The Cornell / Xerox / Commission on Preservation and Access Joint Study in Digital Preservation
Appendix I–Paper Facsimile Comparisons
The sample pages located in Supplement V graphically represent the quality obtained through light lens and digital processes. The samples are reproductions of a standard Facsimile Test Chart that provides a means for evaluating the capture of text, gray scale, line art, photography, and resolution.[*] The Test Chart includes two samples of the Microcopy Resolution Test pattern for evaluating microrecording systems. They are used to measure resolution, which is defined in terms of the number of line pairs per millimeter that can be seen or differentiated on these test patterns. The first measures resolution on both the vertical and horizontal axis; the second measures resolution at right angles. While primarily used to determine the quality of microfilm, the test charts provide a convenient means to compare the paper output of light lens processes to digital scanning.
The results and observations of the text pages were as follows:
- Canon 8580 photocopy (regular setting)
- text: 2 pt. text legible, some characters partially represented
- resolution patterns read: 6.3 (first pattern); 5.0 (second)
- photograph: poor definition
- gray scale: 5 discernible levels
- other: second set of fine lines at top not reproduced
- Canon 8580 photocopy (photo setting)
- text: lower case 2 pt. text partially represented
- resolution patterns read: 6.3 (first pattern); 4.0 (second)
- photograph: marginal quality
- gray scale: 6-8 levels
- other: second set of fine lines at top not reproduced
- Minolta RP605Z reader/printer (printout from microfilm)
- text: 2 pt. text not rendered, 4 pt. and 6 pt. partially represented
- resolution patterns read: 2.5 (first pattern); 2.2 (second)
- photograph: unsatisfactory
- gray scale: 5 levels
- other: second set of fine lines at top partially rendered
- 600 dpi Digital File
- text: upper case 2 pt. text barely legible, lower case 2 pt. text partially rep.
- resolution patterns read: 5.6 (first pattern); 5.0 (second)
- photograph: good definition
- gray scale: 9-11 levels
- other: second set of fine lines at top fully rendered
- 300 dpi Digital File
- text: 2 pt. text illegible, 4 pt. text breaking up
- resolution patterns read: 4.0 (first pattern); 3.2 (second)
- photograph: dotted, like pointillism
- gray scale: 10 levels, but dotted
- other: fine line sets at top partially rendered
- 200 dpi Digital File
- text: 2 pt text illegible, 4 pt. text breaking up; block representation to 12 pt. text
- resolution patterns read: 2.8 (first pattern); 2.8 (second)
- photograph: very grainy, eye-legible dots–gray scale: 9 levels, eye-legible dots
- other: fine line sets at top poorly rendered
The first two sample pages were produced on a Canon 8580 which is used at Cornell to make preservation photocopies of brittle books. The first photocopy was produced on the regular setting and the second was produced using the “photo” setting. This latter shows an improvement in the capture of the gray scale and photograph, but there has been a marked decline in the capture of text and fine lines.
The third sample is a printout from a microfilmed version of the scanner test target. The copy was produced on a Minolta RP605Z reader/printer that uses a dry printing process. The film from which this printout was produced has a high resolution, and the grayscale and photograph are acceptably rendered. The paper printout, however, represents the poorest quality of the six samples. Text, photograph, and gray scale have all clearly suffered. The results indicate that, although the film copy may meet national standards, the quality of the paper use copy does not. More attention should be paid to the quality requirements of the use copy.
The fourth sample is a 600 dpi representation. The gray scale at the top of the test chart and the photograph were windowed during scanning and captured using the photo mode with a selected filter, screen, and Tonal Reproduction Curve (TRC) that rendered them exceptionally well. Eleven shades of gray are discernible and the photograph possesses a depth not present in the photocopied versions. The text and fine lines were captured in the mixed mode, again using a variety of settings, and fall somewhere between the two photocopy versions in terms of text capture. The overall presentation of the 600 dpi scanned version is superior to either of the photocopies and the microfilm printout, and dramatically illustrates digital technology’s advantage over light lens processes for recording illustrated material.
The final two examples are versions scanned at 300 and 200 dpi resolution. Each was scanned using the identical windows and settings that were applied to the 600 dpi copy. The text, gray scale, and the photograph have declined visibly in overall presentation. The quality achieved with 300 dpi resolution is comparable to that of a laser printer; the 200 dpi version resembles the quality achieved with a dot matrix printer. The gray scale and the photograph are rendered as eye legible dots in both copies. While lower resolution scanning can produce satisfactory copies from crisp, high contrast modern documents of 6 point type and larger, these examples graphically illustrate the limited utility of lower resolutions to capture the variety of printing techniques and illustrations found in older research materials.
