Hamilton College Digital Collections


Table: Digitizing Printed Text - Recommended Specification
Use Original Resolution Bit
Depth
Color
Space
Dimen-
sions
Editing/
Processing
File Format/ Compression Storage

Archival
Master
(AM)

Long-term archive

b&w/text - in good-condition (high-contrast)

300 ppi

8-bit

grayscale 1

Varies (100% of original; no less than 4,000 pixels on LS)

Check grayscale factors 2

TIFF, with little endian (IBM PC) Byte Order; no compression

Stored on 2 media: CD DVD, tape, spinning disk

b&w/text: to be OCR'd

400 ppi

b&w/text - in poor condition, tiny print, or with illustr.

600 ppi

with color

24-bit

RGB

Check grayscale and color accuracy3

Working
Master
(WM)

-To create derivatives.

-Publication printing.

Same as AM

Same as AM

Same as AM

Same as AM, but sRGB if color

-Change file format.4

-Saved as sRGB

-Tone & color adjusted

-Sharpened

Same as AM, but may be compressed with LZW compression if saving storage space is essential.

If kept, same as AM.

Static
Delivery
File

Internet delivery (multiple files)

N/A

150 dpi5

24-bit
(b&w: 8-bit or 1-bit)

sRGB
(b&w: grayscale or bitonal)

No less than 640 pixels on wide

-Reduced dimensions
-Sharpened

GIF6

(JPEG, medium quality, 5-7, of 10)

Spinning disk and backup tape

-b&w text may be converted to grayscale/ bitonal.

ITU/CCITTGroup IV compression

Compound
Delivery
File

Continuous reading (one file)

?

?

Append all pages

PDF or PDF/A with minimal compression

Tran-
scription

Keyword searching

Uses Working Master

N/A

N/A

N/A

N/A

OCR;
[TEI encoded]

UTF-8;

TXT (or XML if encoded)

Note: No color or grayscale color profile target needs to be scanned with typed (or type set) textual materials, however, such textual materials with illustrations are scanned as 8-bit grayscale and should be scanned with a grayscale target.

[1] Black and white text documents on white backgrounds should be scanned at 300 ppi, 8-bit grayscale to produce the Archival Original version, however, if the original material was a clean, high-contrast black and white printed document (e.g. laser printed or typeset) and a small delivery file is desired, the Working Master can be converted from to 1-bit bitonal and saved as the Static Delivery File. Moreover, many textual images can be reduced by as much as 50% from the size of the original material without reducing legibility. Only testing will tell for sure.

[2] Grayscale factors: minimum number of visible steps: 18; minimum number of f-stops: 5.5; Y channel noise: <=5%.

[3] Color accuracy: DeltaE < 8, ICC Profile (decrease if image is very large).

[4] If you are using Photoshop to edit the Working Master, change the file's TIFF format to PSD, Photoshop's native file format, which allows for more editing functionality.

[5] 150 dpi is a rather conservative setting for getting a reasonably good print out of most images that will fit on a computer screen. If you change the value in the Resolution box to another dpi (in Photoshop), always readjust the Pixel Dimension appropriately because changing the dpi automatically changes the Pixel Dimensions as well.

[6] Although GIF files tend to be larger than JPEG files, trey display faster in a browser because of their lower bit depth and because if their type of compression decompressed faster. Also, the GIF format is well suited to storing images of discrete data such as printed or typed text and line art.

(Updated: May 6, 2008; December 17, 2010 pjm)