Although I am not sure I am looking at the exact same help reference, I believe when the ENVI help refers to the reference pixel x/y (in Working with Header Files), it is referring to 'image coordinates' and not 'file coordinates'. There is a pretty good discussion of the two in the help if you search for 'Coordinate Systems in ENVI' (in the Additional topics in ENVI Programming). So with regards to your questions: 1) Aren't the reference pixel x and y locations actually one-based with firrst pixel being 1,1? 1a) When working in the ENVI gui or generating header files, you specify the reference pixel in image coordinates, which are 1,1 for the upper left corner coordinate. When working with the file coordinates programmatically, it will be using zero-based numbers (as in an IDL array), and thus the first pixel's file coordinates are 0,0. 2) Won't using a single point reference (map info) and calculating the other three corners from the image size and pixel size yield inaccurate georefencing? Any inaccuracy in pixel size will be magnified by repeatedly adding it to reference point to calculate other points. 2a) Map info should only be defined for images that are in a true map projection thus the pixel size should be constant across the image. When you have map info, you have a reference pixel, a pixel size, and a projection system. From this point, all ther points can be accurately calculated. Now of course, if your pixel size *is* inaccurate or wrong, you will propogate this error across the image. The assumption is that the map info is accurate and correct. 3) The menu selection Edit ENVI Header/Edit Attributes has item Geographic Corners which generates keyword "geo points" containing the geographic locations for the four corners of the data. In general, will these points yield more accurate georeferencing or are they just the same values you can calculate using the map info data? Is editing the header the only way that the geo points are stored in the header file? 3a) 'Geographic corners' are separate and different than Map Info. You can have the geographic locations of the 4 corners of the image defined but they are not the same as having 'Map Info' which means the image has been 'warped' into a true map projection. There are occasions when the 'geographic corners' are populated while the map info is not. This just provides the user a rough estimate of the location of the file but should in no way be more accurate than an image that is truly projected. When the four corners are defined, you can sometimes have map info which shows 'Pseudo Geographic' but again, it should not be confused as a true or accurate projection system. Now if you project your image, or have map info defined, then ENVI will estimate the 4 geographic corners from this information. I hope this is helpful!