Selecting the appropriate hardware for digital scanning of materials is critical for high-quality results. Different types of scanners are required to scan different materials, such as paper documents, photographs, films, and transparencies. Scanning specifications including resolution, color/grayscale depth, optical density, internal sensor and processing, and driver software are important for generating high-quality digital scans. Specialized scanners are required for digitizing unique media, such as films and transparencies, which require a scanner with a high optical density, resolution, and mounting system. Due to the high dynamic range of photographic films, particularly black and white, use of a scanner with a Dmax (maximum absolute density) of 3.8 or greater is needed for best results.
Types of Digital Scanners
Specialized digital scanners are available to handle general documents, large-format, film, and unique media. Specialty models, including drum film, aerial film, and sheet-fed well log, are not discussed herein. Selecting the appropriate scanner for a specific project workflow is critical for producing high-quality, cost-effective scan products.
Document Scanners – available in simple flatbed to high-speed sheet feeders, generally up to U.S. legal size (8.5”x14”), for scanning a variety of paper documents. Inexpensive, consumer models are unsuitable for most data preservation scanning, due to their lower quality sensors and optics, slower speeds, limited driver software functionality, and decreased reliability (known as mean time between failure (MTBF)) than commercial models. High quality scans are critical for accurate optical character recognition (OCR) of text.
Large-Format Scanners – available in flatbed and sheet-fed models for scanning variety of media types and sizes and ideal for large maps, drawings, and photographic prints.
- Flatbed: Large-format flatbed scanners typically have a maximum imaging size of approximately 12” by 17”, maximum optical resolution of approximately 2400 dpi, transparency capability, and are suited for scanning larger paper documents and maps, photographic prints, and aerial photographs. The transparency features enable the scanning of Mylar sheets and cut films.
- Sheet-Fed: Most sheet-fed scanners allow imaging widths of 36 to 42 inches; few models allow widths up to 54 inches. Maximum optical resolution for large format scanners is typically less than 800 dpi. Large scans at higher resolutions would generate extremely large and unwieldy file sizes.
Film Scanners – available in mounted slide/filmstrip and flatbed models, film scanners are the preferred hardware to scan film positives and negatives, due to their mounting systems, higher resolution and optical density, and imaging quality.
- Mounted Slide and Filmstrip: Several types of film scanners are available for mounted slides or filmstrips mounted in a specialized carrier provided by the manufacturer. Most film scanners can perform batch scanning for a stack of slides or series of frames within a filmstrip. Cardboard mounted slides that have warped or curled must be processed carefully as they can become stuck and damaaged in automatic feeding systems. Warped, curled, or separating cardboard mounts can be replaced with plastic mounts for ease of scanning and long-term preservation; however, care must be taken to prevent film damage from finger oils and scratches.
- Flatbed: High-resolution, high optical density, and dimensionally accurate (low distortion) flatbed scanners are the preferred hardware to scan unmounted film negatives and positives at a significantly lower cost than drum scanners. Optical resolutions greater than 6000 dpi are often available, with Dmax values typically ≥ 4.0.
Digital Scanner Specifications
To select the appropriate scanner for a project, reference the manufacturer-provided specification sheets, which provide more accurate information about scanner capabilities than marketing web pages. Web pages often do not provide full specifications, such as optical density, that are critical for informed selection of appropriate scanners.
Resolution – is the ability of a scanner to resolve detail in the scanned object, and reported as optical and/or interpolated resolution in dots per inch (dpi) or pixels per inch (ppi). Optical resolution is determined by the optics and sensor of the scanner. Interpolated resolution is determined by a multiplicative mathematical process, and results in images with no additional detail than the scanner’s optical resolution.
Optical Density – commonly expressed as the Dmax or the maximum density that can be recorded by a scanner sensor, and refers to the dynamic range of light intensity (pure black to pure white). Larger values of Dmax result in scans with more detail in shadows and very light areas, while lower values of Dmax have limited or no detail in shadows and light areas. Since photographic films have a wide dynamic range of light intensity, particularly black and white films, a Dmax of 3.8 or greater is often used.
