When ordering digital scans there are several parameters that must be addressed.
RESOLUTION
The resolution of a digital image can be described using three basic methods, all making reference to the size of the pixels that make up the image. A pixel is one “dot” of information that, together with a lot of its buddies, helps to form the image. Descriptions of resolution can use the following terms:
Pixels Per Inch (ppi): This is the most widely used notation for digital resolution, and the most useful for translating a digital file into a physical photograph. It is notated as a number representing how many pixels there are in one linear inch - simple, right? Wait, there’s more. When making a print from a digital file the resolution of the printing device (dots per inch or dpi) must be taken into account so that there will be enough pixels in the original scan to cover the full size of the enlargement at the print resolution.
For example:
A 3”x 4” section of film is to be printed at 30”x 40” – an enlargement factor of 10x. Our photographic printer is optimized for printing at 250dpi. In order to maximize the quality of the print, we would scan the area of interest at a minimum of 2500ppi (10 x 250).
Where enlargement factors exceed the maximum resolution of our scanner it is possible for us to interpolate more pixels and generate higher resolution files. This works adequately up to a point, but because this amounts to the computer “making up” pixels, we use this only when absolutely necessary.
Note on ppi vs. dpi: In general the term ppi refers to a digital image, while dpi refers to a physical print. In many cases the number of pixels is interchangeable with the number of dots, but not always.
Micron (µm): The second way of denoting image resolution is to measure the physical size of the pixel. The unit commonly used is the micrometer or micron. A micron is one millionth of a meter (0.00004 inches) and symbolized by µ or µm. Standard resolutions for photogrammetric scans range from 12 to 20µm.
A measurement of pixel size in microns is directly translatable to ppi, because both terms measure the physical size of the pixels – a 20µm scan is always 1270ppi.
The third descriptor for resolution refers to the size of the pixel, but also takes into account the scale of the image being scanned.
Pixels in Meters or Feet: This unit of measurement refers to the actual distance on the ground that is covered by one pixel of the image. It is relative not only on the resolution of the image, but the scale at which the image was shot.
For Example:
If an image is shot at a scale of 1”=500’, then scanned at 1000ppi, the effective resolution is a 6 inch pixel.
500 feet / 1000 pixels = .5 feet / 1 pixel = 6 inch pixel
If the image were scanned at 2000ppi we could get to a 3 inch pixel, and so on.
In reality, there is an upper limit imposed on the pixel resolution by the fineness of film grain and the sharpness of the camera lens. Despite the mathematical formula, no amount of high resolution scanning can resolve 6 inch pixels from photography shot at 1”=3000’.
What resolution do I need?
For the end user the question of resolution is one of balancing the detail available in the image versus the ability to use the file in a reasonable work environment. Our photogrammetric scans can be created with a maximum resolution of 12 microns, or approximately 2100 pixels per inch (ppi). Because a full frame of aerial film is 9”x 9”, this generates a file size in excess of a gigabyte. Scans at this resolution will capture nearly all of the detail that is contained on the film, but may present problems when using a standard desktop PC.
RGB file size is determined by the formula ((h x ppi) x (w x ppi))x3 where h=height in inches, w=width in inches, and ppi = pixels per inch. This will calculate the image size in bytes.
For Example: A 9”x 9” image is scanned at 500 ppi.
(9x500) x (9x500) = 20,250,000 This is the number of bytes per channel
20,250,000 x 3 = 60,750,000 This is for color images only as there are three channels – Red, Green, Blue. A black and white image contains only one channel and is therefore one third the size of an RGB image.
Size in MB can be estimated by sliding the decimal six places to the left (dividing by one million). The file size will be approximately 60.75 MB.
Because computers operate in binary numbers instead of base ten, this estimate will be slightly high. To derive a more accurate file size, divide by 1,048,576 (or 10242) to arrive at a file size of 57.9 MB.
The chart below shows file sizes for full frame scans at several resolutions.
| PPI | Full Frame File Size | PPI | Full Frame File Size | |
| 160 | 2.5 MB | 1270 | 432 MB | |
| 500 | 58 MB | 1500 | 521 MB | |
| 850 | 167 MB | 1800 | 751 MB | |
| 1000 | 232 MB | 2000 | 927 MB | |
| 1200 | 334 MB | 2540 | 1728 MB |
These file sizes can be reduced for storage by utilizing compression file formats (issues involving file compression are dealt with in the next section). In most cases however, image software will have to manipulate the image at full size. Image Viewer, Paint, Irfanview or other basic image software will not easily process files above 500 MB, depending on hardware capabilities. For higher resolution files we recommend an updated version of Photoshop or mapping software designed for large files.
FILE FORMATS AND COMPRESSION
There are a variety of file formats available for digital imagery. In most cases, the format is determined by the end needs of the customer. What are the capabilities of the workstation? How much storage space can be dedicated to the files? Is there a native format that operates best with available software?
BPS can generate image pyramids, also known as minifications, for scanned images. These are a series of lower resolution files derived from the full resolution original. Within a software environment that supports them they allow rapid access to different magnifications. These files can be saved as extra files or embedded within the primary image file.
-
Some common file formats are as follows:
- JPEG - The file format most commonly used in consumer digital cameras. JPEG files can highly compressed, allowing for a large amount of information to be stored in a relatively small disk space. JPEG compression however, is considered “lossy” in that repeated resaving of the file will degrade the image. It is not recommended for the storage of commonly used files.
- JPEG 2000 - This is an updated version of the original JPEG format. It is intended to be more a more robust file format, supporting both lossy and lossless compression and reduced artifact creation. Legal disputes regarding its core wavelet compression engine have delayed widespread implementation of the format.
- TIFF - This is an acronym for Tagged Image File Format. In its uncompressed form it is considered to be a “lossless” format. All of the information contained in the original file is maintained every time the file is transferred or resaved. It is also one of the bulkiest ways to store data.
TIFF files can be compressed using a JPEG algorithm that greatly reduces the amount of disk space required to store the image. Compression levels are denoted with a Q number ranging from 1-100, higher numbers indicating lower amounts of compression. We have found that a TIFF with Q90 JPEG compression does not visibly suffer data loss and creates a file close to ¼ the size of an uncompressed TIFF. 90% of our customers choose some variant of the TIFF JPEG format.
TIFF files can also be “tiled”. This is a process of dividing high resolution images into regularly sized tiles for compression and storage. It allows for faster access and more efficient compression.
For more than anyone ever wanted to know about TIFF files, the official standards are available here. A less wordy and possibly less accurate description is available on Wikipedia.
- ECW - ECW (Enhanced Compression Wavelet) is an open format supported by several major image software products. It uses “wavelet compression” which involves a great deal of higher math and can be explored in detail here and here.
This format allows for extremely large files to be accessed and manipulated in a normal desktop environment without excessive RAM requirements. It is able to decompress portions of an image while the rest remains compressed, allowing the panning of a “zoomed” view without needing to raster process the entire image. At high compression values the format is very lossy, though using it can yield acceptable results.
- Vitec - This is the native image format used by Leica scanners. Some customers utilizing Leica photogrammetry software suites opt to use this format.
