-and-white, can be produced
from a 16-bit grayscale file
.
-or-
- 24-bit RGB mode for color and monochrome (e.g. albumen prints
or other historic print processes), can be produced from a
48-bit RGB file.
|
Format range:
- Equal to or larger than 8x10 in. and up to 11x14 in.
Size range:
- Equal to or larger than 80 in. square and up to 154 in.
square
|
Pixel array:
- 6000 pixels across long dimension of image area, excluding
mounts and borders.
Resolution:
- Adjust the scan resolution to meet pixel array specifications, based on the format of the original object – approximately 600 PPI for originals approximately
8x10 in. and ranging down to approximately 430 PPI for 11x14
in. originals.
Bit depth:
- 8-bit grayscale mode for black-and-white, can be produced
from a 16-bit grayscale file
.
-or-
- 24-bit RGB mode for color and monochrome (e.g. albumen prints
or other historic print processes), can be produced from a
48-bit RGB file.
|
Format range:
- Equal to or larger than 11x14 in.
Size range:
- Equal to or larger than 154 in. square
|
Pixel array:
- 8000 pixels across long dimension of image area, excluding
mounts and borders.
Resolution:
- Adjust the scan resolution to meet pixel array specifications, based on the format of the original object– approximately 570 PPI for originals approximately
11x14 in. and ranging down to the appropriate resolution to
produce the desired size file from larger originals.
Bit depth:
- 8-bit grayscale mode for black-and-white, can be produced
from a 16-bit grayscale file
.
-or-
- 24-bit RGB mode for color and monochrome (e.g. albumen prints
or other historic print processes), can be produced from a
48-bit RGB file.
|
For stereograph images and other multiple image prints, modified
recommended scanning specifications are to scan to original size (length
of both photos and mount) and add 2,000 pixels to the long dimension,
in the event that only one of the photographs is requested for high-quality
reproduction. For example, if the stereograph is 8 in. on the long
dimension, a resolution of 500 ppi would be required to achieve 4,000
pixels across the long dimension for that size format. In this case,
adding 2,000 pixels to the long dimension would require that the stereograph
be scanned at 750 ppi to achieve the desired 6,000 pixels across the
long dimension.
For photographic prints, size measurements for determining appropriate
resolution are based on the size of the image area only, excluding
any borders, frames, or mounts. However, in order to show that the
entire record has been captured, it is good practice to capture the
border area in the master scan file. In cases where a small image
is mounted on a large board (particularly where large file sizes may
be an issue), it may be desirable to scan the image area only at the
appropriate resolution for its size, and then scan the entire mount
at a resolution that achieves 4,000 pixels across the long dimension.
Table 3.6.4
| Aerials: Transmissive Originals
(Film, Slides, and Negatives) |
Features of Original Object
|
Digital Master File |
| NOTE: If scans of
aerial photography will be used for oversized reproduction, follow
the scanning recommendations for the next largest format (e.g.,
if your original is 70 mm wide, follow the specifications for
127 mm wide roll film to achieve 8,000 pixels across the long
dimensions). |
Format range:
- 70 mm wide and medium format roll film
Size range:
- Smaller than 10 in. square
|
Pixel array:
- 6000 pixels across long dimension of image area, excluding
mounts and borders.
Resolution:
- Adjust the scan resolution to meet pixel array specifications, based on the format of the original object – approximately 2700 PPI for 70mm originals and
ranging down to the appropriate resolution to produce the
desired size file from larger originals.
Bit depth:
- 8-bit grayscale mode for black-and-white, can be produced
from a 16-bit grayscale file
.
-or-
- 24-bit RGB mode for color and monochrome (stained negatives),
can be produced from a 48-bit RGB file.
|
Format range:
- 127mm wide roll film, 4x5 in. and up to 5x7 in. sheet
film
Size range:
- Equal to or larger than 10 in. and up to 35 in. square
|
Pixel array:
- 8000 pixels across long dimension of image area, excluding
mounts and borders.