Color or Greyscale Depth – is the number of individual levels of color or grey that a scanner can image and reported as bits per pixel or bits per channel/color. The table below provides the color and greyscale bit depths and levels of colors and grey shades. In a typical RGB color space of a scanner, 8 bits per channel is 24 bits per pixel or 16.8 million colors or shades of grey (levels). Common bit depths include 1 bit (duotone or black/white) for text or line drawing only documents, 8 bit (256 colors or greys for general documents and drawings), 16 bit (65,536 colors or greys), 24 bit (16.8 million colors or greys for detailed photographs and related images), or 48 bit (for specialized image processing). Note, the sensor (internal) bit depth may not be the output (external) bit depth. If the internal and external bit depths differ, the image is being processed by the scanner hardware.
Color Space/Profile– the gamut or range of reproducible colors and defined by standards, such as AdobeRGB (1998) and sRGB, and helps to reproduce accurate color on a variety of devices. Choosing a color space is required if color calibration is used. The AdobeRGB color space was developed to encompass most reproducible colors visible to the human eye; whereas the sRGB color space was developed primarily for computer displays and consumer devices, and can encompass fewer reproducible colors. As a result, the AdobeRGB color space should be used whenever possible for color images.
Suitable scanner software should be used that allows for scanner color calibration using targets (reflective and transparent) to develop International Color Consortium (ICC) color profile files to retain consistent color across a range of devices (scanners, printers, monitors). The Gray Gamma 2.2 color space is the most common for grayscale images.
Output File Format – no compression or loss-less compression file formats, such as TIFF and PDF, should be available options in the scanner driver software. If any post-scan processing is to be performed, lossy compression formats, such as JPEG, should not be used, due to compression degradation that occurs each time a file is saved.
Digital Scanner Software
Scanner software is as important as the scanner itself. Without proper drivers and scanning control software, the resultant scans will not be of high quality and reproducibility. While nearly all modern scanners include driver software, most do not have sufficient control of the scanner hardware and scanning specifications for high-quality results. Scanner software should be able to adjust typical parameters including resolution, bit depth, and mode (reflective or transparent), along with imaging focus (when possible via hardware), output color profile, and allow for scanner calibration using targets.
SilverFast from LaserSoft Imaging (http://www.silverfast.com/) is an excellent option for a variety of current and older scanners. SilverFast includes a variety of useful features, which may be limited by individual scanner capabilities, including infrared-based dust and scratch removal (alternative to Digital ICE Technology), automatic IT8 color calibration (IT8 is a set of American National Standards Institute standards for color specifications), focus control, multi-exposure for higher dynamic range, and correct negative to positive conversion (including accounting for common characteristics such as an orange mask inherent in color negative film).
Color correction is often needed for a variety of paper and film photographs, particularly those not stored in ideal conditions, which may degrade image quality. Photographic slides (positive film) often show significant color degredation and/or casts that may be corrected in distribution copies. ColorPerfect from CF Systems (http://www.colorperfect.com) is an inexpensive Adobe Photoshop plug for color correction of various photographic images.
Embedded file metadata may be easily viewed using the open-source ExifTool (http://owl.phy.queensu.ca/~phil/exiftool/) software for a variety of file formats.
Digital File Storage
Digital files from scanning workflows should be stored to appropriate archival storage devices that may include: enterprise storage area networks (SAN), networked attached storage (NAS), storage arrays, archival tape backup systems, and hard drives. Any device or system used should have an appropriate backup/duplication process in place to prevent data loss and a plan to migrate data before the storage, operating system, cabling/connectivity, or storage format becomes obsolete. CD and DVD disks are not acceptable archival formats, due to the limited storage size, questionable lifetime of the dye-based system used, and are quickly becoming legacy storage technologies. Flash memory drives are also not a long-term storage solution, due to the likely possibility of data decay in the memory itself. Purchasers of hard drive-based storage should consider the use of true enterprise-grade hard drives that have a significantly higher mean time between failure rate and overall quality.
Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government.