Imaging Best Practices

Image scanning at UC Berkeley
Scanning for most UC Berkeley Library projects is done in the Library's Digital Imaging Laboratory (DIL). Digital image capture equipment at DIL includes:

  • Betterlight Super 8K-2 Digital camera, which captures images up to 8000 x 10,600 pixels (LxW). This camera is used to capture originals in bound volumes, and originals too large for flatbed scanning.

  • Epson 1640XL flatbed scanners, which can scan originals up to 11x17 inches. These scanners are primarily used with unbound originals, especially document pages and loose photos, and for scanning film negatives 4x5 inches and larger.

  • Zeutschel 10000T overhead camera, used for books and other textual materials.

  • Nikon LS2000 film scanner, for 35mm film scanning

Capture Specifications for Digital Masters
Our philosophy of digital image capture can be summarized as "scan once, re-use for many purposes." We expect that our master files should be useful for disparate, demanding purposes including scholarly study and high-quality publication, so we follow guidelines for resolution, image composition, and file format to ensure the ongoing value of the images.

CDL Image Standards and other Guidelines
We follow the California Digital Library Guidelines that pertain to imaging. See the following documents:


Other published guidelines which inform our work include the following:

In addition to recommendations for resolution, file formats, etc., these various guidelines contain a wealth of information about best practices for imaging.

File format
Digital masters are captured in 24 bit RGB color and stored in uncompressed TIFF format. Scanner- or camera-specific image capture software is used to manage the technical details of image capture, and then Adobe Photoshop is used to save the TIFF file.

Capture resolution and master file size
DIL's preferred capture resolution for reflective originals is 600 pixels per inch, yielding RGB TIFF master files in the 60 to 120 MB size range for originals close to letter size (8.5 x 11 inches). Larger originals are usually captured at a lower resolution, to keep the file size from growing too much beyond 120MB, and smaller originals may be captured at higher resolution.

Included targets
For reflective originals, a one-piece target is imaged at the edge of each capture. It combines the grayscale target and the color patches from a Kodak Q-13 Color Separation Guide and Grayscale with a centimeter scale, all in a compact layout created using a hobby knife and two-sided adhesive tape. The information from the target is intended to provide information about the tonality and scale of the image to scholars and technicians. The "A," "M," and "B" steps of the grayscale are marked with small dots to make them easy to identify for making tonal measurements during capture set-up and file processing. Several different-sized versions of the combined target are suited to the range of sizes of the originals.

Tonal metric
The RGB data in the image files is captured and archived in the native colorspace of the capture device (camera or scanner); an appropriate ICC color profile is embedded in each master TIFF to characterize the native color space of the device. Before capture occurs, the camera (or scanner) operator uses the controls in the scanning software to adjust the color balance, brightness, and contrast of the scan so that the grayscale target in the image has the expected RGB values. These values are as follows: for the white "A" patch, R, G, and B values all at or near 239; for the middle-gray "M" patch, RGB= 98; for the near-black "B" patch, RGB= 31. These expected RGB values are appropriate for a 24 bit RGB image with gamma 1.8.

Color management
DIL uses ColorSynergy software to create ICC color profiles to characterize the color characteristics of each capture device, and these profiles are embedded in the appropriate TIFF masters. Users can use these profiles to view the master files in a calibrated viewing system, and to convert them into standard color spaces such as Adobe RGB or sRGB.

Cropping and background
Originals are depicted entirely, including blank margins, against a suitable background paper (usually white or gray) so that the digital image documents the physical artifact, as well as reproducing the imagery that the artifact portrays. A narrow gap between the grayscale target and the original allows for cropping the grayscale out of the composition if desired for some new purpose.

Storing the digital masters
The digital capture files are grouped on a local server for quality review, technical metadata preparation, and viewing file preparation, and are then archived on Library servers and on removable hard drives.

Making the derivative files
Viewing files are made from the masters using scripted actions in Adobe Photoshop. The master files are opened, converted to sRGB colorpace, downsampled to size, sharpened with unsharp mask, and saved in GIF and JPEG formats. The viewing files are intended for viewing on a typical computer monitor, as represented by the sRGB colorspace.