Resolution:
- Adjust the scan resolution to meet pixel array specifications, based on the format of the original object – approximately 1600 PPI for 4x5 in. originals
and ranging down to approximately 1100 PPI for 5x7 in. originals.
Bit depth:
- 8-bit grayscale mode for black-and-white, can be produced
from a 16-bit grayscale file
.
-or-
- 24-bit RGB mode for color and monochrome (e.g. albumen prints
or other historic print processes), can be produced from a
48-bit RGB file.
|
Format range:
- Larger than 127mm wide roll film and larger than 5x7 in.
sheet film
Size range:
- Equal to or larger than 35 in. square
|
Pixel array:
- 10000 pixels across long dimension of image area, excluding
mounts and borders.
Resolution:
- Adjust the scan resolution to meet pixel array specifications, based on the format of the original object – approximately 2000 PPI for 5x5 in. originals
and ranging down to the appropriate resolution to produce
the desired size file from larger originals.
Bit depth:
- 8-bit grayscale mode for black-and-white, can be produced
from a 16-bit grayscale file
.
-or-
- 24-bit RGB mode for color and monochrome (e.g. albumen prints
or other historic print processes), can be produced from a
48-bit RGB file.
|
Table 3.6.5
| Aerials: Reflective Originals |
Features of Original Object
|
Digital Master File |
| NOTE: If scans of
aerial photography will be used for oversized reproduction, follow
the scanning recommendations for the next largest format (e.g.,
if your original is 8x10 in., follow the specifications for formats
larger than 8x10 in. to achieve 6000 pixels across the long
dimensions. |
| Format range:
Size range:
- Smaller than 80 in. square
|
Pixel array:
- 4000 pixels across long dimension of image area, excluding
mounts and borders.
Resolution:
- Adjust the scan resolution to meet pixel array specifications, based on the format of the original object – approximately 400 PPI for originals approximately
8x10 in. and ranging up to the appropriate resolution to produce
the desired size file from smaller originals, approximately
570 PPI for 5x7 in. and 800 PPI for 4x5 in. originals.
Bit depth:
- 8-bit grayscale mode for black-and-white, can be produced
from a 16-bit grayscale file.
-or-
- 24-bit RGB mode for color and monochrome (e.g. discolored
prints), can be produced from a 48-bit RGB file.
|
Format range:
- Equal to or larger than 8x10 in. and up to 11x14 in.
Size range:
- Equal to or larger than 80 in. square and up to 154 in.
square
|
Pixel array:
- 6000 pixels across long dimension of image area, excluding
mounts and borders.
Resolution:
- Adjust the scan resolution to meet pixel array specifications, based on the format of the original object – approximately 600 PPI for 8x10 in. originals
and ranging down to approximately 430 PPI for 11x14 in. originals.
Bit depth:
- 8-bit grayscale mode for black-and-white, can be produced
from a 16-bit grayscale file
.
-or-
- 24-bit RGB mode for color and monochrome (e.g. albumen prints
or other historic print processes), can be produced from a
48-bit RGB file.
|
Format range:
- Equal to or larger than 11x14 in.
Size range:
- Equal to or larger than 154 in. square
|
Pixel array:
- 8000 pixels across long dimension of image area, excluding
mounts and borders.
Resolution:
- Adjust the scan resolution to meet pixel array specifications, based on the format of the original object – approximately 570 PPI for 11x14 in. originals
and ranging down to the appropriate resolution to produce
the desired size file from larger originals.
Bit depth:
- 8-bit grayscale mode for black-and-white, can be produced
from a 16-bit grayscale file
.
-or-
- 24-bit RGB mode for color and monochrome (e.g. albumen prints
or other historic print processes), can be produced from a
48-bit RGB file.
|
Table 3.6.6
| Objects and Artifacts |
| 10 to 16 megapixel 24-bit RGB mode image, can
be produced from a 48-bit RGB file. If scanning photographic
copies of objects and artifacts, see recommended requirements
in the appropriate photo charts above (Tables
3.6.2-3.6.3).
|