When the AIR package is compiled, it is compiled either as an 8 bit or as a 16 bit version. The information on this page describes the pixel value remapping behavior of AIR when it is compiled in 8 bit format. All 8 bit versions of AIR represent pixel values internally using numbers ranging from zero to 255. Your image generation and display software may require that data be represented differently, so the AIR package allows pixel values to be systematically remapped when loading data from disk. Pixel values are not modified by an 8 bit version of AIR when saving data to disk.

Loading data from disk into an 8 bit version of AIR

When loading type 0 data (8 bit data) from disk, an 8 bit version of the AIR package will:

• make no adjustment

When loading 16 bit type 1 data from disk, an 8 bit version of the AIR package will:

• multiply by (65535 / (global max))
• discard the fractional component
• divide by 256
• discard the fractional component
• note that "global maximum" is stored in the image header and can be changed using the program setheadermax

When loading 16 bit type 2 data from disk, an 8 bit version of the AIR package will:

• set negative values to zero
• multiply by 2
• multiply by (32767/(global max))
• discard the fractional component
• divide by 256
• discard the fractional component
• note that "global maximum" is stored in the image header and can be changed using the program setheadermax

When loading 16 bit type 3 data from disk, an 8 bit version of the AIR package will:

• add 32768
• multiply by (65535/(global max+32768))
• discard the fractional component
• divide by 256
• discard the fractional component
• note that "global maximum" is stored in the image header and can be changed using the program setheadermax

Saving data from an 8 bit version of AIR

All data (except for binary files) generated by an 8 bit version AIR package will be saved to disk as 8 bit values ranging from zero to 255. Consequently, the output data type (always 8 bit) will not necessarily match the input data type (which can be any 8 or 16 bit data type).

• no adjustment is made