The preprocessor variable "DOUBLE_PIXELS" causes the application to be built with double precision for all its internal calculations, if it is defined. Otherwise, the default is to use single precision.

For 8 bit images (webcam) the double precision is needed if the list of images to process exceeds 16384 elements (if you need this let me know ;o). For 12 bit images (DSLR) the double precision is useful for processing more than 1024 images... And at last, for 16 bits images (astronomical CCD) the double precision is useful because the the limit is only 64 images.

Whatever precision Lynkeos is compiled with, it opens Lynkeos documents saved with any precision.