Acquiring Digital Images

Digital images are obtained most commonly by one of three methods: scanning, use of a digital camera or, by obtaining an existing image file.  In all cases copyright rules apply. 
 
Creating your own digital images requires numerous crucial technical decisions. Poor decisions can result in digital files that are inadequate for their intended uses.  By anticipating the most demanding end-use before the image is digitized, it is possible to choose specifications that will result in satisfactory image quality without squandering resources such as time or network storage

Scanning

Scanning is quite complicated by the fact that there are many different types of scanners and associated software.  All have their own settings and user interface.  However regardless of scanner type or brand, there are several basic concepts to keep in mind that will make scanning easier. 

  • Use the right type of scanner - The selection of a scanner will have the greatest impact on quality of images. It is a challenge to select the right scanner as technology, pricing and scanner features changes nearly everyday. Which scanner is right for your project depends on numerous factors including overall project goals and the format, size, and condition of materials to be scanned, as well as available budget.
  • Optical resolution - The optical resolution of scanner should exceed the maximum resolution needed.  For example, a flatbed scanner with an optical resolution of 1200 ppi has sufficient optical resolution to scan an 8x10” print at 600 ppi, but insufficient optical resolution to scan a 2x2 slide at 4000 ppi.

Most scanners are advertised with high resolution.  This is a little misleading in that often these higher resolutions are interpolated; they use a sophisticated guessing system.  Therefore, it is important that you know the true optical resolution of the equipment.

  •   Bit depth -The greater the bit depth, the more information about the source is captured by the array, resulting in a more accurate digital representation of the original. A bit depth of 8 can capture enough information to represent 256 colors or shades of grey. A bit depth of 24 (8 bit channel Red + 8 bit channel Green + 8 bit channel Blue = 24 bits) captures over 16 million colors or shades of grey. Since higher bit depths capture more information, they increase the overall file size of an image.
  • The best resolution to scan your images depends on how it will be used - on screen or in print
  • Digital image quality is dependent upon the source material scanned.  High resolution scanning will not improve the quality of a poor image.
  • Scan once.  If the image needs to be resized after the scan, it's always much better to resize to reduce the image size rather than to resize to increase the image size. If you aren't sure what your future intentions for the image might be, and won't be able to scan it again, then err on the large size. Resizing to a smaller size discards excess pixels. But resizing to a larger size must create new interpolated pixels which were not in the original scan. The results are not at all the same as scanning at the higher resolution.
  • Make color correction with a photo editing program such as Photoshop.  Do not make corrections using the scanner software.
  • Calculate the final image resolution.

There is a direct relationship between resolution and image dimensions.  If a digital image is to be outputted at a size four times greater than the original, then the capture resolution should be four times greater than the final output resolution.

To calculate capture resolution for known output:
Longest dimension of desired output (in inches) multiplied by desired output resolution (in pixels) divided by longest dimension (in inches) of original.

EXAMPLE:
To print an 8”x10” image at 300 dpi, scan the original (at 100%) as follows:

  • 8”x10” print at 300 ppi (10x300/10)
  • 1”x1.5” slide at 2000 ppi (10x300/1.5) 

To calculate capture resolution based on desired pixel dimension:
Longest dimension of desired output (in inches) multiplied by desired output resolution (in pixels) divided by longest dimension (in inches) of original.

EXAMPLE:
To print an 8”x10” image at 300 dpi, scan the original (at 100%) as follows:

  • 8”x10” print at 300 ppi (10x300/10)
  • 1 ”x1.5” slide at 2000 ppi (10x300/1.5) 

To calculate capture resolution based on desired pixel dimension:
Longest dimension (in pixels) of digital image divided by longest dimension of original (in inches) equals capture resolution.
 
EXAMPLE:
For a longest pixel dimension of 4000 pixels, scan

  • 1”x 1.5” slide at 2670 ppi (4000/1.5)

Scanning Standards

See above

File Format and Resolution

There are two basic types of graphic files: vector and raster.  Every file format was developed to serve a specific purpose.  Some formats provide good compression, others scalability, and many work better with a particular operating system or software vendor. 
 
Vector graphics are based on the mathematical calculations of lines and curves.  They are scalable without any loss of resolution.  Vector images are usually line drawings, 3-D models, etc.  Vector graphics are created in drawing programs such as Adobe Illustrator, Corel Draw, Quark.  Vector images have extensions such as .ai, .cdr, .dxf, .cgm, .eps.
 
Raster images are made up of a fixed grid of pixels. Generally they are not scalable.  The quality of a raster image is determined by the total number of pixels (resolution), and the amount of information in each pixel (often called bit depth).  Photographs are always raster images, as are images scanned or acquired by a digital camera.  File extensions include .tif, .jpeg, .gif, and .bmp.
 
Resolution is a measurement of the density of the pixels in a given area.   For example, a 100 pixels-per-inch (ppi) image will have a grid of 10 pixels across and 10 pixels down in every inch; a 400 ppi image will have a grid of 20 pixels across and 20 pixels down in each inch.

Resolution Standards

Common Image Uses Typical Resolutions Bit Depth
   

Color

Greyscale

Master/archival—Recommended when long-term use is uncertain. 600 dpi (or 4000-6000 pixels across the longest dimension 16 bit 8 bit
Large format printing
(16”X20” and bigger)
150 dpi 16 bit 8 bit
Printing 300 dpi 16 bit 8 bit
Web 72 dpi 8 bit 8 bit
Presentation 72 dpi 8 bit 8 bit
Presentation: detailed image study 72 dpi, however image dimensions should be increased 16 8
Art print 1200 dpi 16 bit 8 bit

File Format Standards

  • To ensure long-term accessibility, master files should be saved in uncompressed TIFF (.tif) format.
  • Derivative files to be used for presentations or Web display should be saved in JPEG (.jpg) format.  Note that compression in JPEG files causes some data to be lost each time the file is saved, resulting in image deterioration over time. 
  • Derivative files to be used for printing should be saved in TIFF (.tif) format.
  • Many cameras use proprietary RAW file formats which complicates the conversion to useable formats. It is not recommended that master files be saved in RAW format.

Master files

An image should be captured once (the “master” image), with quality sufficient to meet current and future needs. Copies of the file (“derivatives”) can be generated from the master image for any number of uses including printing, display, and image processing.
 
For example, if the immediate use of an image capture is for a presentation, then a low resolution file is needed.  However, if there is a probability that the file will be professionally printed in the future, then a high resolution file is required.  In this case, it is best to capture the image at the higher resolution and create a derivative low resolution file for the presentation.
 
A digital image can only have as much detail as the original — not more.  For lower quality originals, such as newspaper clippings, there will a point after which increasing capture resolution will not improve the image quality, but will increase the file size unnecessarily. Consider each original source first and then select an appropriate resolution accordingly.
 
Each image regardless of capture method should be saved as a master file without any compression, color correction or any other correction.