Digitizing Text and Images

There are 3 types of files to consider in the digitization process:

Master File-Used for archival purposes; should remain unedited

Access File-used for web display

Thumbnail -small, low resolution image, often displayed with item record

If possible, one of each type should be created for each digital object, but issues such as cost, time, storage, and purpose of the digital object must all be considered when deciding whether or not master files are desired.

Master Files

If you plan to modify, re-use, or print high quality reproductions of an image after initial digitization, it is strongly recommended that you create a master file that will remain unaltered.   Varying sizes and resolutions can then be derived from that one.   Our master image files are typically saved in Tagged Image File Format (TIFF).  TIFF is not proprietary, and lends itself to long term preservation.  Most scanners typically allow files to be saved in TIFF format, as do some digital cameras.  If a digital camera does not allow files to be downloaded and saved as TIFF, use the highest resolution setting possible.  There are emerging file formats, such as JPEG 2000 and Motion JPEG 2000, that offer lossless compression, which means smaller file sizes with no discernable loss of quality.  One current drawback of such formats, and the main reason we do not currently use them at our institution, is that they are not widely supported by most available hardware; therefore, they are often difficult and expensive to create.

Access Files

An access file is typically used for web access and display on a computer monitor.  File sizes are considerably smaller than master files, which allows for faster transmission and download.  The typical file format for access image files is JPEG.  If a JPEG image is saved and re-saved, the image quality will degrade with each generation.  The preferred format for text-based access files is PDF.

Thumbnails

A thumbnail is a very small JPEG or GIF image that is often displayed alongside a bibliographic record if one exists.  A thumbnail acts as sort of an index print that allows the viewer to quickly decide if they want to see the access image.  Thumbnails are usually only useful when dealing with image files, not text. 

Scanning Guidelines

About Scanners

Most scanners can operate in one of three modes depending on types of materials you are working with: 

1. Bitonal-Black and white; Good for black and white, text-only documents
2. Grayscale-Good for black and white photographs
3. Color-Good for full-color image or text documents

Aside from choosing which mode to capture images in, it is also important to consider resolution, bit depth, and dimensions of the digitized image.  Resolution is measured in dots per inch (dpi) or points per inch (ppi)—both terms are interchangeable.   A higher ppi means a higher resolution.

Bit depth refers to the number of colors available to reproduce the digitized image.  Greater bit depth means more colors are used.

1 bit (2¹⁾ = 2 tones-Bitonal

2 bits (2²) = 4 tones

3 bits (2³⁾ = 8 tones

4 bits (2⁴⁾ = 16 tones

8 bits (2⁸) = 256 tones

16 bits (2¹⁶⁾ = 65,536 tones

24 bits (2²⁴ ) = 16.7 million tones-true color

48 bits (2⁴⁸) = billions of tones

It may be worthwhile to scan the back side (verso) of a photograph as a separate object as well, particularly if there is information there that does not exist anywhere else.  The decision of whether or not to scan the verso will depend on the nature of a given collection, the needs of the collection creator, and those of the collection’s intended audience.

Choosing the Correct Resolution

  The proper resolution at which master files should be scanned is dependent upon the size of the source material.   For example, master quality scans of 35mm slides should be 4200 pixels on the long side.  Determine correct scanner resolution in ppi by dividing 4200 pixels by 1.5, because the long side of the physical 35mm slide is about 1.5” (4200/1.5=2800) so when creating a master file scan of a 35mm slide, scanner resolution should be set at 2800 ppi.  The same formula works for photographs as well.  The long side of scanned images should be 4000 pixels, so if the original item is a 4x6 photograph, the formula is 4000/6=667.  If the scanner cannot be set to an odd ppi, choose the closest higher setting.